Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
It'll be fascinating to see if Labour does bite the bullet (sorry) and make a significant increase in defence spending. I mean, given the noises off about lack of cash and the ever ballooning health and pensions bills it doesn't seem likely, but perhaps those who insist that everything will be different once they've dispensed with the need to lure the grey vote into backing them will be proven right?
Given our geography, even if the Government feels it can't afford the spectrum of capabilities that it would like, you'd at least think it could be persuaded to fork out for a functioning navy?
More like use the existing money efficiently, rather than buy frigates without fixed antisubmarine torpedo tubes and ear-killing armoured vehicles when perfectly good designs exist off the shelf from other suppliers.
Do any new frigates have fixed torpedo tubes these days?
I thought it was either via helicopter, or rocket from a vertical cell, in latest versions.
Bog standard, it seems. DA commented, when we were discussing the omission in a recent RN class, that it was an essential backup if the ship's helicopter was ill or the sea state was too rough to launch the helo. And AIUI torpedoes are often wire guided - not something thast can be done with a rocket-delivered one.
What's the backup when they're in port because we can't afford to crew or repair them ?
Another Brexit dividend - my 11yo daughter just got a new passport delivered, and was so disgusted that it was a blue one that she threw it on the ground, where it landed on my toe!
An 11 year old cares about passport colours? What colour was she hoping for?
She liked the colour of the old one.
And that's partly why the current settlement isn't settling. One of the factors in 2016 was the nostalgic one, for the good old days before Ted Heath. Doesn't matter what the balance of good and bad is, nostalgia just is. People who grew up with blue passports wanted them back, people who grew up with burgundy will feel the same. Why shouldn't they? It may be silly, but it's also human.
As time passes, the ratio of nostalgia for pre-EEC and nostalgia for in-EEC/EU will evolve in a predictable way.
Indeed, the blue/black one just looks weird to me. Burgundy is a much nicer colour, and it's just the colour that passports should be. I liked the way that the burgundy one used to let me travel and work freely in a single market of half a billion people too, but I guess that's just me.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
This might go down well with his brain dead base but just highlights to everyone else why he should be nowhere near the WH .
He is on bail, having been inducted for four different crimes. Incitement of violence against the president if a federal offence, and a blatant violation of his bail conditions. Any other individual would be in custody by now.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
“I've just run some planning casework through Claude. It identified within a few seconds the grounds on which a legal challenge could be pursued. This is a skill that only the most experienced people working in the industry have and they will have built it up over careers of many years.”
Indeed. Claude is completely extraordinary. I’ve just been chatting with my lady novelist friend. I told you that she fed her latest novel into Claude and it gave her a detailed summary and analysis in forty five seconds. 😶
Now she has given Claude her editor’s feedback - several pages of deep advice - and Claude has decided the editor is right on this, this, this and not this. All in a few seconds. Stunningly good and incomparably fast
We’re chatting now. I had an idea for her novel and I messaged her a moment ago with the idea. Her reply: “Claude just suggested exactly that”
True story
What’s terrifying here is that if Claude can be this good at editing then I see no reason why Claude cannot actually write books on top of editing them. Given that it understands plot, tension, narrative, character, everything
This is Claude 3. Unless progress suddenly stops - why? - clause 6.5 will write brilliant novels. In seconds
@leon. Indeed - I was using GPT before but Claude is a game changer. After last nights output I have been hypothetically planning out all morning how we can completely redesign the service at work based on this technology. Productivity and accuracy could increase exponentially by effectively outsourcing about 80% of the work in report writing. The hard part of checking written output can also be done through AI. It creates vast opportunities to improve 'customer service' by moving resource in to customer facing roles and deliver on enormously neglected areas of the service.
My question is that how would I build a system that generates these reports that isn't in a chat format, IE I get Claude to 'learn' various rules set out in policy documents, then generates text based on variable inputs that the operator gives it? I've seen some people tout these things on linkedin, how do you build these systems?
(I am not in charge of the department - the above is all a thought experiment). Very interested in feedback.
Pay attention to your own IP (intellectual property). Do your customers (and directors, and regulators) know you are freely sending it outside? As a corollary to that, are you making arrangements to self-host AI?
That is a good point - I can't say too much but for the purposes of what I have been feeding to AI, all the data is open source and publicly available, so not disclosing any protected IP. One hazard with these AI systems though is they are recording all the data you give to them forever and they can certainly also identify you.
And they are using your prompts to train on as well, very often.
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
Can I ask, with these people being turned away at the airport because of impending out of date passports - didn't they think to check the entry requirements for the country first?
When I went to Malta, I checked and they wanted a passport valid for six months. Mine had only three left, so I replaced it before checking in.
Not really hard, and would have saved any hassle at all.
I made this point on the last thread
Since 2003 when our son emigrated to New Zealand we have travelled extensively world wide and this rule applies so we renewed our passports accordingly
There is a tedious nature of the continual bewailing of our previous membership of the EU, not least because 'we have left and there is no prospect of us rejoining the single market or freedom of movement'
The quotes are by one Sir Keir Starmer just this week, and reaffirmed by the EU as well at the same time in connection with the review of the treaty next year when they reaffirned they will not reopen UKs membership
It is a forlorn hope but time to move on on the EU membership and to be fair it seems Starmer agrees
Nah. Rejoin is just off the agenda for the next 5 years, not forever.
It's impossible to resist the polling forever.
Rejoin: 48% (=) Stay out: 32% (=)
How long do you think polling on the death penalty can be resisted?
Pretty well forever. Until 60% of the electorate actually care about it enough to make a difference to the way they vote (they don't), nothing will happen.
Similarly with the EU - though the political salience of that is possibly rather more. It certainly was a decade ago, even if for now it's quiescent, But it's managed completely to wreck the Tory party in the meantime.
I think putting a Referendum on the death penalty in the Tory Manifesto is pretty much their best weapon in GE2024
That just shows what a hopeless state they're in. Still, it would at least be a clear test of how important the issue is to electors.
"Vote for us: we're shit, but we'll kill a handful of people" isn't a particularly convincing platform.
Well yes but I guess from their perspective desperate times desperate policies.
What's the betting SKS would say we will offer one too. Would definitely be good news for LDs/ Greens if so. #Hangmanslackenyournoose
Labour will kill more, sooner. Only Labour can be trusted with State human disposal
Progressive politics is all about having unborn babies put to death, not murderers.
Can I ask, with these people being turned away at the airport because of impending out of date passports - didn't they think to check the entry requirements for the country first?
When I went to Malta, I checked and they wanted a passport valid for six months. Mine had only three left, so I replaced it before checking in.
Not really hard, and would have saved any hassle at all.
I made this point on the last thread
Since 2003 when our son emigrated to New Zealand we have travelled extensively world wide and this rule applies so we renewed our passports accordingly
There is a tedious nature of the continual bewailing of our previous membership of the EU, not least because 'we have left and there is no prospect of us rejoining the single market or freedom of movement'
The quotes are by one Sir Keir Starmer just this week, and reaffirmed by the EU as well at the same time in connection with the review of the treaty next year when they reaffirned they will not reopen UKs membership
It is a forlorn hope but time to move on on the EU membership and to be fair it seems Starmer agrees
Nah. Rejoin is just off the agenda for the next 5 years, not forever.
It's impossible to resist the polling forever.
Rejoin: 48% (=) Stay out: 32% (=)
How long do you think polling on the death penalty can be resisted?
Pretty well forever. Until 60% of the electorate actually care about it enough to make a difference to the way they vote (they don't), nothing will happen.
Similarly with the EU - though the political salience of that is possibly rather more. It certainly was a decade ago, even if for now it's quiescent, But it's managed completely to wreck the Tory party in the meantime.
I think putting a Referendum on the death penalty in the Tory Manifesto is pretty much their best weapon in GE2024
That just shows what a hopeless state they're in. Still, it would at least be a clear test of how important the issue is to electors.
"Vote for us: we're shit, but we'll kill a handful of people" isn't a particularly convincing platform.
Well yes but I guess from their perspective desperate times desperate policies.
What's the betting SKS would say we will offer one too. Would definitely be good news for LDs/ Greens if so. #Hangmanslackenyournoose
Labour will kill more, sooner. Only Labour can be trusted with State human disposal
Progressive politics is all about having unborn babies put to death, not murderers.
PJ O’Rourke commented that it was interesting that you don’t find (in the US)
1) anti-abortion activists who are anti death penalty. 2) pro death penalty people who are pro abortion
“I've just run some planning casework through Claude. It identified within a few seconds the grounds on which a legal challenge could be pursued. This is a skill that only the most experienced people working in the industry have and they will have built it up over careers of many years.”
Indeed. Claude is completely extraordinary. I’ve just been chatting with my lady novelist friend. I told you that she fed her latest novel into Claude and it gave her a detailed summary and analysis in forty five seconds. 😶
Now she has given Claude her editor’s feedback - several pages of deep advice - and Claude has decided the editor is right on this, this, this and not this. All in a few seconds. Stunningly good and incomparably fast
We’re chatting now. I had an idea for her novel and I messaged her a moment ago with the idea. Her reply: “Claude just suggested exactly that”
True story
What’s terrifying here is that if Claude can be this good at editing then I see no reason why Claude cannot actually write books on top of editing them. Given that it understands plot, tension, narrative, character, everything
This is Claude 3. Unless progress suddenly stops - why? - clause 6.5 will write brilliant novels. In seconds
@leon. Indeed - I was using GPT before but Claude is a game changer. After last nights output I have been hypothetically planning out all morning how we can completely redesign the service at work based on this technology. Productivity and accuracy could increase exponentially by effectively outsourcing about 80% of the work in report writing. The hard part of checking written output can also be done through AI. It creates vast opportunities to improve 'customer service' by moving resource in to customer facing roles and deliver on enormously neglected areas of the service.
My question is that how would I build a system that generates these reports that isn't in a chat format, IE I get Claude to 'learn' various rules set out in policy documents, then generates text based on variable inputs that the operator gives it? I've seen some people tout these things on linkedin, how do you build these systems?
(I am not in charge of the department - the above is all a thought experiment). Very interested in feedback.
Pay attention to your own IP (intellectual property). Do your customers (and directors, and regulators) know you are freely sending it outside? As a corollary to that, are you making arrangements to self-host AI?
That is a good point - I can't say too much but for the purposes of what I have been feeding to AI, all the data is open source and publicly available, so not disclosing any protected IP. One hazard with these AI systems though is they are recording all the data you give to them forever and they can certainly also identify you.
And they are using your prompts to train on as well, very often.
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
The first one is a good point - I have been aware that in 'testing the limits' of Claude (they rapidly become apparent - despite our have for the technology) I am helping it develop.
Another Brexit dividend - my 11yo daughter just got a new passport delivered, and was so disgusted that it was a blue one that she threw it on the ground, where it landed on my toe!
An 11 year old cares about passport colours? What colour was she hoping for?
OLB has presumably been filling her head with utter nonsense the poor girl.
This might go down well with his brain dead base but just highlights to everyone else why he should be nowhere near the WH .
He is on bail, having been inducted for four different crimes. Incitement of violence against the president if a federal offence, and a blatant violation of his bail conditions. Any other individual would be in custody by now.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
Can I ask, with these people being turned away at the airport because of impending out of date passports - didn't they think to check the entry requirements for the country first?
When I went to Malta, I checked and they wanted a passport valid for six months. Mine had only three left, so I replaced it before checking in.
Not really hard, and would have saved any hassle at all.
I made this point on the last thread
Since 2003 when our son emigrated to New Zealand we have travelled extensively world wide and this rule applies so we renewed our passports accordingly
There is a tedious nature of the continual bewailing of our previous membership of the EU, not least because 'we have left and there is no prospect of us rejoining the single market or freedom of movement'
The quotes are by one Sir Keir Starmer just this week, and reaffirmed by the EU as well at the same time in connection with the review of the treaty next year when they reaffirned they will not reopen UKs membership
It is a forlorn hope but time to move on on the EU membership and to be fair it seems Starmer agrees
Nah. Rejoin is just off the agenda for the next 5 years, not forever.
It's impossible to resist the polling forever.
Rejoin: 48% (=) Stay out: 32% (=)
How long do you think polling on the death penalty can be resisted?
Pretty well forever. Until 60% of the electorate actually care about it enough to make a difference to the way they vote (they don't), nothing will happen.
Similarly with the EU - though the political salience of that is possibly rather more. It certainly was a decade ago, even if for now it's quiescent, But it's managed completely to wreck the Tory party in the meantime.
Good point by the previous poster. You can't say we should Rejoin with 48% but ignore 60% who want the death penalty restored. (I'm not in favour of it myself).
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
“I've just run some planning casework through Claude. It identified within a few seconds the grounds on which a legal challenge could be pursued. This is a skill that only the most experienced people working in the industry have and they will have built it up over careers of many years.”
Indeed. Claude is completely extraordinary. I’ve just been chatting with my lady novelist friend. I told you that she fed her latest novel into Claude and it gave her a detailed summary and analysis in forty five seconds. 😶
Now she has given Claude her editor’s feedback - several pages of deep advice - and Claude has decided the editor is right on this, this, this and not this. All in a few seconds. Stunningly good and incomparably fast
We’re chatting now. I had an idea for her novel and I messaged her a moment ago with the idea. Her reply: “Claude just suggested exactly that”
True story
What’s terrifying here is that if Claude can be this good at editing then I see no reason why Claude cannot actually write books on top of editing them. Given that it understands plot, tension, narrative, character, everything
This is Claude 3. Unless progress suddenly stops - why? - clause 6.5 will write brilliant novels. In seconds
@leon. Indeed - I was using GPT before but Claude is a game changer. After last nights output I have been hypothetically planning out all morning how we can completely redesign the service at work based on this technology. Productivity and accuracy could increase exponentially by effectively outsourcing about 80% of the work in report writing. The hard part of checking written output can also be done through AI. It creates vast opportunities to improve 'customer service' by moving resource in to customer facing roles and deliver on enormously neglected areas of the service.
My question is that how would I build a system that generates these reports that isn't in a chat format, IE I get Claude to 'learn' various rules set out in policy documents, then generates text based on variable inputs that the operator gives it? I've seen some people tout these things on linkedin, how do you build these systems?
(I am not in charge of the department - the above is all a thought experiment). Very interested in feedback.
Pay attention to your own IP (intellectual property). Do your customers (and directors, and regulators) know you are freely sending it outside? As a corollary to that, are you making arrangements to self-host AI?
That is a good point - I can't say too much but for the purposes of what I have been feeding to AI, all the data is open source and publicly available, so not disclosing any protected IP. One hazard with these AI systems though is they are recording all the data you give to them forever and they can certainly also identify you.
And they are using your prompts to train on as well, very often.
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
The first one is a good point - I have been aware that in 'testing the limits' of Claude (they rapidly become apparent - despite our have for the technology) I am helping it develop.
I too wonder if it uses the stuff we feed it, to develop. Probably it does
Likewise I have seen the limits. 200 questions and it flatlines for 8 hours. Weirdly basic errors, occasionally
However if you tell Claude he got such and such wrong, he swiftly learns and then gets it exactly right. Like a brilliant child
As I’ve said before, if you had two AIs - with different training sets - bouncing off each other, then that would be incredibly powerful, and would likely reduce errors to near zero
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
Another Brexit dividend - my 11yo daughter just got a new passport delivered, and was so disgusted that it was a blue one that she threw it on the ground, where it landed on my toe!
An 11 year old cares about passport colours? What colour was she hoping for?
She liked the colour of the old one.
And that's partly why the current settlement isn't settling. One of the factors in 2016 was the nostalgic one, for the good old days before Ted Heath. Doesn't matter what the balance of good and bad is, nostalgia just is. People who grew up with blue passports wanted them back, people who grew up with burgundy will feel the same. Why shouldn't they? It may be silly, but it's also human.
As time passes, the ratio of nostalgia for pre-EEC and nostalgia for in-EEC/EU will evolve in a predictable way.
Indeed, the blue/black one just looks weird to me. Burgundy is a much nicer colour, and it's just the colour that passports should be. I liked the way that the burgundy one used to let me travel and work freely in a single market of half a billion people too, but I guess that's just me.
The colour of our passports is way too important to be left to the whims of politicians. Only one viable solution, let Claude design it for us.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
“I've just run some planning casework through Claude. It identified within a few seconds the grounds on which a legal challenge could be pursued. This is a skill that only the most experienced people working in the industry have and they will have built it up over careers of many years.”
Indeed. Claude is completely extraordinary. I’ve just been chatting with my lady novelist friend. I told you that she fed her latest novel into Claude and it gave her a detailed summary and analysis in forty five seconds. 😶
Now she has given Claude her editor’s feedback - several pages of deep advice - and Claude has decided the editor is right on this, this, this and not this. All in a few seconds. Stunningly good and incomparably fast
We’re chatting now. I had an idea for her novel and I messaged her a moment ago with the idea. Her reply: “Claude just suggested exactly that”
True story
What’s terrifying here is that if Claude can be this good at editing then I see no reason why Claude cannot actually write books on top of editing them. Given that it understands plot, tension, narrative, character, everything
This is Claude 3. Unless progress suddenly stops - why? - clause 6.5 will write brilliant novels. In seconds
@leon. Indeed - I was using GPT before but Claude is a game changer. After last nights output I have been hypothetically planning out all morning how we can completely redesign the service at work based on this technology. Productivity and accuracy could increase exponentially by effectively outsourcing about 80% of the work in report writing. The hard part of checking written output can also be done through AI. It creates vast opportunities to improve 'customer service' by moving resource in to customer facing roles and deliver on enormously neglected areas of the service.
My question is that how would I build a system that generates these reports that isn't in a chat format, IE I get Claude to 'learn' various rules set out in policy documents, then generates text based on variable inputs that the operator gives it? I've seen some people tout these things on linkedin, how do you build these systems?
(I am not in charge of the department - the above is all a thought experiment). Very interested in feedback.
Pay attention to your own IP (intellectual property). Do your customers (and directors, and regulators) know you are freely sending it outside? As a corollary to that, are you making arrangements to self-host AI?
That is a good point - I can't say too much but for the purposes of what I have been feeding to AI, all the data is open source and publicly available, so not disclosing any protected IP. One hazard with these AI systems though is they are recording all the data you give to them forever and they can certainly also identify you.
And they are using your prompts to train on as well, very often.
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
The first one is a good point - I have been aware that in 'testing the limits' of Claude (they rapidly become apparent - despite our have for the technology) I am helping it develop.
I think my last point is the one that may turn out to be most important one - documents written by machine, will be read by machine. Why not do better?
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
We are already getting stupider. People are amazed when I can arrive at my destination without the aid of a satnav.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s actually quite an interesting parallel. All humanity will become The Welsh. Funny little hairy creatures clinging on to their strange language and seaweed eating
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s actually quite an interesting parallel. All humanity will become The Welsh. Funny little hairy creatures clinging on to their strange language and seaweed eating
If AI is comparable to Covid, @Eadric will have a grandstand view of this from Penarth.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
And that company will be way more productive and rake it in. Which means others will follow suit. And the whole industry will become more profitable. A lot of that will go to the owners, but much will go to the higher salary for the incredible productivity editor. They will spend it on more goods and services in other sectors. Which will expand the economy.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
And that company will be way more productive and rake it in. Which means others will follow suit. And the whole industry will become more profitable. A lot of that will go to the owners, but much will go to the higher salary for the incredible productivity editor. They will spend it on more goods and services in other sectors. Which will expand the economy.
And then AI will replace the sub editors and the designers and the advertisers and, in the end, the writers. All cognitive jobs will be robotised
Will the economy be “expanded” by that? Maybe, but it will also be tremendously challenging for humankind - and anyone who loses a job they actually rather liked. We will be without purpose
But, this is coming. Unless we revolt and pull the plug now
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
She was right to laugh at the thought of anything so destructive and ultimately pointless taking place.
Trident is reliable - The US has been firing them yearly since they came into service. With a very high success rate
Since we are literally using the same missile - it’s on the RN to explain the failures.
Your information is incorrect. The failure of the test under Cameron was a fault in the US guidance systems and Obama asked us to keep it secret because it was detrimental to the reputation of the US programme. We must assume that something similar happened under May when the missiles turned tail and decided to head back to the US.
Whether this points to some difference in systems, potentially allowing the US to alter the outcome of a British missile strike, is entering the realms of conspiracy theory. It is a theory that events aren't entirely inconsistent woth though.
If they had that capability why would they demonstrate it to the world?
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Ha!
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
And that company will be way more productive and rake it in. Which means others will follow suit. And the whole industry will become more profitable. A lot of that will go to the owners, but much will go to the higher salary for the incredible productivity editor. They will spend it on more goods and services in other sectors. Which will expand the economy.
While the other workers lose their jobs thus depressing their spending and shrinking the economy.
If Leon is correct then a period of high unemployment and high inequality is coming.
Similar but opposite to the 1980s - unemployment will be greater among the former middle classes and in southern England.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
We are already getting stupider. People are amazed when I can arrive at my destination without the aid of a satnav.
It's not just AI.
There's been, what, 70 years of deskilling of social reproduction outside of the workplace, and the pace is accelerating.
There's nothing sadder than seeing a long line of pillocks sat in their cars on a motorway, with little screens in front of them showing them a cartoon picture of a road, just so they can feel safe, copy everybody else, and not be viewed as a weirdo.
Mass chipping is around the corner. The population will lap it up. There you go, dears, you won't have to remember all your passwords any more. And you like eating, don't you?
But wait...some white collar jobs will disappear...well who would have thought it? Not as if there hasn't been huge and increasingly visible over-employment in the white collar sector for a couple of generations. Walk past any town hall. They're all in front of networked computer monitors nowadays, but their numbers didn't go down. The surprise is that the sector didn't shrink when it was computerised. It's been floating on a haze of mogadon for years.
Trident is reliable - The US has been firing them yearly since they came into service. With a very high success rate
Since we are literally using the same missile - it’s on the RN to explain the failures.
Your information is incorrect. The failure of the test under Cameron was a fault in the US guidance systems and Obama asked us to keep it secret because it was detrimental to the reputation of the US programme. We must assume that something similar happened under May when the missiles turned tail and decided to head back to the US.
Whether this points to some difference in systems, potentially allowing the US to alter the outcome of a British missile strike, is entering the realms of conspiracy theory. It is a theory that events aren't entirely inconsistent woth though.
If they had that capability why would they demonstrate it to the world?
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Overall makes zero sense
I don't think they would have been planning to demonstrate it to the world. If the capability of rerouting and disarming the warheads in the eventuality of a real strike exists, it could have been malfunctioning. Or the British could have changed the destination target of the test (which I am sure they share with the US) at the last minute without telling the US, so as to make it a 'real test', and it made the failsafe systems click in.
That said, I'm not aware that there are any other potential customers for Trident. The EU would be the only potential one, and the French wouldn't have it. They're not stupid. And UK Governments it seems will continue to be willing buyers regardless of how poor and ineffective the system appears to be.
I always put it this way - if you were the USA, would you give the UK the ability to blow up the world at will without any say? I wouldn't.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
She was right to laugh at the thought of anything so destructive and ultimately pointless taking place.
And it probably wouldn't have happened without pressure from the unions.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Ha!
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
I think I can extrapolate better than almost anyone. I’m in the top 0.001% of extrapolators
Why then can’t I monetise this? I should be able to. My problem is I am too excitable and I over react, and I tend to overlook obstacles that will slow down my predictions, or sometimes falsify them
Cf what3words
I need an AI assistant to say “slow down leon” and I need to double the time element in any prediction
So: I think ASI will arrive by 2030 (as does Elon). In reality that probably means 2035
Also my tendency to exaggerate probably means I am only in the top 0.01% of extrapolators
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
She was right to laugh at the thought of anything so destructive and ultimately pointless taking place.
And it probably wouldn't have happened without pressure from the unions.
There's no shortage of people happy to be paid to stay at home and to restrict the freedoms of others.
And finally whine and blame others when taxes and prices increase.
“I've just run some planning casework through Claude. It identified within a few seconds the grounds on which a legal challenge could be pursued. This is a skill that only the most experienced people working in the industry have and they will have built it up over careers of many years.”
Indeed. Claude is completely extraordinary. I’ve just been chatting with my lady novelist friend. I told you that she fed her latest novel into Claude and it gave her a detailed summary and analysis in forty five seconds. 😶
Now she has given Claude her editor’s feedback - several pages of deep advice - and Claude has decided the editor is right on this, this, this and not this. All in a few seconds. Stunningly good and incomparably fast
We’re chatting now. I had an idea for her novel and I messaged her a moment ago with the idea. Her reply: “Claude just suggested exactly that”
True story
What’s terrifying here is that if Claude can be this good at editing then I see no reason why Claude cannot actually write books on top of editing them. Given that it understands plot, tension, narrative, character, everything
This is Claude 3. Unless progress suddenly stops - why? - clause 6.5 will write brilliant novels. In seconds
@leon. Indeed - I was using GPT before but Claude is a game changer. After last nights output I have been hypothetically planning out all morning how we can completely redesign the service at work based on this technology. Productivity and accuracy could increase exponentially by effectively outsourcing about 80% of the work in report writing. The hard part of checking written output can also be done through AI. It creates vast opportunities to improve 'customer service' by moving resource in to customer facing roles and deliver on enormously neglected areas of the service.
My question is that how would I build a system that generates these reports that isn't in a chat format, IE I get Claude to 'learn' various rules set out in policy documents, then generates text based on variable inputs that the operator gives it? I've seen some people tout these things on linkedin, how do you build these systems?
(I am not in charge of the department - the above is all a thought experiment). Very interested in feedback.
Pay attention to your own IP (intellectual property). Do your customers (and directors, and regulators) know you are freely sending it outside? As a corollary to that, are you making arrangements to self-host AI?
That is a good point - I can't say too much but for the purposes of what I have been feeding to AI, all the data is open source and publicly available, so not disclosing any protected IP. One hazard with these AI systems though is they are recording all the data you give to them forever and they can certainly also identify you.
And they are using your prompts to train on as well, very often.
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
The first one is a good point - I have been aware that in 'testing the limits' of Claude (they rapidly become apparent - despite our have for the technology) I am helping it develop.
I think my last point is the one that may turn out to be most important one - documents written by machine, will be read by machine. Why not do better?
I think the situation in this respect is just an evolution of what already exists. Documents with loads of cut and paste text, totally irrellevant, and never read. Now it moves to machine generated text read by machines... the role of the human decision maker is pretty similar in both scenarios.
I don't think AI can take anything over at the moment, it can just assist with the grunt work. I've had exchanges with decent barristers that have the same view. Everyone wants to know what happens next and how much its capability will evolve... but no one knows the answer to that.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
The prediction needs some idea of timescale.
I just gave it
I think AGI will come in the next 2-3 years, so it’s probably in the next 4-6 years. I think ASI - the singularity - will arrive in 5 years but likely therefore it’s 10
However long before we reach the singularity we will see huge impacts on societies and economies
“I've just run some planning casework through Claude. It identified within a few seconds the grounds on which a legal challenge could be pursued. This is a skill that only the most experienced people working in the industry have and they will have built it up over careers of many years.”
Indeed. Claude is completely extraordinary. I’ve just been chatting with my lady novelist friend. I told you that she fed her latest novel into Claude and it gave her a detailed summary and analysis in forty five seconds. 😶
Now she has given Claude her editor’s feedback - several pages of deep advice - and Claude has decided the editor is right on this, this, this and not this. All in a few seconds. Stunningly good and incomparably fast
We’re chatting now. I had an idea for her novel and I messaged her a moment ago with the idea. Her reply: “Claude just suggested exactly that”
True story
What’s terrifying here is that if Claude can be this good at editing then I see no reason why Claude cannot actually write books on top of editing them. Given that it understands plot, tension, narrative, character, everything
This is Claude 3. Unless progress suddenly stops - why? - clause 6.5 will write brilliant novels. In seconds
@leon. Indeed - I was using GPT before but Claude is a game changer. After last nights output I have been hypothetically planning out all morning how we can completely redesign the service at work based on this technology. Productivity and accuracy could increase exponentially by effectively outsourcing about 80% of the work in report writing. The hard part of checking written output can also be done through AI. It creates vast opportunities to improve 'customer service' by moving resource in to customer facing roles and deliver on enormously neglected areas of the service.
My question is that how would I build a system that generates these reports that isn't in a chat format, IE I get Claude to 'learn' various rules set out in policy documents, then generates text based on variable inputs that the operator gives it? I've seen some people tout these things on linkedin, how do you build these systems?
(I am not in charge of the department - the above is all a thought experiment). Very interested in feedback.
Pay attention to your own IP (intellectual property). Do your customers (and directors, and regulators) know you are freely sending it outside? As a corollary to that, are you making arrangements to self-host AI?
That is a good point - I can't say too much but for the purposes of what I have been feeding to AI, all the data is open source and publicly available, so not disclosing any protected IP. One hazard with these AI systems though is they are recording all the data you give to them forever and they can certainly also identify you.
And they are using your prompts to train on as well, very often.
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
The first one is a good point - I have been aware that in 'testing the limits' of Claude (they rapidly become apparent - despite our have for the technology) I am helping it develop.
I think my last point is the one that may turn out to be most important one - documents written by machine, will be read by machine. Why not do better?
I think the situation in this respect is just an evolution of what already exists. Documents with loads of cut and paste text, totally irrellevant, and never read. Now it moves to machine generated text read by machines... the role of the human decision maker is pretty similar in both scenarios.
I don't think AI can take anything over at the moment, it can just assist with the grunt work. I've had exchanges with decent barristers that have the same view. Everyone wants to know what happens next and how much its capability will evolve... but no one knows the answer to that.
The evolution towards a specific “expert system” on a particular matter, that you set up and the consumer asks questions of seems obvious. In effect this is what is happening with using “AI” to summarise unreadable documents.
Worth noting that Claude still has serious limitations. It can make basic errors of comprehension, and it has to reread an entire conversation - and, in this case, a novel - every time it answers
But even that is quite amazing. It reads a whole novel every time you ask, and it does that in about 1 minute. PB-ers need to try this machine. You need to try the pro opus version
Yes: that's the whole point of the context window.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
An absolute inability to extrapolate. That’s what you have. Like many
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Ha!
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
I think I can extrapolate better than almost anyone. I’m in the top 0.001% of extrapolators
Why then can’t I monetise this? I should be able to. My problem is I am too excitable and I over react, and I tend to overlook obstacles that will slow down my predictions, or sometimes falsify them
Cf what3words
I need an AI assistant to say “slow down leon” and I need to double the time element in any prediction
So: I think ASI will arrive by 2030 (as does Elon). In reality that probably means 2035
Also my tendency to exaggerate probably means I am only in the top 0.01% of extrapolators
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Ha!
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
I think I can extrapolate better than almost anyone. I’m in the top 0.001% of extrapolators
Why then can’t I monetise this? I should be able to. My problem is I am too excitable and I over react, and I tend to overlook obstacles that will slow down my predictions, or sometimes falsify them
Cf what3words
I need an AI assistant to say “slow down leon” and I need to double the time element in any prediction
So: I think ASI will arrive by 2030 (as does Elon). In reality that probably means 2035
Also my tendency to exaggerate probably means I am only in the top 0.01% of extrapolators
Retrospective criticism of lockdowns is a classic case of hindsight, and it's mostly coming from people who will be avid supporters of much worse policies in the future when they're imposed in the name of public safety, public health, national security, or some other two-word phrase.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Ha!
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
I think I can extrapolate better than almost anyone. I’m in the top 0.001% of extrapolators
Why then can’t I monetise this? I should be able to. My problem is I am too excitable and I over react, and I tend to overlook obstacles that will slow down my predictions, or sometimes falsify them
Cf what3words
I need an AI assistant to say “slow down leon” and I need to double the time element in any prediction
So: I think ASI will arrive by 2030 (as does Elon). In reality that probably means 2035
Also my tendency to exaggerate probably means I am only in the top 0.01% of extrapolators
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
Ha!
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
I think I can extrapolate better than almost anyone. I’m in the top 0.001% of extrapolators
Why then can’t I monetise this? I should be able to. My problem is I am too excitable and I over react, and I tend to overlook obstacles that will slow down my predictions, or sometimes falsify them
Cf what3words
I need an AI assistant to say “slow down leon” and I need to double the time element in any prediction
So: I think ASI will arrive by 2030 (as does Elon). In reality that probably means 2035
Also my tendency to exaggerate probably means I am only in the top 0.01% of extrapolators
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
An absolute inability to extrapolate. That’s what you have. Like many
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
I just asked Claude to edit a few pages of prose and got the following message:
'Sure, I can help you edit the text. Please provide the specific changes or edits you would like me to make.'
Trident is reliable - The US has been firing them yearly since they came into service. With a very high success rate
Since we are literally using the same missile - it’s on the RN to explain the failures.
Your information is incorrect. The failure of the test under Cameron was a fault in the US guidance systems and Obama asked us to keep it secret because it was detrimental to the reputation of the US programme. We must assume that something similar happened under May when the missiles turned tail and decided to head back to the US.
Whether this points to some difference in systems, potentially allowing the US to alter the outcome of a British missile strike, is entering the realms of conspiracy theory. It is a theory that events aren't entirely inconsistent woth though.
If they had that capability why would they demonstrate it to the world?
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Overall makes zero sense
I don't think they would have been planning to demonstrate it to the world. If the capability of rerouting and disarming the warheads in the eventuality of a real strike exists, it could have been malfunctioning. Or the British could have changed the destination target of the test (which I am sure they share with the US) at the last minute without telling the US, so as to make it a 'real test', and it made the failsafe systems click in.
That said, I'm not aware that there are any other potential customers for Trident. The EU would be the only potential one, and the French wouldn't have it. They're not stupid. And UK Governments it seems will continue to be willing buyers regardless of how poor and ineffective the system appears to be.
I always put it this way - if you were the USA, would you give the UK the ability to blow up the world at will without any say? I wouldn't.
Excepting that they have, on multiple occasions.
Hell, in Germany, in the late 50s/early 60s, you had American nuclear bombs hanging off German fighter bombers. With the pilot sometimes having a tin tie from the late unpleasantness. And no lock of any kind on the bomb.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
Can I ask, with these people being turned away at the airport because of impending out of date passports - didn't they think to check the entry requirements for the country first?
When I went to Malta, I checked and they wanted a passport valid for six months. Mine had only three left, so I replaced it before checking in.
Not really hard, and would have saved any hassle at all.
I made this point on the last thread
Since 2003 when our son emigrated to New Zealand we have travelled extensively world wide and this rule applies so we renewed our passports accordingly
There is a tedious nature of the continual bewailing of our previous membership of the EU, not least because 'we have left and there is no prospect of us rejoining the single market or freedom of movement'
The quotes are by one Sir Keir Starmer just this week, and reaffirmed by the EU as well at the same time in connection with the review of the treaty next year when they reaffirned they will not reopen UKs membership
It is a forlorn hope but time to move on on the EU membership and to be fair it seems Starmer agrees
Nah. Rejoin is just off the agenda for the next 5 years, not forever.
It's impossible to resist the polling forever.
Rejoin: 48% (=) Stay out: 32% (=)
How long do you think polling on the death penalty can be resisted?
Pretty well forever. Until 60% of the electorate actually care about it enough to make a difference to the way they vote (they don't), nothing will happen.
Similarly with the EU - though the political salience of that is possibly rather more. It certainly was a decade ago, even if for now it's quiescent, But it's managed completely to wreck the Tory party in the meantime.
I think putting a Referendum on the death penalty in the Tory Manifesto is pretty much their best weapon in GE2024
That just shows what a hopeless state they're in. Still, it would at least be a clear test of how important the issue is to electors.
"Vote for us: we're shit, but we'll kill a handful of people" isn't a particularly convincing platform.
Well yes but I guess from their perspective desperate times desperate policies.
What's the betting SKS would say we will offer one too. Would definitely be good news for LDs/ Greens if so. #Hangmanslackenyournoose
Labour will kill more, sooner. Only Labour can be trusted with State human disposal
Progressive politics is all about having unborn babies put to death, not murderers.
Retrospective criticism of lockdowns is a classic case of hindsight, and it's mostly coming from people who will be avid supporters of much worse policies in the future when they're imposed in the name of public safety, public health, national security, or some other two-word phrase.
I don't think you're right Donkeys. I think those who favour restrictions tend to generally favour them, and those who oppose them tend to generally oppose them.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
An absolute inability to extrapolate. That’s what you have. Like many
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
I just asked Claude to edit a few pages of prose and got the following message:
'Sure, I can help you edit the text. Please provide the specific changes or edits you would like me to make.'
People who are too stupid to use AI giving their negative opinions of AI is one of my favourite genres
I had people saying this shit to me when I first used Stable Diffusion. “Oh it only gives me terrrible images”
THATS BECAUSE YOU ARE AN IDIOT
Now Midjourney produces flawless images - even for idiots - and everyone just shrugs and accepts it
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
I think you mean
that. liz. truss
I don't no where that is but I know it is a. desolate. wilderness
Here’s an actual example of Claude at work on a book. Just sent me by my lady writer friend. This is easily as good and useful as having a professional editor sitting besides you as you work
This is Claude discussing a proposed new plot twist in the novel
My friend says Claude is actually having excellent ideas. It’s not just editing and helping - it is creating and imagining
Now apply that - plus the skill ease and cheapness - to any cognitive task
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
You weren't here in 2020. There was a poster called Eadric who had a similar writing style to you but I don't believe he mentioned the corona virus until people started dropping like flies in China and Italy, at which point he f***** off to Wales and proposed a lab leak theory. Eadric was no Nostradamus.
In late January 2020 I told the mother of my older daughter that “they will close the schools”
She literally laughed in my face. Three months later she apologised and said “you were right”
She was right to laugh at the thought of anything so destructive and ultimately pointless taking place.
And it probably wouldn't have happened without pressure from the unions.
There's no shortage of people happy to be paid to stay at home and to restrict the freedoms of others.
And finally whine and blame others when taxes and prices increase.
This is all getting too surreal.
The unions locked us down? If only we'd had a great, inch-perfect, Churchillian War Prime Minister to lead us through the wilderness of COVID.
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
An absolute inability to extrapolate. That’s what you have. Like many
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
I just asked Claude to edit a few pages of prose and got the following message:
'Sure, I can help you edit the text. Please provide the specific changes or edits you would like me to make.'
People who are too stupid to use AI giving their negative opinions of AI is one of my favourite genres
I had people saying this shit to me when I first used Stable Diffusion. “Oh it only gives me terrrible images”
THATS BECAUSE YOU ARE AN IDIOT
Now Midjourney produces flawless images - even for idiots - and everyone just shrugs and accepts it
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
I’m so old I can remember when you said all AI art is “shit” and like “crappy album covers” and it would always be shit. It was about a year ago
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
An absolute inability to extrapolate. That’s what you have. Like many
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
I just asked Claude to edit a few pages of prose and got the following message:
'Sure, I can help you edit the text. Please provide the specific changes or edits you would like me to make.'
People who are too stupid to use AI giving their negative opinions of AI is one of my favourite genres
I had people saying this shit to me when I first used Stable Diffusion. “Oh it only gives me terrrible images”
THATS BECAUSE YOU ARE AN IDIOT
Now Midjourney produces flawless images - even for idiots - and everyone just shrugs and accepts it
Okay, I'll persevere.
You don’t just feed something in and say “edit that”
You have to work with it. Instruct it. Learn how it responds. Think of it as a dog you have to train, and command, but with the brain of Einstein
Claude also stores your conversations. So you can leave it and log out for an hour or a week. Then you return to the dialogue and it re-reads it all in one minute and resumes the brainstorming instantly, as smart and alert as ever, whenever you want
No editor can do that. They get tired, they menstruate, they sleep, they get ill or bored or move jobs
They’d better move jobs because editing is over as a job
I would imagine they are not just 'editing' the work though - there is also an element of guiding it towards commercial success, which is presumably the business model of the publisher.
Yes, but Claude has an “intuitive” sense of what is commercial. All writing is, in the end, an algorithm
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
You may be proved right in the end, but I think you are probably making an assumption here that Claude is preferable to an editor which seems to be based to a large extent on speed of response. It could also be manipulating you re the merits of what it is telling you to do - one feature of all the AI systems I have used is that their responses are superficially extremely convincing but can rapidly then fall apart when they are properly scutinised. One of my biggest fears about AI is that it will make us all more stupid.
Remember this tech is only going to get better
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
The only way it will be better than a human editor is that it will be faster. We are not about to enter a golden age of publishing.
An absolute inability to extrapolate. That’s what you have. Like many
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
I just asked Claude to edit a few pages of prose and got the following message:
'Sure, I can help you edit the text. Please provide the specific changes or edits you would like me to make.'
People who are too stupid to use AI giving their negative opinions of AI is one of my favourite genres
I had people saying this shit to me when I first used Stable Diffusion. “Oh it only gives me terrrible images”
THATS BECAUSE YOU ARE AN IDIOT
Now Midjourney produces flawless images - even for idiots - and everyone just shrugs and accepts it
Can you get it to give odds and reasoning behind them for the next GE and next Prime Minister market?
Trident is reliable - The US has been firing them yearly since they came into service. With a very high success rate
Since we are literally using the same missile - it’s on the RN to explain the failures.
Your information is incorrect. The failure of the test under Cameron was a fault in the US guidance systems and Obama asked us to keep it secret because it was detrimental to the reputation of the US programme. We must assume that something similar happened under May when the missiles turned tail and decided to head back to the US.
Whether this points to some difference in systems, potentially allowing the US to alter the outcome of a British missile strike, is entering the realms of conspiracy theory. It is a theory that events aren't entirely inconsistent woth though.
If they had that capability why would they demonstrate it to the world?
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Overall makes zero sense
I don't think they would have been planning to demonstrate it to the world. If the capability of rerouting and disarming the warheads in the eventuality of a real strike exists, it could have been malfunctioning. Or the British could have changed the destination target of the test (which I am sure they share with the US) at the last minute without telling the US, so as to make it a 'real test', and it made the failsafe systems click in.
That said, I'm not aware that there are any other potential customers for Trident. The EU would be the only potential one, and the French wouldn't have it. They're not stupid. And UK Governments it seems will continue to be willing buyers regardless of how poor and ineffective the system appears to be.
I always put it this way - if you were the USA, would you give the UK the ability to blow up the world at will without any say? I wouldn't.
Excepting that they have, on multiple occasions.
Hell, in Germany, in the late 50s/early 60s, you had American nuclear bombs hanging off German fighter bombers. With the pilot sometimes having a tin tie from the late unpleasantness. And no lock of any kind on the bomb.
Your knowledge of military history is far superior to mine. But going off what I see now, I do not see today's USA being happy to provide a fully independent nuclear deterrent (that could be used against the USA itself, quite apart from someone the USA didn't want to be fried) without profound ways of influencing the course of those events.
If I were today's UK politicians, I would move away from Trident, ensuring that I was spending enough on other US weapons to minimise the fury. I would consider gifting Trident to the EU, but I don't think even they would be silly enough to take it, grandiose military ambitions or no.
Another Brexit dividend - my 11yo daughter just got a new passport delivered, and was so disgusted that it was a blue one that she threw it on the ground, where it landed on my toe!
An 11 year old cares about passport colours? What colour was she hoping for?
She liked the colour of the old one.
And you’ve trained her to respond to change by having a tantrum?
Trident is reliable - The US has been firing them yearly since they came into service. With a very high success rate
Since we are literally using the same missile - it’s on the RN to explain the failures.
Your information is incorrect. The failure of the test under Cameron was a fault in the US guidance systems and Obama asked us to keep it secret because it was detrimental to the reputation of the US programme. We must assume that something similar happened under May when the missiles turned tail and decided to head back to the US.
Whether this points to some difference in systems, potentially allowing the US to alter the outcome of a British missile strike, is entering the realms of conspiracy theory. It is a theory that events aren't entirely inconsistent woth though.
If they had that capability why would they demonstrate it to the world?
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Overall makes zero sense
I don't think they would have been planning to demonstrate it to the world. If the capability of rerouting and disarming the warheads in the eventuality of a real strike exists, it could have been malfunctioning. Or the British could have changed the destination target of the test (which I am sure they share with the US) at the last minute without telling the US, so as to make it a 'real test', and it made the failsafe systems click in.
That said, I'm not aware that there are any other potential customers for Trident. The EU would be the only potential one, and the French wouldn't have it. They're not stupid. And UK Governments it seems will continue to be willing buyers regardless of how poor and ineffective the system appears to be.
I always put it this way - if you were the USA, would you give the UK the ability to blow up the world at will without any say? I wouldn't.
Excepting that they have, on multiple occasions.
Hell, in Germany, in the late 50s/early 60s, you had American nuclear bombs hanging off German fighter bombers. With the pilot sometimes having a tin tie from the late unpleasantness. And no lock of any kind on the bomb.
Your knowledge of military history is far superior to mine. But going off what I see now, I do not see today's USA being happy to provide a fully independent nuclear deterrent (that could be used against the USA itself, quite apart from someone the USA didn't want to be fried) without profound ways of influencing the course of those events.
If I were today's UK politicians, I would move away from Trident, ensuring that I was spending enough on other US weapons to minimise the fury. I would consider gifting Trident to the EU, but I don't think even they would be silly enough to take it, grandiose military ambitions or no.
Why would the EU want Trident, when France had its own independent nuclear deterrent?
Trident is reliable - The US has been firing them yearly since they came into service. With a very high success rate
Since we are literally using the same missile - it’s on the RN to explain the failures.
Your information is incorrect. The failure of the test under Cameron was a fault in the US guidance systems and Obama asked us to keep it secret because it was detrimental to the reputation of the US programme. We must assume that something similar happened under May when the missiles turned tail and decided to head back to the US.
Whether this points to some difference in systems, potentially allowing the US to alter the outcome of a British missile strike, is entering the realms of conspiracy theory. It is a theory that events aren't entirely inconsistent woth though.
If they had that capability why would they demonstrate it to the world?
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Overall makes zero sense
I don't think they would have been planning to demonstrate it to the world. If the capability of rerouting and disarming the warheads in the eventuality of a real strike exists, it could have been malfunctioning. Or the British could have changed the destination target of the test (which I am sure they share with the US) at the last minute without telling the US, so as to make it a 'real test', and it made the failsafe systems click in.
That said, I'm not aware that there are any other potential customers for Trident. The EU would be the only potential one, and the French wouldn't have it. They're not stupid. And UK Governments it seems will continue to be willing buyers regardless of how poor and ineffective the system appears to be.
I always put it this way - if you were the USA, would you give the UK the ability to blow up the world at will without any say? I wouldn't.
Excepting that they have, on multiple occasions.
Hell, in Germany, in the late 50s/early 60s, you had American nuclear bombs hanging off German fighter bombers. With the pilot sometimes having a tin tie from the late unpleasantness. And no lock of any kind on the bomb.
Your knowledge of military history is far superior to mine. But going off what I see now, I do not see today's USA being happy to provide a fully independent nuclear deterrent (that could be used against the USA itself, quite apart from someone the USA didn't want to be fried) without profound ways of influencing the course of those events.
If I were today's UK politicians, I would move away from Trident, ensuring that I was spending enough on other US weapons to minimise the fury. I would consider gifting Trident to the EU, but I don't think even they would be silly enough to take it, grandiose military ambitions or no.
Why would the EU want Trident, when France had its own independent nuclear deterrent?
So they could paint an EU flag on it? I agree they wouldn't want it - I do think they'd have liked one of our aircraft carriers though.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
Interesting and surprisingly sympathetic dissection in today's Guardian of the total failure that is Liz Truss
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
I am sure Liz Truss's version is the bondage version - I think the chain attaches to the bale differently, but even that construction doesn't seem to be wholly uncommon.
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
Interesting and surprisingly sympathetic dissection in today's Guardian of the total failure that is Liz Truss
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
Interesting and surprisingly sympathetic dissection in today's Guardian of the total failure that is Liz Truss
Half-hearted cattiness (about the fat folk in the Popcon audience etc.) aisde, this article is quite good - raises some interesting points and identifies some genuine flaws in Truss's argument. Why did she not realise she would have to work within the confines of the blob for a while before starting to change it? One could almost argue that she intended to fail, and make a lucrative living on the right wing shock jock circuit (which she seems to have done very successfully).
However, I still think that really, she believed that her Prime Ministerial mandate would see her through, and presumably that the Tory Party would let her ride it out until her policies started to bear economical fruit by the election.
There's a story on the BBC website about the use of AI to show people with potentially terminal illnesses their 'lost future' by generating images of future events:
Another Brexit dividend - my 11yo daughter just got a new passport delivered, and was so disgusted that it was a blue one that she threw it on the ground, where it landed on my toe!
An 11 year old cares about passport colours? What colour was she hoping for?
She liked the colour of the old one.
And you’ve trained her to respond to change by having a tantrum?
I've not trained her to do anything, she's her own person and prone to theatrical gestures. She's a quite brilliant and talented child and destined for great things, I suspect. Thanks for your concern.
Legendary labor leader Cesar Chavez’s family is outraged that RFK Jr is exploiting their dad’s legacy by using an iconic photo of him to promote a ‘Cesar Chavez Rally with RFK Jr’ in LA tomorrow. They are all supporting Biden. https://twitter.com/RonFilipkowski/status/1773806050786849105
I think a June/July election has got to be value at this point.
Why? The Conservatives will get tonked in the locals, then what? Tories pressing Rishi to resign will not be doing so because they want to see Starmer in Number 10. They will be hoping for someone to turn things round, which surely gets us through the summer.
THREAD: What is happening in UK universities is astonishing. Right now about quarter of universities are laying off staff. Almost all universities expect to make a loss this year. And the entire thing is happening in an effort to deceive the population about migration statistics. https://twitter.com/Prolapsarian/status/1774030136737730661
This makes about as much sense was banning foreign tourism to bring down the immigration figures. It’s just self destructive.
Legendary labor leader Cesar Chavez’s family is outraged that RFK Jr is exploiting their dad’s legacy by using an iconic photo of him to promote a ‘Cesar Chavez Rally with RFK Jr’ in LA tomorrow. They are all supporting Biden. https://twitter.com/RonFilipkowski/status/1773806050786849105
I think a June/July election has got to be value at this point.
Why? The Conservatives will get tonked in the locals, then what? Tories pressing Rishi to resign will not be doing so because they want to see Starmer in Number 10. They will be hoping for someone to turn things round, which surely gets us through the summer.
I guess the thinking is May humping, vote of no confidence, Sunak survives but realises he can't hang on till November and keep the party in one fractured piece so goes straight Into an election. Especially if he's had enough and fancies buggering off to the States shortly into the new parliament
Edit - and it will be first week of July so he can say 'see? Second half of the year'
THREAD: What is happening in UK universities is astonishing. Right now about quarter of universities are laying off staff. Almost all universities expect to make a loss this year. And the entire thing is happening in an effort to deceive the population about migration statistics. https://twitter.com/Prolapsarian/status/1774030136737730661
This makes about as much sense was banning foreign tourism to bring down the immigration figures. It’s just self destructive.
Is it not the case that there are currently about 50% more foreign students at UK universities than there were pre-Brexit?
If they are in financial trouble it suggests there has been some ponzi-style accounting.
I can share the experiences of others confirming it's a very pleasant afternoon with plenty of people out and about.
That said, glad to be back home and ready to enter the "fray" on here.
As I don't believe in bringing a knife to a gun fight, I don't comment on AI though the domination of topic after topic by one individual on one subject is in itself a sign.
The problem is there's very little happening in politics currently. The polls aren't moving yet and the Conservatives seem to have mentally given up - "sleepwalking to disaster" as someone once said. A significant Labour majority seems probablr though the implications of that for the medium term political outlook seem uncertain.
As in early 1997, there's a view among some supporters of the Government the incoming Labour administration will swiftly beome unpopular and the country will suddenly clamour for the return of the Tories. That seems improbable given Labour can heap blams on the outgoing Tories for most of what is happening and some (perhaps not many) of those Conservatives responsble will be on the opposition benches.
One of those things that is by its very nature a bit of a tricky one. Jeffrey Donaldson's charges.
1. Whilst its talked about as being of a historical nature, its not quite as far back as that term suggests 2. The fact that there are two women coming forward is a pointer 3. Despite how some police forces seem to work these days, tipping off the media before anyone is even charged, this didnt leak at all. I understand that any idea that other senior UK politicians might have been aware is incorrect. 4. Despite the usual online speculation that this is a dark forces conspiracy/stitch up/malevolent elites secnario, there is no evidence of that 5. Questioning to charge was extremely quick 5. Whether guilty or not, the nature of the allegations when they come out means there is not a hope in hell of him coming back. The only chance of a comeback of any sort at all is the case is dropped before it reaches court and that seems very unlikely.
THREAD: What is happening in UK universities is astonishing. Right now about quarter of universities are laying off staff. Almost all universities expect to make a loss this year. And the entire thing is happening in an effort to deceive the population about migration statistics. https://twitter.com/Prolapsarian/status/1774030136737730661
This makes about as much sense was banning foreign tourism to bring down the immigration figures. It’s just self destructive.
Is it not the case that there are currently about 50% more foreign students at UK universities than there were pre-Brexit?
If they are in financial trouble it suggests there has been some ponzi-style accounting.
Or that they were told by the government that the plan was to expand the number of foreign students by 50% and budgeted on that basis?
Everyone on PB should be using Claude 3 opus. Whatever your job is, this tech is going to transform it
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
Indeed and absolutely right
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
That’s pretty interesting. I know people mock you but your comments and articles do provoke thought. Far more than many others. I find peoples blasé attitude towards this surprising. The risks are plain to see.
Elon Musk reckons AGI by 2030.
I was mocked when I said “covid is coming” in late January 2020
Now I am saying: AGI is coming
For balance you were also mocked for promoting Putin as defender of Judaeo-Christian values and suggesting Truss might surprise on the upside.
That was the Liz truss in the necklace, that woman? The necklace I spotted within 2 minutes of watching the debate, when the rest of you were utterly clueless?
That Liz truss?
The actuality of PM Truss was quite a bit after your necklace burbling. I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
Interesting and surprisingly sympathetic dissection in today's Guardian of the total failure that is Liz Truss
If that article is sympathetic I'd like to read an unsympathetic one. But interesting to learn that the Trumpites think she's a total fraud. And is she really at risk in South West Norfolk?
Comments
And it will probably take your job away in 5-10 years, possibly sooner than that
Those that benefit from it will be those that master it early, and get ahead of the curve, and use it while it is still best as an assistant, rather than doing everything itself (which it will do)
https://claude.ai/login
All these suckers taking degrees, taking in debt they will never pay off, hoping for a career when AI will decimate many of their prospective jobs and replace them who god knows what.
Far better to get an apprenticeship and a trade.
One arm, straight out horizontally* while shouting "My Lords!" and "Dreadful!".
*You need to be careful with this bit, as getting it wrong makes people think that you are a Nazi.
Incitement of violence against the president if a federal offence, and a blatant violation of his bail conditions.
Any other individual would be in custody by now.
Claude is good at algorithms. Formulae. That’s what Claude does best, right now - and it can do the job of 10 professional editors - much faster - for £20 a month. You don’t need to be a computer to crunch the economic maths there
Properly segregated systems for commercial use are surprisingly lagging. And there is a serious trust issue.
The other joker is - who will be reading the documents at the other end? Yes, people are working on….
Which comes to an interesting idea - a shared info base, rather than documents. So you *ask* a system, as the consumer, for specific information. So instead of a telephone directory to be inhaled, you ask a specific question from a knowledge base. And get a few sentences on that specific request.
1) anti-abortion activists who are anti death penalty.
2) pro death penalty people who are pro abortion
Head full of utter nonsense, you say ?
I fear for my kids. I’m not completely delighted by this tech, half of me is terrified, but I’m not going to live in denial. It’s here and it’s real and it’s not going away
On the upside, we possibly live in the most interesting moment in human history. The creation of a superior and alien intelligence
I am reminded of this prescient article in the Spectator, 16 months ago, comparing the arrival of AI to the arrival of Europeans in Australia
“A few decades after the first contact art appeared in Arnhemland, the Aboriginal people were largely swept from their homelands by the economic forces and political changes brought by European settlers. But not everyone deserted these silent red canyons and bright lilied waterways. In the 2000s, the last native speaker of the local Amurdag language, Charlie Mungulda, returned to the Mount Borradaile rockfaces to paint a symbolic and final hand, a scarlet print which says ‘remember us, because we were here’. With this ritual, he signalled an end to 50,000 continuous years of Aboriginal art in West Arnhemland. Nothing has been painted in Mount Borrodaile since. Nor will it be.”
https://www.spectator.co.uk/article/how-australian-rock-art-warns-us-about-2023/
Likewise I have seen the limits. 200 questions and it flatlines for 8 hours. Weirdly basic errors, occasionally
However if you tell Claude he got such and such wrong, he swiftly learns and then gets it exactly right. Like a brilliant child
As I’ve said before, if you had two AIs - with different training sets - bouncing off each other, then that would be incredibly powerful, and would likely reduce errors to near zero
As for the editing, a good human editor is PROBABLY better than Claude right now, albeit much much slower and a thousand times more expensive
For me the economics points one way and one way only. The first publisher to sack 96% of its editors and replace them with AIs guided by just one editor, will save squillions of pounds that it can then spend on advertising, and better AI, thereby defeating its competitors
It’s inevitable and inexorable
Elon Musk reckons AGI by 2030.
Now I am saying: AGI is coming
Will the economy be “expanded” by that? Maybe, but it will also be tremendously challenging for humankind - and anyone who loses a job they actually rather liked. We will be without purpose
But, this is coming. Unless we revolt and pull the plug now
She literally laughed in my face. Three months later she apologised and said “you were right”
Lab 40%
Con 21%
RefUK 16%
LD 10%
Green 8%
https://en.wikipedia.org/wiki/Opinion_polling_for_the_next_United_Kingdom_general_election#National_poll_results
Undermines the attractiveness of US weapons to other buyers
Undermines the US’s most consistent ally
Provides the US’s adversaries with useful intelligence
Overall makes zero sense
So you are not Eadric who didn't have a Scooby Doo. My apologies for suggesting you were.
If Leon is correct then a period of high unemployment and high inequality is coming.
Similar but opposite to the 1980s - unemployment will be greater among the former middle classes and in southern England.
There's been, what, 70 years of deskilling of social reproduction outside of the workplace, and the pace is accelerating.
There's nothing sadder than seeing a long line of pillocks sat in their cars on a motorway, with little screens in front of them showing them a cartoon picture of a road, just so they can feel safe, copy everybody else, and not be viewed as a weirdo.
Where we're headed reminds me of this scene from Monty Python and the Holy Grail.
Mass chipping is around the corner. The population will lap it up. There you go, dears, you won't have to remember all your passwords any more. And you like eating, don't you?
But wait...some white collar jobs will disappear...well who would have thought it? Not as if there hasn't been huge and increasingly visible over-employment in the white collar sector for a couple of generations. Walk past any town hall. They're all in front of networked computer monitors nowadays, but their numbers didn't go down. The surprise is that the sector didn't shrink when it was computerised. It's been floating on a haze of mogadon for years.
That said, I'm not aware that there are any other potential customers for Trident. The EU would be the only potential one, and the French wouldn't have it. They're not stupid. And UK Governments it seems will continue to be willing buyers regardless of how poor and ineffective the system appears to be.
I always put it this way - if you were the USA, would you give the UK the ability to blow up the world at will without any say? I wouldn't.
Why then can’t I monetise this? I should be able to. My problem is I am too excitable and I over react, and I tend to overlook obstacles that will slow down my predictions, or sometimes falsify them
Cf what3words
I need an AI assistant to say “slow down leon” and I need to double the time element in any prediction
So: I think ASI will arrive by 2030 (as does Elon). In reality that probably means 2035
Also my tendency to exaggerate probably means I am only in the top 0.01% of extrapolators
We are not about to enter a golden age of publishing.
And finally whine and blame others when taxes and prices increase.
I don't think AI can take anything over at the moment, it can just assist with the grunt work. I've had exchanges with decent barristers that have the same view. Everyone wants to know what happens next and how much its capability will evolve... but no one knows the answer to that.
I think AGI will come in the next 2-3 years, so it’s probably in the next 4-6 years. I think ASI - the singularity - will arrive in 5 years but likely therefore it’s 10
However long before we reach the singularity we will see huge impacts on societies and economies
A faster editor IS a much better editor. Right now my friend has to wait weeks for her editor to get round to her novel and do the job. Pointless weeks. The AI editor is there and waiting and excellent, all the time. The AI editor will work at 3am if needs be. It will brainstorm in real time. It is inexhaustible
Now extrapolate from that. And remember the AI will only improve and remember that you could have two different AIs working on your book, honing it to perfection
lock.down.parties
'Sure, I can help you edit the text. Please provide the specific changes or edits you would like me to make.'
Hell, in Germany, in the late 50s/early 60s, you had American nuclear bombs hanging off German fighter bombers. With the pilot sometimes having a tin tie from the late unpleasantness. And no lock of any kind on the bomb.
That Liz truss?
https://yougov.co.uk/topics/politics/trackers/should-women-have-the-right-to-an-abortion
I had people saying this shit to me when I first used Stable Diffusion. “Oh it only gives me terrrible images”
THATS BECAUSE YOU ARE AN IDIOT
Now Midjourney produces flawless images - even for idiots - and everyone just shrugs and accepts it
I think it’s fair to say she pretty much surprised on the downside, even for those of us with low expectations.
that. liz. truss
I don't no where that is but I know it is
a. desolate. wilderness
This is Claude discussing a proposed new plot twist in the novel
My friend says Claude is actually having excellent ideas. It’s not just editing and helping - it is creating and imagining
Now apply that - plus the skill ease and cheapness - to any cognitive task
The unions locked us down? If only we'd had a great, inch-perfect, Churchillian War Prime Minister to lead us through the wilderness of COVID.
You don’t say that so much, anymore
You have to work with it. Instruct it. Learn how it responds. Think of it as a dog you have to train, and command, but with the brain of Einstein
If I were today's UK politicians, I would move away from Trident, ensuring that I was spending enough on other US weapons to minimise the fury. I would consider gifting Trident to the EU, but I don't think even they would be silly enough to take it, grandiose military ambitions or no.
(And I'm serious )
Temporary ban on AI/AGI discussions.
Sorry @Leon
https://www.theguardian.com/politics/2024/mar/30/she-still-carries-an-aura-of-spectacular-failure-why-hasnt-liz-truss-gone-away
I am sure Liz Truss's version is the bondage version - I think the chain attaches to the bale differently, but even that construction doesn't seem to be wholly uncommon.
However, I still think that really, she believed that her Prime Ministerial mandate would see her through, and presumably that the Tory Party would let her ride it out until her policies started to bear economical fruit by the election.
https://www.bbc.co.uk/news/uk-wales-68609431
So the people that liked this tweet - @Foxy, @bondegezou and @Mexicanpete - don't actually have anything interesting to say, about anything else
Legendary labor leader Cesar Chavez’s family is outraged that RFK Jr is exploiting their dad’s legacy by using an iconic photo of him to promote a ‘Cesar Chavez Rally with RFK Jr’ in LA tomorrow. They are all supporting Biden.
https://twitter.com/RonFilipkowski/status/1773806050786849105
Very uplifting.
Don't let Labour ruin it.
Fully unmanned drone combat in Donetsk Oblast.
Seen here, Ukrainian FPV munitions from the 47th Mechanized Brigade destroy a pair of Russian unmanned ground vehicles that attempted to push into Berdychi.
https://twitter.com/Osinttechnical/status/1774062457746604463
https://twitter.com/Prolapsarian/status/1774030136737730661
This makes about as much sense was banning foreign tourism to bring down the immigration figures. It’s just self destructive.
Edit - and it will be first week of July so he can say 'see? Second half of the year'
If they are in financial trouble it suggests there has been some ponzi-style accounting.
I can share the experiences of others confirming it's a very pleasant afternoon with plenty of people out and about.
That said, glad to be back home and ready to enter the "fray" on here.
As I don't believe in bringing a knife to a gun fight, I don't comment on AI though the domination of topic after topic by one individual on one subject is in itself a sign.
The problem is there's very little happening in politics currently. The polls aren't moving yet and the Conservatives seem to have mentally given up - "sleepwalking to disaster" as someone once said. A significant Labour majority seems probablr though the implications of that for the medium term political outlook seem uncertain.
As in early 1997, there's a view among some supporters of the Government the incoming Labour administration will swiftly beome unpopular and the country will suddenly clamour for the return of the Tories. That seems improbable given Labour can heap blams on the outgoing Tories for most of what is happening and some (perhaps not many) of those Conservatives responsble will be on the opposition benches.
1. Whilst its talked about as being of a historical nature, its not quite as far back as that term suggests
2. The fact that there are two women coming forward is a pointer
3. Despite how some police forces seem to work these days, tipping off the media before anyone is even charged, this didnt leak at all. I understand that any idea that other senior UK politicians might have been aware is incorrect.
4. Despite the usual online speculation that this is a dark forces conspiracy/stitch up/malevolent elites secnario, there is no evidence of that
5. Questioning to charge was extremely quick
5. Whether guilty or not, the nature of the allegations when they come out means there is not a hope in hell of him coming back. The only chance of a comeback of any sort at all is the case is dropped before it reaches court and that seems very unlikely.