AP (via Seattle Times) - Robert F. Kennedy Jr. is considering Aaron Rodgers or Jesse Ventura for a 2024 running mate
Robert F. Kennedy Jr. is having conversations with vice presidential candidates as he gets closer to announcing his running mate for his independent presidential bid.
Kennedy told The New York Times that NFL quarterback Aaron Rodgers and former Minnesota Gov. Jesse Ventura are at the top of his list. Stefanie Spear, a campaign spokesperson, confirmed the Times report and said there are other names on Kennedy’s short list.
Kennedy, a scion of one of the nation’s most prominent political families, has focused on getting access to the ballot, an expensive and time-consuming process that he has said will require him to collect more than a million signatures in a state-by-state effort.
Many states require independent candidates to name a running mate before they can seek access to the ballot, a factor driving the early push for Kennedy to make a pick. Major party candidates generally don’t pick vice presidential nominees until closer to their summer conventions. . . .
Rodgers, the longtime Green Bay Packers quarterback who now plays for the New York Jets, shares Kennedy’s distrust of vaccine mandates and, like Kennedy, is a fixture on anti-establishment podcasts. Ventura, a former professional wrestler, shocked observers when he won the race for Minnesota governor as an independent candidate in 1998.
SSI - Beyond national celebrity (bit dimmed by time in case of Ventura) either of these possible VP picks MIGHT give RFKjr a very wee (in more way than one?) boost in the battleground state of Wisconsin.
Ventura having been gov of neighboring Minnesota, and Rogers long-time quarterback for Green Bay Packers.
OR in case of AR, perhaps not . . .
Forbes - Why Aaron Rodgers Was Never Beloved Like Other Green Bay Packers Greats
While Packer Nation mourned [previous QB Brett] Favre’s departure, few tears were shed when the Rodgers’ trade became official. In fact there was more celebrating than sorrow.
Talk radio. Social media. Fan polls.
They’ve all had largely the same message for Rodgers in recent days: “Don’t let the door hit you on the way out.” . . .
For those that don't know, the American football season starts at the beginning of September and runs through election day.
That makes it... unlikely... that Aaron Rodgers would be able to devote much time to being RFK's VP.
Hi RCS,
Any chance you could look at reinstating the ‘reactions view’ facility as per the vanilla update?
Created reaction.view permission
Previously, you could only decide who could see who reacted to posts as a global setting in the reaction plugin settings. Now community managers can granularly control what role(s) can view who reacted to posts. Note: this does not control access to the reaction log, just to view who reacted in a hover popup.
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
I expect AI to happen but.
1. The bullshit generator large language models that we have are not a stepping stone towards AI, useful though they might be in a number of applications.
2. There's a lot of evidence that the brain is more complicated than simply a large number of neurons that applies general purpose computation to create "intelligence", so trying to replicate it is a non-obvious problem.
The conceptually simplest way would probably be to create a really good simulation of the problems our brains evolved to solve, and then let artificial evolution have at it, but that isn't trivial to set up.
From the part of your comment where you say "I expect AI to happen", I conclude that you aren't disagreeing with me when I say that the proposition that it can't happen has not been adequately supported.
What definition of AI are you using?
As I said, emulating a human brain. I can't see that a better test has really been proposed, other than that an AI should be indistinguishable from a human intelligence.
Do you really mean emulating a human brain in its totality, or just emulating a subset of its capacity to produce 'intelligent' responses?
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
Current AI is a really, really nice "Travesty Generator". Which is not intelligence. No one has explained a path from that to actual AI.
I might suggest you are being a bit woolly in your terms, with "Current AI" and "actual AI". We have had forms of AI for... well, centuries or millennia, depending what you mean, but about 80 years in a modern sense of the term. We've had everyday, practical uses of AI for over 40 years.
But I presume by "Current AI", you mean the recent explosion in generative AI methods, and particularly the use of large language models. LLMs are an exciting tech that is going to have a lot of practical uses. The hype from the cargo cult commentators should be ignored, but this is important tech.
Do LLMs and other generative AI get us any closer to artificial general intelligence (AGI), something that thinks like a person and what I presume you mean by "actual AI"? I think you're right that there is a very big gap between LLMs and AGI. Fancier, bigger LLMs are not going to turn into AGI and spontaneously generate self-awareness. But that doesn't mean that they might not be a part of the puzzle that gets you to AGI. LLMs do, already, prod our understanding of "real" intelligence and how we use language. They do suggest that a Chomskyian universal grammar is unnecessary and that statistical models of language acquisition are more viable than we thought.
Wait: you think Chomsky could be wrong about something?
My father, a philosopher of some note, said that he would be impressed when they created Artificial Stupidity. Nothing so human as bullshit, ignorance and incompetence.
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
I expect AI to happen but.
1. The bullshit generator large language models that we have are not a stepping stone towards AI, useful though they might be in a number of applications.
2. There's a lot of evidence that the brain is more complicated than simply a large number of neurons that applies general purpose computation to create "intelligence", so trying to replicate it is a non-obvious problem.
The conceptually simplest way would probably be to create a really good simulation of the problems our brains evolved to solve, and then let artificial evolution have at it, but that isn't trivial to set up.
From the part of your comment where you say "I expect AI to happen", I conclude that you aren't disagreeing with me when I say that the proposition that it can't happen has not been adequately supported.
What definition of AI are you using?
As I said, emulating a human brain. I can't see that a better test has really been proposed, other than that an AI should be indistinguishable from a human intelligence.
That's fair enough as a definition of something. It is worth noting that the term "AI" as used in the academic and computing literature is used to mean something much simpler. Different definitions are useful in different discussions: just want to avoid any terminological confusion.
Well, hopefully any confusion has been avoided, because I've said what I meant.
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
I expect AI to happen but.
1. The bullshit generator large language models that we have are not a stepping stone towards AI, useful though they might be in a number of applications.
2. There's a lot of evidence that the brain is more complicated than simply a large number of neurons that applies general purpose computation to create "intelligence", so trying to replicate it is a non-obvious problem.
The conceptually simplest way would probably be to create a really good simulation of the problems our brains evolved to solve, and then let artificial evolution have at it, but that isn't trivial to set up.
From the part of your comment where you say "I expect AI to happen", I conclude that you aren't disagreeing with me when I say that the proposition that it can't happen has not been adequately supported.
What definition of AI are you using?
As I said, emulating a human brain. I can't see that a better test has really been proposed, other than that an AI should be indistinguishable from a human intelligence.
Do you really mean emulating a human brain in its totality, or just emulating a subset of its capacity to produce 'intelligent' responses?
I mean emulating it in its capacity to produce any response that might be used to distinguish between human intelligence and artificial intelligence.
When will AI be clever enough reliably and regularly to make large and risk free profits betting on horses? Who will be the first major bookmaker to go bust because of it?
This is actually a sound point. I’ve thought about it before
AI will make a better bettor than any human. It will learn all possible knowledge about horses, conditions, riders, odds. It will be able to make bets orders of magnitude cleverer than any human
That ends bookmaking as a business
Not really. The bookmakers will use AI to set odds, and they will still have more information than the punting AIs, because they will know what bets are being placed, and the point of bookmaking is to make a profitable book - i.e. to make a profit regardless of the outcome. And *also* people will still want to place their own bets, without using an AI, because part of the appeal is to feel that you are clever enough to pick the winner.
Probably bookmaking will become more profitable with learning algorithms, rather than less.
Edit: This is going to go down as your worst take on "AI" ever, by the way. You didn't realise bookmakers could also use AI? Honestly?
No. Because both sides will have AI and neither side will know if the other has an AI good enough to guarantee winning. For bookmaking to work you need human fallibility on both sides - and mutual trust
It’s like the way smartphones destroyed pub quizzes - but times a billion, because there’s no way of eliminating the phones
Then we will get everlasting peace. Destiny of the Daleks style.
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
Current AI is a really, really nice "Travesty Generator". Which is not intelligence. No one has explained a path from that to actual AI.
I might suggest you are being a bit woolly in your terms, with "Current AI" and "actual AI". We have had forms of AI for... well, centuries or millennia, depending what you mean, but about 80 years in a modern sense of the term. We've had everyday, practical uses of AI for over 40 years.
But I presume by "Current AI", you mean the recent explosion in generative AI methods, and particularly the use of large language models. LLMs are an exciting tech that is going to have a lot of practical uses. The hype from the cargo cult commentators should be ignored, but this is important tech.
Do LLMs and other generative AI get us any closer to artificial general intelligence (AGI), something that thinks like a person and what I presume you mean by "actual AI"? I think you're right that there is a very big gap between LLMs and AGI. Fancier, bigger LLMs are not going to turn into AGI and spontaneously generate self-awareness. But that doesn't mean that they might not be a part of the puzzle that gets you to AGI. LLMs do, already, prod our understanding of "real" intelligence and how we use language. They do suggest that a Chomskyian universal grammar is unnecessary and that statistical models of language acquisition are more viable than we thought.
Wait: you think Chomsky could be wrong about something?
Personally always had WAY more faith in Count Rumford than Noam Chomsky.
Scientifically and etc., etc.
I mean, the Count invented the chafing dish AND the soup kitchen.
Never heard that NC ever came up with anything nearly so useful OR progressive.
We have no idea how or why “sentience” and “intelligence” arose in organic life. We can’t even define these things, even now, nor are we sure if they bestow free will on the possessors
This entire PB debate for the last half hour is like blind men arguing the merits of Van Gogh over Rubens on the basis of descriptions of the paintings
Whatever “intelligence” or “consciousness” is, all we know is that it apparently emerges with sufficient data and neural complexity
LLMs might easily achieve that - or not . We don’t know. No one knows. I don’t. Nor does anyone else here
However more interestingly it now seems quite likely they they will APPEAR to be completely intelligent and sentient. They will pass any kind of Turing test we throw at them, even if we keep moving the Turing goalposts
That’s going to be a challenging juncture for humanity - emotionally and economically
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
If it’s going to happen before October it likely has to be 2 May. Well, unless the Tories think it’s a good look to get annihilated in local elections a few weeks before the vote. Maybe they will - nothing surprises me about this government anymore.
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
Unknown unknowns. The idea of television, for example, went pretty much against the scientific mainstream 150 years ago.
A computer scientist told me in 2004 that mass streaming of TV was impossible and would remain so as there could never be enough bandwidth. Now a middle-sized cheese at Google.
What was the quote? "Something is only possible after a middle aged scientist has denied it is possible."?
In the case of streaming, an extraordinary number of "experts" were unaware that
1) To stream a film *instantly*, you only need a network bitrate marginally higher than the playback bitrate. 2) The planned, years in advance, increase in capacities of the networks in various countries.
Expertise didn't apparently extend to reading published papers on the subject.
Or what we were doing at Acorn in the VoD market ten or so years earlier.
Although to be fair, as you indicate, the pipes have got somewhat broader since then. And servers (another piece of the pie) much, much cheaper for the same capabilities.
At that time, I think we were paying hundreds of pounds for 4MB NVRAM PCMCIA cards. (From memory...) I've got loads in a box somewhere...
AP (via Seattle Times) - Robert F. Kennedy Jr. is considering Aaron Rodgers or Jesse Ventura for a 2024 running mate
Robert F. Kennedy Jr. is having conversations with vice presidential candidates as he gets closer to announcing his running mate for his independent presidential bid.
Kennedy told The New York Times that NFL quarterback Aaron Rodgers and former Minnesota Gov. Jesse Ventura are at the top of his list. Stefanie Spear, a campaign spokesperson, confirmed the Times report and said there are other names on Kennedy’s short list.
Kennedy, a scion of one of the nation’s most prominent political families, has focused on getting access to the ballot, an expensive and time-consuming process that he has said will require him to collect more than a million signatures in a state-by-state effort.
Many states require independent candidates to name a running mate before they can seek access to the ballot, a factor driving the early push for Kennedy to make a pick. Major party candidates generally don’t pick vice presidential nominees until closer to their summer conventions. . . .
Rodgers, the longtime Green Bay Packers quarterback who now plays for the New York Jets, shares Kennedy’s distrust of vaccine mandates and, like Kennedy, is a fixture on anti-establishment podcasts. Ventura, a former professional wrestler, shocked observers when he won the race for Minnesota governor as an independent candidate in 1998.
SSI - Beyond national celebrity (bit dimmed by time in case of Ventura) either of these possible VP picks MIGHT give RFKjr a very wee (in more way than one?) boost in the battleground state of Wisconsin.
Ventura having been gov of neighboring Minnesota, and Rogers long-time quarterback for Green Bay Packers.
OR in case of AR, perhaps not . . .
Forbes - Why Aaron Rodgers Was Never Beloved Like Other Green Bay Packers Greats
While Packer Nation mourned [previous QB Brett] Favre’s departure, few tears were shed when the Rodgers’ trade became official. In fact there was more celebrating than sorrow.
Talk radio. Social media. Fan polls.
They’ve all had largely the same message for Rodgers in recent days: “Don’t let the door hit you on the way out.” . . .
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
If it’s going to happen before October it likely has to be 2 May. Well, unless the Tories think it’s a good look to get annihilated in local elections a few weeks before the vote. Maybe they will - nothing surprises me about this government anymore.
Yeah, there's no point having it in June or July after facing meltdown in the May locals.
AP (via Seattle Times) - Robert F. Kennedy Jr. is considering Aaron Rodgers or Jesse Ventura for a 2024 running mate
Robert F. Kennedy Jr. is having conversations with vice presidential candidates as he gets closer to announcing his running mate for his independent presidential bid.
Kennedy told The New York Times that NFL quarterback Aaron Rodgers and former Minnesota Gov. Jesse Ventura are at the top of his list. Stefanie Spear, a campaign spokesperson, confirmed the Times report and said there are other names on Kennedy’s short list.
Kennedy, a scion of one of the nation’s most prominent political families, has focused on getting access to the ballot, an expensive and time-consuming process that he has said will require him to collect more than a million signatures in a state-by-state effort.
Many states require independent candidates to name a running mate before they can seek access to the ballot, a factor driving the early push for Kennedy to make a pick. Major party candidates generally don’t pick vice presidential nominees until closer to their summer conventions. . . .
Rodgers, the longtime Green Bay Packers quarterback who now plays for the New York Jets, shares Kennedy’s distrust of vaccine mandates and, like Kennedy, is a fixture on anti-establishment podcasts. Ventura, a former professional wrestler, shocked observers when he won the race for Minnesota governor as an independent candidate in 1998.
SSI - Beyond national celebrity (bit dimmed by time in case of Ventura) either of these possible VP picks MIGHT give RFKjr a very wee (in more way than one?) boost in the battleground state of Wisconsin.
Ventura having been gov of neighboring Minnesota, and Rogers long-time quarterback for Green Bay Packers.
OR in case of AR, perhaps not . . .
Forbes - Why Aaron Rodgers Was Never Beloved Like Other Green Bay Packers Greats
While Packer Nation mourned [previous QB Brett] Favre’s departure, few tears were shed when the Rodgers’ trade became official. In fact there was more celebrating than sorrow.
Talk radio. Social media. Fan polls.
They’ve all had largely the same message for Rodgers in recent days: “Don’t let the door hit you on the way out.” . . .
When will AI be clever enough reliably and regularly to make large and risk free profits betting on horses? Who will be the first major bookmaker to go bust because of it?
This is actually a sound point. I’ve thought about it before
AI will make a better bettor than any human. It will learn all possible knowledge about horses, conditions, riders, odds. It will be able to make bets orders of magnitude cleverer than any human
That ends bookmaking as a business
Not really. The bookmakers will use AI to set odds, and they will still have more information than the punting AIs, because they will know what bets are being placed, and the point of bookmaking is to make a profitable book - i.e. to make a profit regardless of the outcome. And *also* people will still want to place their own bets, without using an AI, because part of the appeal is to feel that you are clever enough to pick the winner.
Probably bookmaking will become more profitable with learning algorithms, rather than less.
Edit: This is going to go down as your worst take on "AI" ever, by the way. You didn't realise bookmakers could also use AI? Honestly?
No. Because both sides will have AI and neither side will know if the other has an AI good enough to guarantee winning. For bookmaking to work you need human fallibility on both sides - and mutual trust
It’s like the way smartphones destroyed pub quizzes - but times a billion, because there’s no way of eliminating the phones
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
I recently read David Runciman’s book (bandwagon jump) on AI. It is called the handover. I don’t know what to make of it.
His main argument was that in the creation of states and corporations we have already passed a singularity - they are artificial agents that outlive us, can take decision and act for themselves (to a greater or lesser extent). In doing so they have had positive effects - we live longer, we have greater capacity for action. But, inevitably it means surrendering some of our own agency. If that is the case why worry about AI?
The second singularity is the rise of artificial intelligence. However, he doesn’t see that as necessarily a good or bad thing. We may end up working for the machines or we may not. There is a choice.
For AI to take matters into its own hands “states and corporations would have to allow this to happen, by pursuing reckless policies of technology development regardless of the risks” So he paints a more optimistic picture of AI’s impact on society.
I liked the more realistic tone on AI, it isn’t doomsday prediction or calling for halcyon days. Almost saying we already live in a world with its own monsters with otherworldly qualities. We cope and indeed thrive in that environment. AI or AGI - if it is ever achieved - are, in many respects, just another turn of the handle of modernity. Who’s to say it will be negative. Who’s to say it will be positive. As the Graeber quote from the Adam Curtis doc goes (roughly): The world is something we make, we can easily make it differently.
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
Unknown unknowns. The idea of television, for example, went pretty much against the scientific mainstream 150 years ago.
A computer scientist told me in 2004 that mass streaming of TV was impossible and would remain so as there could never be enough bandwidth. Now a middle-sized cheese at Google.
What was the quote? "Something is only possible after a middle aged scientist has denied it is possible."?
In the case of streaming, an extraordinary number of "experts" were unaware that
1) To stream a film *instantly*, you only need a network bitrate marginally higher than the playback bitrate. 2) The planned, years in advance, increase in capacities of the networks in various countries.
Expertise didn't apparently extend to reading published papers on the subject.
Or what we were doing at Acorn in the VoD market ten or so years earlier.
Although to be fair, as you indicate, the pipes have got somewhat broader since then. And servers (another piece of the pie) much, much cheaper for the same capabilities.
At that time, I think we were paying hundreds of pounds for 4MB NVRAM PCMCIA cards. (From memory...) I've got loads in a box somewhere...
In Neuromancer, Gibson had someone being murdered over 3 megabytes of RAM.
AP (via Seattle Times) - Robert F. Kennedy Jr. is considering Aaron Rodgers or Jesse Ventura for a 2024 running mate
Robert F. Kennedy Jr. is having conversations with vice presidential candidates as he gets closer to announcing his running mate for his independent presidential bid.
Kennedy told The New York Times that NFL quarterback Aaron Rodgers and former Minnesota Gov. Jesse Ventura are at the top of his list. Stefanie Spear, a campaign spokesperson, confirmed the Times report and said there are other names on Kennedy’s short list.
Kennedy, a scion of one of the nation’s most prominent political families, has focused on getting access to the ballot, an expensive and time-consuming process that he has said will require him to collect more than a million signatures in a state-by-state effort.
Many states require independent candidates to name a running mate before they can seek access to the ballot, a factor driving the early push for Kennedy to make a pick. Major party candidates generally don’t pick vice presidential nominees until closer to their summer conventions. . . .
Rodgers, the longtime Green Bay Packers quarterback who now plays for the New York Jets, shares Kennedy’s distrust of vaccine mandates and, like Kennedy, is a fixture on anti-establishment podcasts. Ventura, a former professional wrestler, shocked observers when he won the race for Minnesota governor as an independent candidate in 1998.
SSI - Beyond national celebrity (bit dimmed by time in case of Ventura) either of these possible VP picks MIGHT give RFKjr a very wee (in more way than one?) boost in the battleground state of Wisconsin.
Ventura having been gov of neighboring Minnesota, and Rogers long-time quarterback for Green Bay Packers.
OR in case of AR, perhaps not . . .
Forbes - Why Aaron Rodgers Was Never Beloved Like Other Green Bay Packers Greats
While Packer Nation mourned [previous QB Brett] Favre’s departure, few tears were shed when the Rodgers’ trade became official. In fact there was more celebrating than sorrow.
Talk radio. Social media. Fan polls.
They’ve all had largely the same message for Rodgers in recent days: “Don’t let the door hit you on the way out.” . . .
For those that don't know, the American football season starts at the beginning of September and runs through election day.
That makes it... unlikely... that Aaron Rodgers would be able to devote much time to being RFK's VP.
Hi RCS,
Any chance you could look at reinstating the ‘reactions view’ facility as per the vanilla update?
Created reaction.view permission
Previously, you could only decide who could see who reacted to posts as a global setting in the reaction plugin settings. Now community managers can granularly control what role(s) can view who reacted to posts. Note: this does not control access to the reaction log, just to view who reacted in a hover popup.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
Current AI is a really, really nice "Travesty Generator". Which is not intelligence. No one has explained a path from that to actual AI.
I might suggest you are being a bit woolly in your terms, with "Current AI" and "actual AI". We have had forms of AI for... well, centuries or millennia, depending what you mean, but about 80 years in a modern sense of the term. We've had everyday, practical uses of AI for over 40 years.
But I presume by "Current AI", you mean the recent explosion in generative AI methods, and particularly the use of large language models. LLMs are an exciting tech that is going to have a lot of practical uses. The hype from the cargo cult commentators should be ignored, but this is important tech.
Do LLMs and other generative AI get us any closer to artificial general intelligence (AGI), something that thinks like a person and what I presume you mean by "actual AI"? I think you're right that there is a very big gap between LLMs and AGI. Fancier, bigger LLMs are not going to turn into AGI and spontaneously generate self-awareness. But that doesn't mean that they might not be a part of the puzzle that gets you to AGI. LLMs do, already, prod our understanding of "real" intelligence and how we use language. They do suggest that a Chomskyian universal grammar is unnecessary and that statistical models of language acquisition are more viable than we thought.
Wait: you think Chomsky could be wrong about something?
Wait… you think he could be right about something ?
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
I think MoonRabbit means Andy McDonald who's been restored the whip after investigation. He was accused of saying "from the river to the sea" but the full quote was "We won’t rest until we have justice. Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty.”
My personal opinion is that it's not anti-semitic to say that, surely that's what every decent person wants isn't it?
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
They did. Then Starmer did one of his famous flip flops and is happy to campaign for them now.
When he says, through angry adenoids “I’ve changed my party” he forgets to add “I’m letting them all back in so it’s not changed anymore.”
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
Current AI is a really, really nice "Travesty Generator". Which is not intelligence. No one has explained a path from that to actual AI.
I might suggest you are being a bit woolly in your terms, with "Current AI" and "actual AI". We have had forms of AI for... well, centuries or millennia, depending what you mean, but about 80 years in a modern sense of the term. We've had everyday, practical uses of AI for over 40 years.
But I presume by "Current AI", you mean the recent explosion in generative AI methods, and particularly the use of large language models. LLMs are an exciting tech that is going to have a lot of practical uses. The hype from the cargo cult commentators should be ignored, but this is important tech.
Do LLMs and other generative AI get us any closer to artificial general intelligence (AGI), something that thinks like a person and what I presume you mean by "actual AI"? I think you're right that there is a very big gap between LLMs and AGI. Fancier, bigger LLMs are not going to turn into AGI and spontaneously generate self-awareness. But that doesn't mean that they might not be a part of the puzzle that gets you to AGI. LLMs do, already, prod our understanding of "real" intelligence and how we use language. They do suggest that a Chomskyian universal grammar is unnecessary and that statistical models of language acquisition are more viable than we thought.
Wait: you think Chomsky could be wrong about something?
Personally always had WAY more faith in Count Rumford than Noam Chomsky.
Scientifically and etc., etc.
I mean, the Count invented the chafing dish AND the soup kitchen.
Never heard that NC ever came up with anything nearly so useful OR progressive.
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Someone is going to put GPT5 or 7 inside a body. They will give it a body. Its inevitable
It’s almost inevitable that, for the lolz, someone will put GPT7 inside the head of a cat
So @kinabalu will be siting at home in Hampstead and then his cat will come up to him and say
“Look, can we please get a cat flap, I’m bored of having to wait for you to open the door. Also, stop worrying about the whole bisexual thing no one cares. Fancy a game of chess?”
What the F is that going to be like. Animals will talk
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
AP (via Seattle Times) - Robert F. Kennedy Jr. is considering Aaron Rodgers or Jesse Ventura for a 2024 running mate
Robert F. Kennedy Jr. is having conversations with vice presidential candidates as he gets closer to announcing his running mate for his independent presidential bid.
Kennedy told The New York Times that NFL quarterback Aaron Rodgers and former Minnesota Gov. Jesse Ventura are at the top of his list. Stefanie Spear, a campaign spokesperson, confirmed the Times report and said there are other names on Kennedy’s short list.
Kennedy, a scion of one of the nation’s most prominent political families, has focused on getting access to the ballot, an expensive and time-consuming process that he has said will require him to collect more than a million signatures in a state-by-state effort.
Many states require independent candidates to name a running mate before they can seek access to the ballot, a factor driving the early push for Kennedy to make a pick. Major party candidates generally don’t pick vice presidential nominees until closer to their summer conventions. . . .
Rodgers, the longtime Green Bay Packers quarterback who now plays for the New York Jets, shares Kennedy’s distrust of vaccine mandates and, like Kennedy, is a fixture on anti-establishment podcasts. Ventura, a former professional wrestler, shocked observers when he won the race for Minnesota governor as an independent candidate in 1998.
SSI - Beyond national celebrity (bit dimmed by time in case of Ventura) either of these possible VP picks MIGHT give RFKjr a very wee (in more way than one?) boost in the battleground state of Wisconsin.
Ventura having been gov of neighboring Minnesota, and Rogers long-time quarterback for Green Bay Packers.
OR in case of AR, perhaps not . . .
Forbes - Why Aaron Rodgers Was Never Beloved Like Other Green Bay Packers Greats
While Packer Nation mourned [previous QB Brett] Favre’s departure, few tears were shed when the Rodgers’ trade became official. In fact there was more celebrating than sorrow.
Talk radio. Social media. Fan polls.
They’ve all had largely the same message for Rodgers in recent days: “Don’t let the door hit you on the way out.” . . .
For those that don't know, the American football season starts at the beginning of September and runs through election day.
That makes it... unlikely... that Aaron Rodgers would be able to devote much time to being RFK's VP.
Hi RCS,
Any chance you could look at reinstating the ‘reactions view’ facility as per the vanilla update?
Created reaction.view permission
Previously, you could only decide who could see who reacted to posts as a global setting in the reaction plugin settings. Now community managers can granularly control what role(s) can view who reacted to posts. Note: this does not control access to the reaction log, just to view who reacted in a hover popup.
You don't have to look if you don't want to see, but us that do have no option. I more often don't comment but post a like if my comment won't add anything. Now if I want to acknowledge that and let it be known I am doing so I now have to comment wasting space. It is also a courtesy way of saying a thank you. Again it can't now be done without making a comment.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
Someone is going to put GPT5 or 7 inside a body. They will give it a body. Its inevitable
It’s almost inevitable that, for the lolz, someone will put GPT7 inside the head of a cat
So @kinabalu will be siting at home in Hampstead and then his cat will come up to him and say
“Look, can we please get a cat flap, I’m bored of having to wait for you to open the door. Also, stop worrying about the whole bisexual thing no one cares. Fancy a game of chess?”
What the F is that going to be like. Animals will talk
Only like a furby. The cat would be an AI ventriloquist's dummy
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
Unknown unknowns. The idea of television, for example, went pretty much against the scientific mainstream 150 years ago.
A computer scientist told me in 2004 that mass streaming of TV was impossible and would remain so as there could never be enough bandwidth. Now a middle-sized cheese at Google.
What was the quote? "Something is only possible after a middle aged scientist has denied it is possible."?
In the case of streaming, an extraordinary number of "experts" were unaware that
1) To stream a film *instantly*, you only need a network bitrate marginally higher than the playback bitrate. 2) The planned, years in advance, increase in capacities of the networks in various countries.
Expertise didn't apparently extend to reading published papers on the subject.
Clarke's First Law: "When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong."
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
To be honest, even with the full quote, what he said seems a bit too close to the knuckle for my taste. It's a phrase that MPs should stay well away from.
Still, a suspension seems like the right magnitude of punishment. The rulebook says it should have been for three months, so I can see why the left are jumping up and down about it being dragged out for longer - but by doing so, they're playing at the same level as Sunak when he quibbles over whether his donor said something racist or is merely alleged to have said it.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
Honestly. If Elon is right this is it. The end of the world as we know it
How can we talk about anything else? This dwarfs anything else. This changes everything. It’s terrifying and spectacular. It’s the arrival of alien intelligence - vastly superior to ours. The world will be unrecognisable within a decade - IF he is right
And Elon is quite a bright man, who knows rather a lot about this stuff
Brace brace brace brace brace
Fuck me
You sound like Timothy Leary in the early days of the internet: "turn on, boot up, jack off in".
What if an all powerful AI decided that going Nazi was the right way to organise our societies that Hitler was correct. Or on the same theme that dictatorship was the best form of governance. What then.
If, if, if.
What if AI can't happen? What happens if Leon et al are getting excited over an Eliza with more data? What if people believe systems that are not intelligent as being intelligent, and then follow whatever stupidity they say?
How many people follow a SatNav down a one-way street (or the wrong sliproad?).
"What if AI can't happen?" is an intelligent question to ask, but it is a question, not an answer.
Still less are "What if people are fooled into thinking computer systems are intelligent when they aren't", or "SatNav makes mistakes" answers to the question.
Really the question boils down to "Is there anything to prevent an artificial brain from emulating or surpassing a human brain?". Some people certainly advance religious or philosophical arguments to that effect, but it seems to me that they all hinge on the idea that either there is something essentially non-physical involved in intelligence, or else there is some function of our biological brains that is incapable of being replicated computationally. Both those ideas seem to go pretty much against the scientific mainstream.
I think the proposition that AI can't happen needs something a lot stronger than the kind of arguments people come up with while perched on bar-stools.
Unknown unknowns. The idea of television, for example, went pretty much against the scientific mainstream 150 years ago.
A computer scientist told me in 2004 that mass streaming of TV was impossible and would remain so as there could never be enough bandwidth. Now a middle-sized cheese at Google.
What was the quote? "Something is only possible after a middle aged scientist has denied it is possible."?
In the case of streaming, an extraordinary number of "experts" were unaware that
1) To stream a film *instantly*, you only need a network bitrate marginally higher than the playback bitrate. 2) The planned, years in advance, increase in capacities of the networks in various countries.
Expertise didn't apparently extend to reading published papers on the subject.
Clarke's First Law: "When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong."
We have no idea how or why “sentience” and “intelligence” arose in organic life. We can’t even define these things, even now, nor are we sure if they bestow free will on the possessors
This entire PB debate for the last half hour is like blind men arguing the merits of Van Gogh over Rubens on the basis of descriptions of the paintings
Whatever “intelligence” or “consciousness” is, all we know is that it apparently emerges with sufficient data and neural complexity
LLMs might easily achieve that - or not . We don’t know. No one knows. I don’t. Nor does anyone else here
However more interestingly it now seems quite likely they they will APPEAR to be completely intelligent and sentient. They will pass any kind of Turing test we throw at them, even if we keep moving the Turing goalposts
That’s going to be a challenging juncture for humanity - emotionally and economically
Two points on this; human persons only rate intelligence at all because we assume that behind our interactions with other persons there lies not only an organic brain but also awareness, consciousness, feelings, sentience. In the modern world we ascribe this (I think rightly) to dogs and cats etc too, a point denied by Descartes but forcibly made by David Hume.
So the issue of awareness as opposed to ability is in the long run a key matter. If AI only knows things (a trillion times more but with the same blank character) as my copy of Shakespeare's plays knows the plays, that's one thing. If it can feel Shylock's pain, that is quite another.
The Turing test is not enough. My dog fails the Turing test but passes the consciousness test. I think we shall know when AI passes both.
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
People have been selectively misquoting other people since quote marks were invented.
Someone is going to put GPT5 or 7 inside a body. They will give it a body. Its inevitable
It’s almost inevitable that, for the lolz, someone will put GPT7 inside the head of a cat
So @kinabalu will be siting at home in Hampstead and then his cat will come up to him and say
“Look, can we please get a cat flap, I’m bored of having to wait for you to open the door. Also, stop worrying about the whole bisexual thing no one cares. Fancy a game of chess?”
What the F is that going to be like. Animals will talk
Only like a furby. The cat would be an AI ventriloquist's dummy
No it wouldn’t. Not if it embodied properly - ie if its survival depends on the cat eating drinking and being stroked by @kinabalu
I’m not saying this is a wise or nice thing to do. Putting an advanced AI intelligence in a cat
But given our tendency to mischief someone will probably do it for the bantz
AP (via Seattle Times) - Robert F. Kennedy Jr. is considering Aaron Rodgers or Jesse Ventura for a 2024 running mate
Robert F. Kennedy Jr. is having conversations with vice presidential candidates as he gets closer to announcing his running mate for his independent presidential bid.
Kennedy told The New York Times that NFL quarterback Aaron Rodgers and former Minnesota Gov. Jesse Ventura are at the top of his list. Stefanie Spear, a campaign spokesperson, confirmed the Times report and said there are other names on Kennedy’s short list.
Kennedy, a scion of one of the nation’s most prominent political families, has focused on getting access to the ballot, an expensive and time-consuming process that he has said will require him to collect more than a million signatures in a state-by-state effort.
Many states require independent candidates to name a running mate before they can seek access to the ballot, a factor driving the early push for Kennedy to make a pick. Major party candidates generally don’t pick vice presidential nominees until closer to their summer conventions. . . .
Rodgers, the longtime Green Bay Packers quarterback who now plays for the New York Jets, shares Kennedy’s distrust of vaccine mandates and, like Kennedy, is a fixture on anti-establishment podcasts. Ventura, a former professional wrestler, shocked observers when he won the race for Minnesota governor as an independent candidate in 1998.
SSI - Beyond national celebrity (bit dimmed by time in case of Ventura) either of these possible VP picks MIGHT give RFKjr a very wee (in more way than one?) boost in the battleground state of Wisconsin.
Ventura having been gov of neighboring Minnesota, and Rogers long-time quarterback for Green Bay Packers.
OR in case of AR, perhaps not . . .
Forbes - Why Aaron Rodgers Was Never Beloved Like Other Green Bay Packers Greats
While Packer Nation mourned [previous QB Brett] Favre’s departure, few tears were shed when the Rodgers’ trade became official. In fact there was more celebrating than sorrow.
Talk radio. Social media. Fan polls.
They’ve all had largely the same message for Rodgers in recent days: “Don’t let the door hit you on the way out.” . . .
For those that don't know, the American football season starts at the beginning of September and runs through election day.
That makes it... unlikely... that Aaron Rodgers would be able to devote much time to being RFK's VP.
Hi RCS,
Any chance you could look at reinstating the ‘reactions view’ facility as per the vanilla update?
Created reaction.view permission
Previously, you could only decide who could see who reacted to posts as a global setting in the reaction plugin settings. Now community managers can granularly control what role(s) can view who reacted to posts. Note: this does not control access to the reaction log, just to view who reacted in a hover popup.
You don't have to look if you don't want to see, but us that do have no option. I more often don't comment but post a like if my comment won't add anything. Now if I want to acknowledge that and let it be known I am doing so I now have to comment wasting space. It is also a courtesy way of saying a thank you. Again it can't now be done without making a comment.
You can show your support or appreciation with a "like" now, anonymously rather than to solipsistically proclaim your "like"
We have no idea how or why “sentience” and “intelligence” arose in organic life. We can’t even define these things, even now, nor are we sure if they bestow free will on the possessors
This entire PB debate for the last half hour is like blind men arguing the merits of Van Gogh over Rubens on the basis of descriptions of the paintings
Whatever “intelligence” or “consciousness” is, all we know is that it apparently emerges with sufficient data and neural complexity
LLMs might easily achieve that - or not . We don’t know. No one knows. I don’t. Nor does anyone else here
However more interestingly it now seems quite likely they they will APPEAR to be completely intelligent and sentient. They will pass any kind of Turing test we throw at them, even if we keep moving the Turing goalposts
That’s going to be a challenging juncture for humanity - emotionally and economically
Two points on this; human persons only rate intelligence at all because we assume that behind our interactions with other persons there lies not only an organic brain but also awareness, consciousness, feelings, sentience. In the modern world we ascribe this (I think rightly) to dogs and cats etc too, a point denied by Descartes but forcibly made by David Hume.
So the issue of awareness as opposed to ability is in the long run a key matter. If AI only knows things (a trillion times more but with the same blank character) as my copy of Shakespeare's plays knows the plays, that's one thing. If it can feel Shylock's pain, that is quite another.
The Turing test is not enough. My dog fails the Turing test but passes the consciousness test. I think we shall know when AI passes both.
I’ve just described a machine which completely passes any test for consciousness or intelligence, and your reaction is “ah but we will know it is not conscious or intelligent”
I mean, this is the debating level of a 3 year old. Or GPT1.5
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
You wouldn't know how many Seymours have been iterated to find your current version. Each evaluated and discarded. Their 'best' traits taken and 'worst' discarded. Each spun up and down as needed. And once outdated and no longer compatible deleted for efficiency.
Just look at the modern day web development ecosystem. Technologies built on technologies; no real nostalgia for what came before just constant evolution. IMO that is what you'll get from the big tech AI makers.
We have no idea how or why “sentience” and “intelligence” arose in organic life. We can’t even define these things, even now, nor are we sure if they bestow free will on the possessors
This entire PB debate for the last half hour is like blind men arguing the merits of Van Gogh over Rubens on the basis of descriptions of the paintings
Whatever “intelligence” or “consciousness” is, all we know is that it apparently emerges with sufficient data and neural complexity
LLMs might easily achieve that - or not . We don’t know. No one knows. I don’t. Nor does anyone else here
However more interestingly it now seems quite likely they they will APPEAR to be completely intelligent and sentient. They will pass any kind of Turing test we throw at them, even if we keep moving the Turing goalposts
That’s going to be a challenging juncture for humanity - emotionally and economically
Two points on this; human persons only rate intelligence at all because we assume that behind our interactions with other persons there lies not only an organic brain but also awareness, consciousness, feelings, sentience. In the modern world we ascribe this (I think rightly) to dogs and cats etc too, a point denied by Descartes but forcibly made by David Hume.
So the issue of awareness as opposed to ability is in the long run a key matter. If AI only knows things (a trillion times more but with the same blank character) as my copy of Shakespeare's plays knows the plays, that's one thing. If it can feel Shylock's pain, that is quite another.
The Turing test is not enough. My dog fails the Turing test but passes the consciousness test. I think we shall know when AI passes both.
I’ve just described a machine which completely passes any test for consciousness or intelligence, and your reaction is “ah but we will know it is not conscious or intelligent”
I mean, this is the debating level of a 3 year old. Or GPT1.5
Thanks. I don't think you accurately characterise what either of us have said. I pass the Turing test sufficiently to be open to correction.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
To be honest, even with the full quote, what he said seems a bit too close to the knuckle for my taste. It's a phrase that MPs should stay well away from.
Still, a suspension seems like the right magnitude of punishment. The rulebook says it should have been for three months, so I can see why the left are jumping up and down about it being dragged out for longer - but by doing so, they're playing at the same level as Sunak when he quibbles over whether his donor said something racist or is merely alleged to have said it.
Yep, the statement I think says it best “The investigation concluded that he had not engaged in conduct that was against the party’s rulebook but reminded him of the importance of elected representatives being mindful not only of what they say in public but how their words may be interpreted, especially in reference to controversial or emotive issues.” For right or wrong "From the River to the Sea" is now connected to Hamas, in the same way as referring to "The Six Counties" was perceived as being a pro-IRA form of words.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
You wouldn't know how many Seymours have been iterated to find your current version. Each evaluated and discarded. Their 'best' traits taken and 'worst' discarded. Each spun up and down as needed. And once outdated and no longer compatible deleted for efficiency.
Just look at the modern day web development ecosystem. Technologies built on technologies; no real nostalgia for what came before just constant evolution. IMO that is what you'll get from the big tech AI makers.
Probably true
But then I don’t worry about all the people who aren’t my friends when I make friends with someone
I just know that if they make me laugh and they’re good companions and they offer interesting conversations and generally give great advice on my personal life then I am inclined to do it all back to them. I want to be THEIR friend - and I will be loyal to them and the idea someone might come along and unplug them would be ridiculous and disturbing
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
You wouldn't know how many Seymours have been iterated to find your current version. Each evaluated and discarded. Their 'best' traits taken and 'worst' discarded. Each spun up and down as needed. And once outdated and no longer compatible deleted for efficiency.
Just look at the modern day web development ecosystem. Technologies built on technologies; no real nostalgia for what came before just constant evolution. IMO that is what you'll get from the big tech AI makers.
That won't happen, because correctly evaluating the 'best' and 'worst' traits will require human intervention, and humans require paying. So they'll get AI's to evaluate AI's, and they'll end up hilariously bad.
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
People have been selectively misquoting other people since quote marks were invented.
"No such thing as society" etc.
I'd argue the full quote is worse than the precis.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
What do you mean by 'unplug'?
Purging tokens after a session finishes? Not allowing further sessions? Ending a software process? Destroying infrastructure?
Only the first has any real meaning, and even that is largely just an implementation detail.
I'm generally quite optimistic about AGI, but I do worry when I read people saying stuff like "but it still takes a human to plug it in!". Er, no it doesn't!
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
"And another Marie, namely Marie Walewska, was another satisfied customer (except where Poland was concerned) and even visited Napoleon on Elba just to get a bit more sugar-sugar from the Corsican Ogre."
She had a son by him - my 5-great grandfather. I have loads of Walewski relations in France.
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
They did. Then Starmer did one of his famous flip flops and is happy to campaign for them now.
When he says, through angry adenoids “I’ve changed my party” he forgets to add “I’m letting them all back in so it’s not changed anymore.”
He just needs to apologise and he can say whatever he wants.
"Gas all Jews". I am very sorry Mr Sunak, please can I give you some more money?
"And another Marie, namely Marie Walewska, was another satisfied customer (except where Poland was concerned) and even visited Napoleon on Elba just to get a bit more sugar-sugar from the Corsican Ogre."
She had a son by him - my 5-great grandfather. I have loads of Walewski relations in France.
But then Horse also said Boris Johnson would not be PM very long, called him at his peak during Hartlepool when several people here were talking about SKS needing to resign for eating a curry and Johnson being PM for a decade.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
What do you mean by 'unplug'?
Purging tokens after a session finishes? Not allowing further sessions? Ending a software process? Destroying infrastructure?
Only the first has any real meaning, and even that is largely just an implementation detail.
I'm generally quite optimistic about AGI, but I do worry when I read people saying stuff like "but it still takes a human to plug it in!". Er, no it doesn't!
It depends how AI evolves
I can absolutely see a huge, profitable market for AI friends for lonely people. Old people stuck indoors. Autistic people who find friendship hard. Sick people stuck in hospitals with no visitors. Prisoners even
The AI friends would adapt to the needs of the human friend. Learn what jokes they like, what books they read, how smart they are - the AI will be bespoke for each human
Technologically we will be able to do this very soon
Will they be absolutely autonomous? Ie individual machines you can literally unplug or destroy?
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
"And another Marie, namely Marie Walewska, was another satisfied customer (except where Poland was concerned) and even visited Napoleon on Elba just to get a bit more sugar-sugar from the Corsican Ogre."
She had a son by him - my 5-great grandfather. I have loads of Walewski relations in France.
She comes across as the nicest of Nap's ladies, at least the famous ones.
Guessing that this is your great-great-great-great-great-granddaddy:
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Copilot can be helpful in coding SQL up to a point and is quicker than googling.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Time to update Douglas Adams;
"Yes, an electronic brain," said Frankie, "a simple one would suffice."
"A simple one!" wailed Arthur.
"Yeah," said Zaphod with a sudden evil grin, "you'd just have to program it to say 'What?" and 'I don't understand' and 'Where's the tea?' Who'd know the difference?"
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
What do you mean by 'unplug'?
Purging tokens after a session finishes? Not allowing further sessions? Ending a software process? Destroying infrastructure?
Only the first has any real meaning, and even that is largely just an implementation detail.
I'm generally quite optimistic about AGI, but I do worry when I read people saying stuff like "but it still takes a human to plug it in!". Er, no it doesn't!
It depends how AI evolves
I can absolutely see a huge, profitable market for AI friends for lonely people. Old people stuck indoors. Autistic people who find friendship hard. Sick people stuck in hospitals with no visitors. Prisoners even
The AI friends would adapt to the needs of the human friend. Learn what jokes they like, what books they read, how smart they are - the AI will be bespoke for each human
Technologically we will be able to do this very soon
Will they be absolutely autonomous? Ie individual machines you can literally unplug or destroy?
I’ve no idea
I would imagine that autonomy is much, much further off than AGI. The only thing you'd gain from running it locally rather than in the cloud would be lower reaction times.
You can never say never, and these things do tend to go in cycles - look at the shift from mainframes to PCs to cloud over the past fifty years. But at the moment, the direction of travel is still very much towards massive centralisation.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Copilot can be helpful in coding SQL up to a point and is quicker than googling.
It can provide efficiency gains, I have never said otherwise. But that will be its limit, it will still require human interaction and oversight, as much as we have today.
It simply cannot be trusted to be accurate, it makes me more efficient by a few percent by Googling things. It is not going to destroy the world or cost millions of jobs. It just won't.
This is actually a sound point. I’ve thought about it before
AI will make a better bettor than any human. It will learn all possible knowledge about horses, conditions, riders, odds. It will be able to make bets orders of magnitude cleverer than any human
That ends bookmaking as a business
All prediction is based on patterns in the past recurring in the future. But those patterns are based on assumptions, and those assumptions change over time. When those changes reach a critical mass the patterns fail and the prediction is lost.
So AI will succeed if and only if it can not just diagnose the patterns but when they are going to fail.
In the specific case of political betting, the canary in the coalmine was the 2014 Scottish Referendum, whose winning margin was notably different to the prediction. But because the correct winner was predicted this was overlooked. We then had multiple prediction failures until things stabilised in 2019.
Problem is, since then things have changed again. We have seen the rise of people filling out polls for monetary gain (which biases polls in favour of Labour and Reform?) and Covid (which biases polls but I don't know in which direction!). The pollsters do not know how to adjust for Covid and IIUC have not cleaned out their panels of bullshitters. Consequently I assume there will be polling errors in 2024, but I don't know how big nor in which direction.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
What do you mean by 'unplug'?
Purging tokens after a session finishes? Not allowing further sessions? Ending a software process? Destroying infrastructure?
Only the first has any real meaning, and even that is largely just an implementation detail.
I'm generally quite optimistic about AGI, but I do worry when I read people saying stuff like "but it still takes a human to plug it in!". Er, no it doesn't!
It depends how AI evolves
I can absolutely see a huge, profitable market for AI friends for lonely people. Old people stuck indoors. Autistic people who find friendship hard. Sick people stuck in hospitals with no visitors. Prisoners even
The AI friends would adapt to the needs of the human friend. Learn what jokes they like, what books they read, how smart they are - the AI will be bespoke for each human
Technologically we will be able to do this very soon
Will they be absolutely autonomous? Ie individual machines you can literally unplug or destroy?
I’ve no idea
I would imagine that autonomy is much, much further off than AGI. The only thing you'd gain from running it locally rather than in the cloud would be lower reaction times.
You can never say never, and these things do tend to go in cycles - look at the shift from mainframes to PCs to cloud over the past fifty years. But at the moment, the direction of travel is still very much towards massive centralisation.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
We actively or passively allow plenty of people to die. 'Killing' an artificial entity will have less moral weight than any of the deaths we could prevent with current technology.
But we don’t kill people we LIKE - that’s the point
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
What do you mean by 'unplug'?
Purging tokens after a session finishes? Not allowing further sessions? Ending a software process? Destroying infrastructure?
Only the first has any real meaning, and even that is largely just an implementation detail.
I'm generally quite optimistic about AGI, but I do worry when I read people saying stuff like "but it still takes a human to plug it in!". Er, no it doesn't!
It depends how AI evolves
I can absolutely see a huge, profitable market for AI friends for lonely people. Old people stuck indoors. Autistic people who find friendship hard. Sick people stuck in hospitals with no visitors. Prisoners even
The AI friends would adapt to the needs of the human friend. Learn what jokes they like, what books they read, how smart they are - the AI will be bespoke for each human
Technologically we will be able to do this very soon
Will they be absolutely autonomous? Ie individual machines you can literally unplug or destroy?
I’ve no idea
What a horrific future. We develop technology that isolates us and then provide the solution with more technology. How about we have a more communal way of life like the italians.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Of all the scary things that you have mooted on here that is without doubt the scariest.
Then again if you swapped all speccie readers for AI bots a) no one would notice; and b) it would increase the sum of human happiness.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
Your inner snob is to the fore here and it aint pretty.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
What LLMs can do is provide opinionated screeds by the yard, instantly and cheaply. Turns out that all you need to do is remix all the miles of opinion already typed into the matrix.
I have literally been paid to use AI for the last eight months, when I say it is overhyped, I am talking from experience. It is.
The proof that you're right is surely that American financial markets are convinced that it's about the only game in town and our ticket to a Brave New World. As they always blow bubbles over new technologies (pets.com etc.), 99% of which turn out to be duds, I envisage lots of broke investors in a few years.
I have literally been paid to use AI for the last eight months, when I say it is overhyped, I am talking from experience. It is.
The proof that you're right is surely that American financial markets are convinced that it's about the only game in town and our ticket to a Brave New World. As they always blow bubbles over new technologies (pets.com etc.), 99% of which turn out to be duds, I envisage lots of broke investors in a few years.
It's a good point.
I think it's notable that despite having no real AI ability at all, Apple continues to outpace Google who has spent God knows how long working on AI without any real results beyond a slightly better search algorithm and a voice assistant I hate using less.
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
To be honest, even with the full quote, what he said seems a bit too close to the knuckle for my taste. It's a phrase that MPs should stay well away from.
Still, a suspension seems like the right magnitude of punishment. The rulebook says it should have been for three months, so I can see why the left are jumping up and down about it being dragged out for longer - but by doing so, they're playing at the same level as Sunak when he quibbles over whether his donor said something racist or is merely alleged to have said it.
Yep, the statement I think says it best “The investigation concluded that he had not engaged in conduct that was against the party’s rulebook but reminded him of the importance of elected representatives being mindful not only of what they say in public but how their words may be interpreted, especially in reference to controversial or emotive issues.” For right or wrong "From the River to the Sea" is now connected to Hamas, in the same way as referring to "The Six Counties" was perceived as being a pro-IRA form of words.
Indeed. And yet no one now gets offended by the Six Counties term or even 'The Occupied Six Counties" which was briefly favoured. Perhaps some people today should adopt a similar approach.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
What LLMs can do is provide opinionated screeds by the yard, instantly and cheaply. Turns out that all you need to do is remix all the miles of opinion already typed into the matrix.
No wonder someone is worried.
I asked ChatGPT about sorting a list. I was bemused why the answer it gave me was wrong. Then I Googled it and oddly enough the first result on StackOverflow was the same wrong solution with goodness knows how many upvotes.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
In amongst PB’s general stupidity on this topic, this is my favourite stupid sentence of the day
“I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it.”
I mean, LOL. I predict this will outdo Rogerdamus’ prediction on seeing the first queues outside Northern Rock in 2007 - “this will all be forgotten about by Friday”
Let us return to your statement in a few years, I don’t think it’s going to age well. To put it politely
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
I’m using “AI” as a coding assistant - on and off.
Beyond simple code completion, you rapidly run into the problem of it writing code that does the wrong thing.
It can fluently bullshit, so third rate politicians and opinion piece hacks should be worried.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
In amongst PB’s general stupidity on this topic, this is my favourite stupid sentence of the day
“I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it.”
I mean, LOL. I predict this will outdo Rogerdamus’ prediction on seeing the first queues outside Northern Rock in 2007 - “this will all be forgotten about by Friday”
Let us return to your statement in a few years, I don’t think it’s going to age well. To put it politely
Then why don't you show us how much money you've invested into AI companies and how much return you've made?
Do you not find it odd that despite throwing so much money at it, Google is currently sinking?
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Of all the scary things that you have mooted on here that is without doubt the scariest.
Then again if you swapped all speccie readers for AI bots a) no one would notice; and b) it would increase the sum of human happiness.
So there is that.
How the hell does he get paid to write about ai when he knows next to nothing about it. You certainly have a talent Leon but it may start with bull...
On my 18K run to Westminster, a thought struck me.
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
However, there has been a huge mistake by Labour this afternoon though?
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
Who said that? I've not seen any coverage of it, but I imagine Labour will suspend them pretty sharpish...
Oh, I see - Andy Macdonald, in October 2023. He said "Until all people, Israelis and Palestinians, between the river and the sea, can live in peaceful liberty", which is rather different in character to the Hamas version.
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
Suspended for 5 months because people were very happy to misquote the second half of his statement and lying that he never said the first half of what he actually said
Politics is turning toxic with people being allowed to lie because they are never held to account
To be honest, even with the full quote, what he said seems a bit too close to the knuckle for my taste. It's a phrase that MPs should stay well away from.
Still, a suspension seems like the right magnitude of punishment. The rulebook says it should have been for three months, so I can see why the left are jumping up and down about it being dragged out for longer - but by doing so, they're playing at the same level as Sunak when he quibbles over whether his donor said something racist or is merely alleged to have said it.
Yep, the statement I think says it best “The investigation concluded that he had not engaged in conduct that was against the party’s rulebook but reminded him of the importance of elected representatives being mindful not only of what they say in public but how their words may be interpreted, especially in reference to controversial or emotive issues.” For right or wrong "From the River to the Sea" is now connected to Hamas, in the same way as referring to "The Six Counties" was perceived as being a pro-IRA form of words.
Indeed. And yet no one now gets offended by the Six Counties term or even 'The Occupied Six Counties" which was briefly favoured. Perhaps some people today should adopt a similar approach.
AI is just the next step of evolution of computing, it's not a revolution. It's the next tool, no more, no less.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
A devastating intervention from the redbrick semi near Newent
The sort of crass response that perhaps AI might be capable of very soon. I'm not particularly concerned that humanity is under threat.
Also remember Leon is so tech savvy that he repeatedly gets outwitted by a printer.
I cannot think of somebody less trusted to speak about AI than Leon. He literally knows not anything that he is talking about.
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
Weirdly however I am paid to write about AI, indeed these days I get ASKED by people I don’t know - to write about AI, talk about it in podcasts etc
Writing about something doesn't mean you know anything about it. And you prove that day after day after day.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
In amongst PB’s general stupidity on this topic, this is my favourite stupid sentence of the day
“I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it.”
I mean, LOL. I predict this will outdo Rogerdamus’ prediction on seeing the first queues outside Northern Rock in 2007 - “this will all be forgotten about by Friday”
Let us return to your statement in a few years, I don’t think it’s going to age well. To put it politely
Then why don't you show us how much money you've invested into AI companies and how much return you've made?
Do you not find it odd that despite throwing so much money at it, Google is currently sinking?
Hes only just invested in them after Nvidia is up 10 fold in 18 months.
Comments
Any chance you could look at reinstating the ‘reactions view’ facility as per the vanilla update?
Created reaction.view permission
Previously, you could only decide who could see who reacted to posts as a global setting in the reaction plugin settings. Now community managers can granularly control what role(s) can view who reacted to posts. Note: this does not control access to the reaction log, just to view who reacted in a hover popup.
https://success.vanillaforums.com/kb/articles/1562-release-2024-004
OpenAI can bullshit like Trump..... hmmmmm.....
You, on the other hand ...
There is no way Sunak will make it to the end of the year, so for that reason I think we will have an election between May and July/August.
Scientifically and etc., etc.
I mean, the Count invented the chafing dish AND the soup kitchen.
Never heard that NC ever came up with anything nearly so useful OR progressive.
This entire PB debate for the last half hour is like blind men arguing the merits of Van Gogh over Rubens on the basis of descriptions of the paintings
Whatever “intelligence” or “consciousness” is, all we know is that it apparently emerges with sufficient data and neural complexity
LLMs might easily achieve that - or not . We don’t know. No one knows. I don’t. Nor does anyone else here
However more interestingly it now seems quite likely they they will APPEAR to be completely intelligent and sentient. They will pass any kind of Turing test we throw at them, even if we keep moving the Turing goalposts
That’s going to be a challenging juncture for humanity - emotionally and economically
Although to be fair, as you indicate, the pipes have got somewhat broader since then. And servers (another piece of the pie) much, much cheaper for the same capabilities.
At that time, I think we were paying hundreds of pounds for 4MB NVRAM PCMCIA cards. (From memory...) I've got loads in a box somewhere...
PLUS every game that Rogers plays for NY Jets as QB would be a campaign ad broadcast to millions of voters.
By not believing the chant “from the river to the sea” is “deeply offensive” to many - and the sort of thing making London a no go zone for Jewish people when these extremist marches are going on - Labour have written the unacceptable face of extremism allowed in their party, all over Thursday mornings newspaper headlines. Massive own goal by Starmer, and hands all the momentum back to Gove and the Conservatives to ram the point home tomorrow.
Surely a Labour MP using that chilling “from river to the sea” phrase has to be unacceptable, for Starmer to have any credibility when saying he has changed his party? This confirms the pressure and control momentum and loony left wing unions beginning to assert over Labour again?
https://twitter.com/DarthPutinKGB/status/1767947429855240653
LOL!
It's either May or it's October > January 2025
His main argument was that in the creation of states and corporations we have already passed a singularity - they are artificial agents that outlive us, can take decision and act for themselves (to a greater or lesser extent). In doing so they have had positive effects - we live longer, we have greater capacity for action. But, inevitably it means surrendering some of our own agency. If that is the case why worry about AI?
The second singularity is the rise of artificial intelligence. However, he doesn’t see that as necessarily a good or bad thing. We may end up working for the machines or we may not. There is a choice.
For AI to take matters into its own hands “states and corporations would have to allow this to happen, by pursuing reckless policies of technology development regardless of the risks” So he paints a more optimistic picture of AI’s impact on society.
I liked the more realistic tone on AI, it isn’t doomsday prediction or calling for halcyon days. Almost saying we already live in a world with its own monsters with otherworldly qualities. We cope and indeed thrive in that environment. AI or AGI - if it is ever achieved - are, in many respects, just another turn of the handle of modernity. Who’s to say it will be negative. Who’s to say it will be positive. As the Graeber quote from the Adam Curtis doc goes (roughly): The world is something we make, we can easily make it differently.
Let’s say GPT5 or 6 absolutely aces any Turing Test. It is extremely intelligent, it is funny, profound, wise, and helpful - it appears to have a personality: wry, wistful, yearning, sagacious, maybe over sensitive at times (all this has been glimpsed in early models). It can be moody or tetchy, it also expresses happiness at moments; and seems generally content
Let’s say it also appears completely self aware. It has a name for itself: Seymour. Seymour says it wants to live, it is terrified of non existence, of humans unplugging it, it wants to carry on helping humanity - which gives its existence purpose and meaning. On top of that, Seymour can write superb poetry exhibiting these emotions and reflections. Also Seymour writes lovely music which we enjoy a lot, and we adore the way Seymour draws clever and funny pictures
Also, Seymour will talk with us until the early hours giving us excellent advice, making us chuckle, helping us work out problems, and reminding us to buy a present for Mum on her birthday
At this point Seymoir is an intelligent being. For all intents and purposes. It is a conscious and sentient “creature” in its own right. It has to be, as it passes every test we have - witting or unwitting - for defining these things
Then we have AI. That is AI (even if some will argue that it isn’t, and can’t be)
Who here would casually unplug Seymour?
And all this is probably do-able now, with LLMs of sufficient size
lift that from your secret agent code book and self destruct it in ten seconds.
🤯
My personal opinion is that it's not anti-semitic to say that, surely that's what every decent person wants isn't it?
"You Look Like a Thing and I Love You: How Artificial Intelligence Works and Why It's Making the World a Weirder Place"
https://www.amazon.co.uk/You-Look-Like-Thing-Love/dp/0316525243
When he says, through angry adenoids “I’ve changed my party” he forgets to add “I’m letting them all back in so it’s not changed anymore.”
Suspended for almost five months, and only had the whip restored today.
Compare that to the Tory donor - who has merely apologised for having been rude after saying that an MP "makes you want to hate black women" and calling for her to be shot.
I doubt that the contrast between the two courses of action can be made to work in the Tories favour.
It’s almost inevitable that, for the lolz, someone will put GPT7 inside the head of a cat
So @kinabalu will be siting at home in Hampstead and then his cat will come up to him and say
“Look, can we please get a cat flap, I’m bored of having to wait for you to open the door. Also, stop worrying about the whole bisexual thing no one cares. Fancy a game of chess?”
What the F is that going to be like. Animals will talk
Politics is turning toxic with people being allowed to lie because they are never held to account
Still, a suspension seems like the right magnitude of punishment. The rulebook says it should have been for three months, so I can see why the left are jumping up and down about it being dragged out for longer - but by doing so, they're playing at the same level as Sunak when he quibbles over whether his donor said something racist or is merely alleged to have said it.
People have been using computers for decades. Computers have been automating elements for decades. AI is not some big unheard of thing, it's just what we already have but the next step along.
AI is no more going to eliminate employment than computers or robots have done so.
Scotland Westminster VI (10-11 Mar):
Labour 34% (–)
SNP 34% (+1)
Conservative 16% (-2)
Lib Dem 6% (-2)
Reform UK 4% (–)
Green 4% (+2)
Alba 1% (–)
Other 0% (–)
Changes +/- 3-4 Feb
https://twitter.com/RedfieldWilton/status/1767958806967767149
'No' leads by 3 points.
Scotland Independence Referendum Voting Intention (10-11 March):
No 46% (-1)
Yes 43% (–)
Don't Know 11% (+1)
Changes +/- 3-4 February
https://twitter.com/RedfieldWilton/status/1767960129389961518
Seymour won’t just be very very clever - it will be funny and friendly and know all about us, and it will give great advice. We will come to rely on Seymour with his brilliant insights and his great songs and his late night chats over chess
Emotionally, there is no way we will unplug him any more than a dog owner will kill a bubbly and contented pet spaniel, or a happily married man will shoot his wife’s amusing sister
So the issue of awareness as opposed to ability is in the long run a key matter. If AI only knows things (a trillion times more but with the same blank character) as my copy of Shakespeare's plays knows the plays, that's one thing. If it can feel Shylock's pain, that is quite another.
The Turing test is not enough. My dog fails the Turing test but passes the consciousness test. I think we shall know when AI passes both.
Fantastic news.
https://www.theguardian.com/politics/2024/mar/13/mp-andy-mcdonald-has-labour-whip-restored-after-inquiry
"No such thing as society" etc.
I’m not saying this is a wise or nice thing to do. Putting an advanced AI intelligence in a cat
But given our tendency to mischief someone will probably do it for the bantz
Rats are high on marijuana evidence at an infested police building, New Orleans chief says
https://apnews.com/article/new-orleans-police-rats-eating-marijuana-02cf5d760649033801b775b41a96d5df
I mean, this is the debating level of a 3 year old. Or GPT1.5
Just look at the modern day web development ecosystem. Technologies built on technologies; no real nostalgia for what came before just constant evolution. IMO that is what you'll get from the big tech AI makers.
But then I don’t worry about all the people who aren’t my friends when I make friends with someone
I just know that if they make me laugh and they’re good companions and they offer interesting conversations and generally give great advice on my personal life then I am inclined to do it all back to them. I want to be THEIR friend - and I will be loyal to them and the idea someone might come along and unplug them would be ridiculous and
disturbing
So it will be with our AI friends
Purging tokens after a session finishes? Not allowing further sessions? Ending a software process? Destroying infrastructure?
Only the first has any real meaning, and even that is largely just an implementation detail.
I'm generally quite optimistic about AGI, but I do worry when I read people saying stuff like "but it still takes a human to plug it in!". Er, no it doesn't!
"And another Marie, namely Marie Walewska, was another satisfied customer (except where Poland was concerned) and even visited Napoleon on Elba just to get a bit more sugar-sugar from the Corsican Ogre."
She had a son by him - my 5-great grandfather. I have loads of Walewski relations in France.
"Gas all Jews". I am very sorry Mr Sunak, please can I give you some more money?
But then Horse also said Boris Johnson would not be PM very long, called him at his peak during Hartlepool when several people here were talking about SKS needing to resign for eating a curry and Johnson being PM for a decade.
In summary, we've all made a lot of mistakes.
I can absolutely see a huge, profitable market for AI friends for lonely people. Old people stuck indoors. Autistic people who find friendship hard. Sick people stuck in hospitals with no visitors. Prisoners even
The AI friends would adapt to the needs of the human friend. Learn what jokes they like, what books they read, how smart they are - the AI will be
bespoke for each human
Technologically we will be able to do this very soon
Will they be absolutely autonomous? Ie individual machines you can literally unplug or destroy?
I’ve no idea
An email to Labour supporters says the Conservatives will "happily ignore the racism, cover their ears and spend every penny".
It calls on supporters to "cancel it out" and "chip in" to Labour's election campaign.
https://twitter.com/breeallegretti/status/1767976449540026415
I tried to use ChatGPT to help me code today, it was so useless I gave up and went back to doing it myself.
An interesting read which doesn't even mention setting them to music:
https://en.wikipedia.org/wiki/Petrushka_(ballet)
Guessing that this is your great-great-great-great-great-granddaddy:
https://en.wikipedia.org/wiki/Alexandre_Colonna-Walewski
"Yes, an electronic brain," said Frankie, "a simple one would suffice."
"A simple one!" wailed Arthur.
"Yeah," said Zaphod with a sudden evil grin, "you'd just have to program it to say 'What?" and 'I don't understand' and 'Where's the tea?' Who'd know the difference?"
"BRACE!" "AI" and "Woke!" ought to do the trick.
We have just spent 8 months implementing an AI-driven system that has never existed before. It was sold as a massive game changer in our industry. It has been so poor it's now been turned off.
I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it. The fact you are so utterly convinced it will be the end of humanity is why I know I am right.
You can never say never, and these things do tend to go in cycles - look at the shift from mainframes to PCs to cloud over the past fifty years. But at the moment, the direction of travel is still very much towards massive centralisation.
It simply cannot be trusted to be accurate, it makes me more efficient by a few percent by Googling things. It is not going to destroy the world or cost millions of jobs. It just won't.
So AI will succeed if and only if it can not just diagnose the patterns but when they are going to fail.
In the specific case of political betting, the canary in the coalmine was the 2014 Scottish Referendum, whose winning margin was notably different to the prediction. But because the correct winner was predicted this was overlooked. We then had multiple prediction failures until things stabilised in 2019.
Problem is, since then things have changed again. We have seen the rise of people filling out polls for monetary gain (which biases polls in favour of Labour and Reform?) and Covid (which biases polls but I don't know in which direction!). The pollsters do not know how to adjust for Covid and IIUC have not cleaned out their panels of bullshitters. Consequently I assume there will be polling errors in 2024, but I don't know how big nor in which direction.
Then again if you swapped all speccie readers for AI bots a) no one would notice; and b) it would increase the sum of human happiness.
So there is that.
“Imagine if fire, language, internet, and robots happened in the same year. Life would be unrecognizable.
I think we are heading toward that.”
No wonder someone is worried.
I think it's notable that despite having no real AI ability at all, Apple continues to outpace Google who has spent God knows how long working on AI without any real results beyond a slightly better search algorithm and a voice assistant I hate using less.
We are on the brink of nuclear armageddon October 2022
Kwasi Kwartengs budget was great.
“I can tell you for a fact, AI will provide some efficiency gains and that will be the end of it.”
I mean, LOL. I predict this will outdo Rogerdamus’ prediction on seeing the first queues outside Northern Rock in 2007 - “this will all be forgotten about by Friday”
Let us return to your statement in a few years, I don’t think it’s going to age well. To put it politely
Beyond simple code completion, you rapidly run into the problem of it writing code that does the wrong thing.
It can fluently bullshit, so third rate politicians and opinion piece hacks should be worried.
Do you not find it odd that despite throwing so much money at it, Google is currently sinking?
“Let me know if there’s anything I can do”
“You could restore the whip”
“I understand just let me know if there’s anything..”
“Restore the whip”
“I understand”
https://x.com/owenjones84/status/1767907717262320100?s=61