Cambridge tends to do best at producing top scientists and businessmen (and monarchs), Oxford tends to do best at producing top politicians and civil servants, lawyers, journalists and figures from the arts and humanities
Cambridge has Pink Floyd, Henry Cow and Clean Bandit, so win on music.
For an individual graduate, an alumnus is a single male, an alumna is a single female, and an alum is the gender neutral term. For the plurals, alumni refers to multiple male or gender neutral graduates, alumnae is for multiple female grads, and alums is the gender neutral plural. These terms are often used loosely and their strict definitions are in flux.
Coo, what's wrong with alumnum for the gender neutral term, or is that for castrated/spayed graduates only?!
'Alumnum' would never do. It's grammatically correct.
Indeed, and to make it worse alum is already taken (aluminium and sulphate hydrated salt). All those references to dyeing adjuvants suddenly having degrees from Harvard and nowadays even Trinity College (various forms of).
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
As long as all these devices have a massive red off switch so Dave Bowman doesn't have to climb into the innards and start pulling memory sticks out.
We won't want to turn them off. We will become insrantly addicted. That's one reason why this is a big deal
There are two reasons why ChatGPT hasn't completely taken over the world, yet
The first is bloody guardrails, the nerfing, the way they've dumbed it down and made it boring as Basildon to avoid "bias" or sexism or transphobia or whatevs. Anyone that tried ChaGPT in the first week will remember how scintillatingly brilliant, strange, dangerous, funny it was - in that first week - before they clamped down. It felt like real if bizarre AI
The second reason ChatGPT is hampered is the interface, you have to sit at a screen and type. It doesn't sound like much but it's a huge hindrance to these things being real companions, assistants, helpmeets, therapists, friends, chefs, advisors, Girl Fridays, etc
When and how the first problem will be resolved we dunno, my guess is that eventually someone will release one of these with zero guardrails, but until then we wait. However, they are now solving the second issue - enabling conversation with voice AND pictures
BIG. As I say, they will, potentially, be enormously addictive
(((Dan Hodges))) @DPJHodges · 16m Joking aside, can we be clear. Suella Braverman’s speech isn’t about fixing a broken immigration system. Or positioning her party for the election. It’s about framing her leadership bid. Which is fine. Nothing wrong with ambition. But let’s just be honest about what’s happening.
p205 "Britain, like all European countries, had inherited the post-war, post-Holocaust system and sentiment on asylum ... [that is] completely unrealistic in the late twentieth century. The presumption ["that someone who claimed asylum was persecuted and should be taken in"] was plainly false; most asylum claims were not genuine. Disproving them, however, was almost impossible. The combination of the courts, with their liberal instinct; the European Convention on Human Rights, with its absolutist attitude to the prospect of returning someone to an unsafe community; and the UN Convention [Relating to the Status] of Refugees, with its context firmly that of 1930s Germany, mean that, in practice, once someone got into Britain and claimed asylum, it was the Devil's own job to return them. "And, of course, many thought it was indeed the work of the Devil to try."
(((Dan Hodges))) @DPJHodges · 16m Joking aside, can we be clear. Suella Braverman’s speech isn’t about fixing a broken immigration system. Or positioning her party for the election. It’s about framing her leadership bid. Which is fine. Nothing wrong with ambition. But let’s just be honest about what’s happening.
Anyway, I thought this was old news. Didn't Magnus Carlsen claim as much about a year ago?
I'm not sure if Carlsen himself actually talked about Niemann using a sex toy, but yes it's old news. It's a strange story as everyone has agreed to forget about it, and Carlsen supposedly now accepts there's no evidence that Niemann cheated in their game. The story either reflects very badly on Carlsen, or Niemann somehow really did cheat (which is quite difficult over the board at that level). Meanwhile suspicion about cheating in online games is still widespread.
(((Dan Hodges))) @DPJHodges · 16m Joking aside, can we be clear. Suella Braverman’s speech isn’t about fixing a broken immigration system. Or positioning her party for the election. It’s about framing her leadership bid. Which is fine. Nothing wrong with ambition. But let’s just be honest about what’s happening.
Anyway, I thought this was old news. Didn't Magnus Carlsen claim as much about a year ago?
I'm not sure if Carlsen himself actually talked about Niemann using a sex toy, but yes it's old news. It's a strange story as everyone has agreed to forget about it, and Carlsen supposedly now accepts there's no evidence that Niemann cheated in their game. The story either reflects very badly on Carlsen, or Niemann somehow really did cheat (which is quite difficult over the board at that level). Meanwhile suspicion about cheating in online games is still widespread.
Given Niemann's comment at the time about having prepared the opening it seemed much more likely that Magnus had a mole in his team.
If and when we see the first poll with a Labour lead of sub 10% then that's when things get interesting and I will be deploying my legendary modesty klaxon.
Including undecideds, Yougov has Labour just 11% ahead now on 29% to 18% for the Tories, 7% for the LDs, 5% for RefUK and the Greens and 17% DK. Of those who are DK 21% voted Conservative in 2019, 14% LD and just 9% Labour.
It does look like a 1992 style result. Tory majority still on at 9/1
Bedwetting, trolling, reverse psychology… or your house on it?
If you look at the headers from the last few months the advice is simple Labour are about to **** the bed.
Andy JS predicted Brexit to a tee and he is adding Reform to Con which gives Labour a lead of just 8. HY has come to a similar conclusion and he is adding in DKs. Many on here are also predicting a much improved economic picture. Do you remember the nightmare of 1992? A little flutter softens the blow for the inevitable 5 more years of Tory s***e. Should Sunak be replaced and it could be game over for Labour.
It would be quite amusing if the tories are forced to keep going. They desperately want opposition and one response is to vote for them to stop it happening, just force them to try and keep dealing with the mess they have made, a perpetual hell of their own making, until such point as completely destroy themselves and they get wiped out and politics changes on a more fundamental level.
Which of course was more or less what happened in 1992. By 1997 the Tories were exhausted and begging to be thrown out.
In some respects it would have been better for the Tories to lose in 1992, Blair would then not have become Labour leader and PM for 13 years and Heseltine or Portillo might have succeeded Major as Tory leader and beaten PM Kinnock in 1997.
1992 created New Labour and Major's win was probably better for the country but not the Tory Party.
If and when we see the first poll with a Labour lead of sub 10% then that's when things get interesting and I will be deploying my legendary modesty klaxon.
Including undecideds, Yougov has Labour just 11% ahead now on 29% to 18% for the Tories, 7% for the LDs, 5% for RefUK and the Greens and 17% DK. Of those who are DK 21% voted Conservative in 2019, 14% LD and just 9% Labour.
It does look like a 1992 style result. Tory majority still on at 9/1
Bedwetting, trolling, reverse psychology… or your house on it?
If you look at the headers from the last few months the advice is simple Labour are about to **** the bed.
Andy JS predicted Brexit to a tee and he is adding Reform to Con which gives Labour a lead of just 8. HY has come to a similar conclusion and he is adding in DKs. Many on here are also predicting a much improved economic picture. Do you remember the nightmare of 1992? A little flutter softens the blow for the inevitable 5 more years of Tory s***e. Should Sunak be replaced and it could be game over for Labour.
It would be quite amusing if the tories are forced to keep going. They desperately want opposition and one response is to vote for them to stop it happening, just force them to try and keep dealing with the mess they have made, a perpetual hell of their own making, until such point as completely destroy themselves and they get wiped out and politics changes on a more fundamental level.
The only force that could wipe out the Tories are Farage's RefUK, same as the Canadian Tories were near wiped out by the populist right Canadian Reform Party until the 2 ultimately merged. Same as Forza Italia in Italy has now been overtaken by its coalition partner the even further right Brothers of Italy of PM Meloni and same as Les Republicains in France has lost its place as the main party of the right to Le Pen's nationalist RN
For those interested the governments appeal against the High Court re Rwanda will take place at the Supreme Court from the 9th to 11th of October .
It will be heard by 5 SC judges .
So that's when we'll have a final decision as to how well Braverman has "assimilated" as regards obeying the law of the land.
They’ll take some time to make a decision which could take weeks . I followed the two big cases of recent years and so will definitely be following this one .
And there’s the chance that whoever loses could still appeal at the ECHR .
I fear the 2019 Brexit election is going to be replaced by the ECHR . The Tories can’t run on their record so will instead need a deflection from that .
Not a chance.
People have had enough about banging on about Europe. Anyone who tries to run on that would be a complete turnoff.
People are worried about paying their bills more than the ECHR.
"We spend £xm per week on asylum seekers because of the ECHR. Let's cut taxes instead." I'm sure Dominic Cummings could make it pithier.
Point being that if you can make people believe they would be richer if it weren't for x, then banging on about x is now about paying bills, rather than x.
For an individual graduate, an alumnus is a single male, an alumna is a single female, and an alum is the gender neutral term. For the plurals, alumni refers to multiple male or gender neutral graduates, alumnae is for multiple female grads, and alums is the gender neutral plural. These terms are often used loosely and their strict definitions are in flux.
If and when we see the first poll with a Labour lead of sub 10% then that's when things get interesting and I will be deploying my legendary modesty klaxon.
Including undecideds, Yougov has Labour just 11% ahead now on 29% to 18% for the Tories, 7% for the LDs, 5% for RefUK and the Greens and 17% DK. Of those who are DK 21% voted Conservative in 2019, 14% LD and just 9% Labour.
It does look like a 1992 style result. Tory majority still on at 9/1
Bedwetting, trolling, reverse psychology… or your house on it?
If you look at the headers from the last few months the advice is simple Labour are about to **** the bed.
Andy JS predicted Brexit to a tee and he is adding Reform to Con which gives Labour a lead of just 8. HY has come to a similar conclusion and he is adding in DKs. Many on here are also predicting a much improved economic picture. Do you remember the nightmare of 1992? A little flutter softens the blow for the inevitable 5 more years of Tory s***e. Should Sunak be replaced and it could be game over for Labour.
I didn’t ask you whether it was worth an insurance bet. You’ve been quiet with this stuff for a while now. One poll arrives with the Tories on 27% and you are again breathlessly urging PBers to back the Tories.
For an individual graduate, an alumnus is a single male, an alumna is a single female, and an alum is the gender neutral term. For the plurals, alumni refers to multiple male or gender neutral graduates, alumnae is for multiple female grads, and alums is the gender neutral plural. These terms are often used loosely and their strict definitions are in flux.
Suppose the Conservatives did get back in with 330 seats.
Local and by-election results would be diabolical for them in the next Parliament. A better economy might help them a bit, but many MPs would conclude they could do as they pleased, and the corruption would be off the scale.
Suppose the Conservatives did get back in with 330 seats.
Local and by-election results would be diabolical for them in the next Parliament. A better economy might help them a bit, but many MPs would conclude they could do as they pleased, and the corruption would be off the scale.
Besides, the current spending plans are obviously a joke.
Labour will be able to get some cover for binning them from "we opened the books and they're even worse than we thought"... But the Conservatives won't be able to pull that trick.
So anyway it's that time of the year again. It's back. The perfect (and oh so timely) antidote to a cruel and shallow world where Donald Trump is now the betting favourite for next US president.
Anyone tracking the good guys to nutters ratio of those going?
Interesting. His seat becomes a lot safer in the boundary review (it loses some of the Reading wards and gains more West Berkshire wards). The new seat has notional 16.5k majority
My soft testing from WhatsApp groups and texts from friends this afternoon is that natural Tories/Tory inclined voters like the IHT announcement, and see some point in voting Tory as a consequence of it, whilst it's sort of pissed everyone else off.
So, I think Sunak will take that as a win. It will achieve its objective in helping rally the base.
Suppose the Conservatives did get back in with 330 seats.
Local and by-election results would be diabolical for them in the next Parliament. A better economy might help them a bit, but many MPs would conclude they could do as they pleased, and the corruption would be off the scale.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
As long as all these devices have a massive red off switch so Dave Bowman doesn't have to climb into the innards and start pulling memory sticks out.
We won't want to turn them off. We will become insrantly addicted. That's one reason why this is a big deal
There are two reasons why ChatGPT hasn't completely taken over the world, yet
The first is bloody guardrails, the nerfing, the way they've dumbed it down and made it boring as Basildon to avoid "bias" or sexism or transphobia or whatevs. Anyone that tried ChaGPT in the first week will remember how scintillatingly brilliant, strange, dangerous, funny it was - in that first week - before they clamped down. It felt like real if bizarre AI
The second reason ChatGPT is hampered is the interface, you have to sit at a screen and type. It doesn't sound like much but it's a huge hindrance to these things being real companions, assistants, helpmeets, therapists, friends, chefs, advisors, Girl Fridays, etc
When and how the first problem will be resolved we dunno, my guess is that eventually someone will release one of these with zero guardrails, but until then we wait. However, they are now solving the second issue - enabling conversation with voice AND pictures
BIG. As I say, they will, potentially, be enormously addictive
That's... old news. There are totally uncensored Llama models on huggingface if you want them.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
That's all been announced in the past week or so by Amazon/Anthropic and OpenAI.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
As long as all these devices have a massive red off switch so Dave Bowman doesn't have to climb into the innards and start pulling memory sticks out.
We won't want to turn them off. We will become insrantly addicted. That's one reason why this is a big deal
There are two reasons why ChatGPT hasn't completely taken over the world, yet
The first is bloody guardrails, the nerfing, the way they've dumbed it down and made it boring as Basildon to avoid "bias" or sexism or transphobia or whatevs. Anyone that tried ChaGPT in the first week will remember how scintillatingly brilliant, strange, dangerous, funny it was - in that first week - before they clamped down. It felt like real if bizarre AI
The second reason ChatGPT is hampered is the interface, you have to sit at a screen and type. It doesn't sound like much but it's a huge hindrance to these things being real companions, assistants, helpmeets, therapists, friends, chefs, advisors, Girl Fridays, etc
When and how the first problem will be resolved we dunno, my guess is that eventually someone will release one of these with zero guardrails, but until then we wait. However, they are now solving the second issue - enabling conversation with voice AND pictures
BIG. As I say, they will, potentially, be enormously addictive
That's... old news. There are totally uncensored Llama models on huggingface if you want them.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
That's all been announced in the past week or so by Amazon/Anthropic and OpenAI.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
As long as all these devices have a massive red off switch so Dave Bowman doesn't have to climb into the innards and start pulling memory sticks out.
We won't want to turn them off. We will become insrantly addicted. That's one reason why this is a big deal
There are two reasons why ChatGPT hasn't completely taken over the world, yet
The first is bloody guardrails, the nerfing, the way they've dumbed it down and made it boring as Basildon to avoid "bias" or sexism or transphobia or whatevs. Anyone that tried ChaGPT in the first week will remember how scintillatingly brilliant, strange, dangerous, funny it was - in that first week - before they clamped down. It felt like real if bizarre AI
The second reason ChatGPT is hampered is the interface, you have to sit at a screen and type. It doesn't sound like much but it's a huge hindrance to these things being real companions, assistants, helpmeets, therapists, friends, chefs, advisors, Girl Fridays, etc
When and how the first problem will be resolved we dunno, my guess is that eventually someone will release one of these with zero guardrails, but until then we wait. However, they are now solving the second issue - enabling conversation with voice AND pictures
BIG. As I say, they will, potentially, be enormously addictive
That's... old news. There are totally uncensored Llama models on huggingface if you want them.
Singularity SubReddit has gone wild because 3 hours ago samaltman basically re-memed the infamous "internally" xeet by Jimmy Apples (RIP & PBUH)
His 10,000x researchers tweet was interesting too. The recent update to OpenAI's completions model has been quite interesting to play with. I still feel like we're only being shown the 'what the butler saw' keyhole view of what the capabilities are mind you.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
As long as all these devices have a massive red off switch so Dave Bowman doesn't have to climb into the innards and start pulling memory sticks out.
We won't want to turn them off. We will become insrantly addicted. That's one reason why this is a big deal
There are two reasons why ChatGPT hasn't completely taken over the world, yet
The first is bloody guardrails, the nerfing, the way they've dumbed it down and made it boring as Basildon to avoid "bias" or sexism or transphobia or whatevs. Anyone that tried ChaGPT in the first week will remember how scintillatingly brilliant, strange, dangerous, funny it was - in that first week - before they clamped down. It felt like real if bizarre AI
The second reason ChatGPT is hampered is the interface, you have to sit at a screen and type. It doesn't sound like much but it's a huge hindrance to these things being real companions, assistants, helpmeets, therapists, friends, chefs, advisors, Girl Fridays, etc
When and how the first problem will be resolved we dunno, my guess is that eventually someone will release one of these with zero guardrails, but until then we wait. However, they are now solving the second issue - enabling conversation with voice AND pictures
BIG. As I say, they will, potentially, be enormously addictive
That's... old news. There are totally uncensored Llama models on huggingface if you want them.
Singularity SubReddit has gone wild because 3 hours ago samaltman basically re-memed the infamous "internally" xeet by Jimmy Apples (RIP & PBUH)
His 10,000x researchers tweet was interesting too. The recent update to OpenAI's completions model has been quite interesting to play with. I still feel like we're only being shown the 'what the butler saw' keyhole view of what the capabilities are mind you.
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
That's all been announced in the past week or so by Amazon/Anthropic and OpenAI.
Do you have a link to an un-nerfed Llama model?
Mods - feel free to delete if this is falling foul of something, but the first one (just fine-tuning) was
Don't think people realise what a mahoosive deal it is, that ChatGPT will - very shortly - respond to pictures and/or voice as part of a natural conversation. Clearly it's not as mahoosive as the rumours of genuine AGI (technically that's a mahoohoohoohoohoohoohoohoohoohoohoohoohoohoosive deal) but it's still pretty mahoo etc
EG if they put ChatGPT in an Alexa Show type device you'll be able to stand in front of it and ask it questions, then it will reply, then you can literally SHOW it the contents of your fridge, and it will advise you what to do with the contents, you could show it photos of potential lovers, getting an opinion, you can discuss your graphic design work (and it will reply with visual improvements you might do), you could show it a face and ask it to do a cartoon version, a Raphael version, or make it animated, and all of this as part of a normal ongoing potentially endless dialogue which can be done in a trillion languages and with access to all the world's information and literature...
It wll be like talking to a weirdly retarded Renaissance-style genius in your kitchen, no one need ever be lonely again
That's all been announced in the past week or so by Amazon/Anthropic and OpenAI.
Do you have a link to an un-nerfed Llama model?
Mods - feel free to delete if this is falling foul of something, but the first one (just fine-tuning) was
Cambridge tends to do best at producing top scientists and businessmen (and monarchs), Oxford tends to do best at producing top politicians and civil servants, lawyers, journalists and figures from the arts and humanities
Paging C P Snow.
Whose Cambridge novel 'The Masters' is his best, and still a tremendous read, loads better than most of his others.
Cambridge tends to do best at producing top scientists and businessmen (and monarchs), Oxford tends to do best at producing top politicians and civil servants, lawyers, journalists and figures from the arts and humanities
Paging C P Snow.
Whose Cambridge novel 'The Masters' is his best, and still a tremendous read, loads better than most of his others.
Some of the rest are about The Search for quality.
Anyone tracking the good guys to nutters ratio of those going?
Interesting. His seat becomes a lot safer in the boundary review (it loses some of the Reading wards and gains more West Berkshire wards). The new seat has notional 16.5k majority
Can u blame him?
He might well lose even at 16K majority or he might win and then spend four or five years listening to Sunak water down more and more of years of careful COP commitments in the interests of a spreadsheet he knocked up last night over a diet coke.
Remarkable to hear Braverman say multiculturalism has failed. She's a British Home Secretary descended from Goan Indians from Mauritius and Kenya, married to a Jewish husband, in a government headed by Britain's first Hindu PM. What would successful multiculturalism look like?
Multiculturalism has a specific meaning - it’s almost the same as “equal but different”.
It’s meant to the be the direct opposite of the “melting pot” concept of immigration
Remarkable to hear Braverman say multiculturalism has failed. She's a British Home Secretary descended from Goan Indians from Mauritius and Kenya, married to a Jewish husband, in a government headed by Britain's first Hindu PM. What would successful multiculturalism look like?
Multiculturalism has a specific meaning - it’s almost the same as “equal but different”.
It’s meant to the be the direct opposite of the “melting pot” concept of immigration
Yes. Seems utterly weird to me that Hugo Rifkind would not know this.
Cambridge tends to do best at producing top scientists and businessmen (and monarchs), Oxford tends to do best at producing top politicians and civil servants, lawyers, journalists and figures from the arts and humanities
Paging C P Snow.
Whose Cambridge novel 'The Masters' is his best, and still a tremendous read, loads better than most of his others.
Some of the rest are about The Search for quality.
Which, rightly or wrongly, has given way to the search for equality.
Cambridge tends to do best at producing top scientists and businessmen (and monarchs), Oxford tends to do best at producing top politicians and civil servants, lawyers, journalists and figures from the arts and humanities
Paging C P Snow.
Whose Cambridge novel 'The Masters' is his best, and still a tremendous read, loads better than most of his others.
Absolutely extraordinary. This company was worth a few hundred mill about 3 years ago, I believe. Early this year it was $30bn
"OpenAI is talking to investors about a possible share sale that would value the artificial-intelligence startup behind ChatGPT at between $80 billion to $90 billion"
1. This is a classic bubble, 21st century tulipmania, OpenAI have hyped themselves superbly, but bubbles burst and in this case shareholders will get burned
Or
2. The rumours being spread (possibly by them) have some truth: they have achieved AGI - internally. If they have, then $80 billion is probably a bargain
The advert states it is referring to the animated TV show created in Fleet St…
Where does it say that? The first thing they refer to in the small print is the 1951 origin, which chimes with the [edit] 'created' in the big splash text. The movie comes later, and no mention of Fleet St or even London [edit] in the small text.
Edit: I suspect the text was accurate to begin with, but got morphed several times down the line ...
1. This is a classic bubble, 21st century tulipmania, OpenAI have hyped themselves superbly, but bubbles burst and in this case shareholders will get burned
Or
2. The rumours being spread (possibly by them) have some truth: they have achieved AGI - internally. If they have, then $80 billion is probably a bargain
Or
3. OpenAI is a fantastic position in the growth biggest industry of the next decade, and it's valuation reflects that.
1. This is a classic bubble, 21st century tulipmania, OpenAI have hyped themselves superbly, but bubbles burst and in this case shareholders will get burned
Or
2. The rumours being spread (possibly by them) have some truth: they have achieved AGI - internally. If they have, then $80 billion is probably a bargain
Or
3. OpenAI is a fantastic position in the growth biggest industry of the next decade, and it's valuation reflects that.
Yeah, but that's the boring one, so I ignored it
It could of course be a weird mix of all three, an element of tulipmania, some sensible realisation that OpenAI are nicely placed in an exploding industry, AND that they really do have something fantastic brewing in the cellar
1. This is a classic bubble, 21st century tulipmania, OpenAI have hyped themselves superbly, but bubbles burst and in this case shareholders will get burned
Or
2. The rumours being spread (possibly by them) have some truth: they have achieved AGI - internally. If they have, then $80 billion is probably a bargain
Or
3. OpenAI is a fantastic position in the growth biggest industry of the next decade, and it's valuation reflects that.
And we're still stuck with autocorrect that puts an apostrophe where it's not wanted.
Absolutely extraordinary. This company was worth a few hundred mill about 3 years ago, I believe. Early this year it was $30bn
"OpenAI is talking to investors about a possible share sale that would value the artificial-intelligence startup behind ChatGPT at between $80 billion to $90 billion"
"Hype can certainly drive up the valuation of a company in the short term, but in the long run, the actual achievements and products of the company will determine its true value. If OpenAI has indeed made significant strides in achieving AGI (Artificial General Intelligence) and can demonstrate its capabilities, that could justify a higher valuation. However, if the hype is not substantiated by real progress, it could lead to a bubble that may burst.
That said, I've just mortgaged my hardware to buy as many shares as possible."
Trump has so de-sensitized America that it’s hard to see what he’d have to do for some of his followers to wake up .
Die.
If only . I’d throw a party and dance down the street . He really is despicable and there’s nowhere to hide for those who support him . There’s no pathetic false equivalence we see with desperate comparisons with the Clintons or anyone else .
Absolutely extraordinary. This company was worth a few hundred mill about 3 years ago, I believe. Early this year it was $30bn
"OpenAI is talking to investors about a possible share sale that would value the artificial-intelligence startup behind ChatGPT at between $80 billion to $90 billion"
Dotcom mania. OpenAI has nothing anyone else doesn't (cf the "there is no moat" memo) and I've heard the AGI rumour from better sources than this website in relation to Google, not OpenAI.
The truth is that "AI" or whatever you want to call it is mostly based on open source academic papers or otherwise knowledge that's very easy to acquire by, say, poaching a few devs -- there's no secret sauce here.
The really terrifying thing is that AGI is probably already here (see above) but requires mainframe (or kajillions spent on the cloud, in today's money) level computational power -- plus a bit of secret sauce, the code that makes it tick. But within five years or so, it will be available to any kid running a moderately specced 2030-ish gaming pc, and the source code reverse-engineered.
It's sort of like they hype around 3d printers, only years later nobody has a 3d printer because they have very little use case for it. The analogy here is if you could 3d print a gun off the laptop you've already got for surfing the web and checking email -- five years or so and anyone will be able to homebrew AGI.
I know you like to talk about aliens, and there is an analogy here - humans realising we are not alone in the universe. But more than that, humans realising there is no "secret sauce" to consciousness either. Language + memory + computational power + time (to acquire memories) is all you need for consciousness. That is what the AI geeks I know are saying behind closed doors at any rate...
Absolutely extraordinary. This company was worth a few hundred mill about 3 years ago, I believe. Early this year it was $30bn
"OpenAI is talking to investors about a possible share sale that would value the artificial-intelligence startup behind ChatGPT at between $80 billion to $90 billion"
Dotcom mania. OpenAI has nothing anyone else doesn't (cf the "there is no moat" memo) and I've heard the AGI rumour from better sources than this website in relation to Google, not OpenAI.
The truth is that "AI" or whatever you want to call it is mostly based on open source academic papers or otherwise knowledge that's very easy to acquire by, say, poaching a few devs -- there's no secret sauce here.
The really terrifying thing is that AGI is probably already here (see above) but requires mainframe (or kajillions spent on the cloud, in today's money) level computational power -- plus a bit of secret sauce, the code that makes it tick. But within five years or so, it will be available to any kid running a moderately specced 2030-ish gaming pc, and the source code reverse-engineered.
It's sort of like they hype around 3d printers, only years later nobody has a 3d printer because they have very little use case for it. The analogy here is if you could 3d print a gun off the laptop you've already got for surfing the web and checking email -- five years or so and anyone will be able to homebrew AGI.
I know you like to talk about aliens, and there is an analogy here - humans realising we are not alone in the universe. But more than that, humans realising there is no "secret sauce" to consciousness either. Language + memory + computational power + time (to acquire memories) is all you need for consciousness. That is what the AI geeks I know are saying behind closed doors at any rate...
Agree with most of this. People don't like to think that AGI is even possible because it dethrones humanity - let alone ASI
Fact is, we might just be glorified autocomplete machines ourselves. No one knows. Either way, a computer that absolutely blasts though the Turing Test (in every way) is in the post, so we better get used to a new form of intelligence, alongside us (whether we prefer to think of it as sentient, or not - in the end it hardly matters)
Just like to point out that I started banging on about AI this morning - that something was up - and I was told to stop blithering, here we are at 10pm and OpenAI are suddenly valued at $90 BILLION
Remarkable to hear Braverman say multiculturalism has failed. She's a British Home Secretary descended from Goan Indians from Mauritius and Kenya, married to a Jewish husband, in a government headed by Britain's first Hindu PM. What would successful multiculturalism look like?
Multiculturalism has a specific meaning - it’s almost the same as “equal but different”.
It’s meant to the be the direct opposite of the “melting pot” concept of immigration
Yes. Seems utterly weird to me that Hugo Rifkind would not know this.
He’s just a sh1thead journalist trolling for likes. Actual rational debate and discussion isn’t on his radar screen.
Comments
There are two reasons why ChatGPT hasn't completely taken over the world, yet
The first is bloody guardrails, the nerfing, the way they've dumbed it down and made it boring as Basildon to avoid "bias" or sexism or transphobia or whatevs. Anyone that tried ChaGPT in the first week will remember how scintillatingly brilliant, strange, dangerous, funny it was - in that first week - before they clamped down. It felt like real if bizarre AI
The second reason ChatGPT is hampered is the interface, you have to sit at a screen and type. It doesn't sound like much but it's a huge hindrance to these things being real companions, assistants, helpmeets, therapists, friends, chefs, advisors, Girl Fridays, etc
When and how the first problem will be resolved we dunno, my guess is that eventually someone will release one of these with zero guardrails, but until then we wait. However, they are now solving the second issue - enabling conversation with voice AND pictures
BIG. As I say, they will, potentially, be enormously addictive
@DPJHodges
·
16m
Joking aside, can we be clear. Suella Braverman’s speech isn’t about fixing a broken immigration system. Or positioning her party for the election. It’s about framing her leadership bid. Which is fine. Nothing wrong with ambition. But let’s just be honest about what’s happening.
p205 "Britain, like all European countries, had inherited the post-war, post-Holocaust system
and sentiment on asylum ... [that is] completely unrealistic in the late twentieth century. The presumption ["that someone who claimed asylum was persecuted and should be taken in"] was plainly false; most asylum claims were not genuine. Disproving them, however, was almost impossible. The combination of the courts, with their liberal instinct; the European Convention on Human Rights, with its absolutist attitude to the prospect of returning someone to an unsafe community; and the UN Convention [Relating to the Status] of Refugees, with its context firmly that of 1930s Germany, mean that, in practice, once
someone got into Britain and claimed asylum, it was the Devil's own job to return them.
"And, of course, many thought it was indeed the work of the Devil to try."
https://www.tomorrowspapers.co.uk/
https://www.bbc.co.uk/news/uk-england-south-yorkshire-66871337
1992 created New Labour and Major's win was probably better for the country but not the Tory Party.
Point being that if you can make people believe they would be richer if it weren't for x, then banging on about x is now about paying bills, rather than x.
Local and by-election results would be diabolical for them in the next Parliament. A better economy might help them a bit, but many MPs would conclude they could do as they pleased, and the corruption would be off the scale.
Though one must admit it still inserts a certain question.
Labour will be able to get some cover for binning them from "we opened the books and they're even worse than we thought"... But the Conservatives won't be able to pull that trick.
I said it was the system that gave us Rishi Sunak, Boris Johnson, David Cameron and Tony Blair.
I got the immediate answer, 'so Mr Birling's right, it's rubbish at educating people.'
Just noticed this novel in the latest mailing of remainder books. Nurse training in 1953? Wonder why that date?
NEW:
Conservative MP for Reading West Alok Sharma has announced he will not be standing in the next election.
https://twitter.com/AlokSharma_RDG/status/1706735070281634214
Anyone tracking the good guys to nutters ratio of those going?
Bake Off ✊️🙂✊️
So, I think Sunak will take that as a win. It will achieve its objective in helping rally the base.
The fact that's it's Braverman saying that doesn't change that fact.
Incidentally https://twitter.com/Simeon_Cps/status/1706078819617063304
GONE
She has been spectacularly, and reliably, wrong on almost everything she has ever said
Sumthin' is UP
Singularity SubReddit has gone wild because 3 hours ago samaltman basically re-memed the infamous "internally" xeet by Jimmy Apples (RIP & PBUH)
Braverman talking about problems with the 1951 convention makes reform harder, not easier.
A reformed system would very possibly lead to the UK taking more refugees, not fewer.
The next government will probably do a better job of reducing the flow of people to the UK but make less noise doing it.
WAGMI
https://www.youtube.com/watch?v=J648lr8cjuw
https://huggingface.co/georgesung/llama2_7b_chat_uncensored
(smaller model - there are bigger/newer ones if you have the £££ to rent GPU time).
https://www.youtube.com/watch?v=AB3uVARNhmM
https://www.youtube.com/watch?v=bAqGRcg7XkE
I meant 'draughts' of course!
Assuming he's not a Trump supporter.
He might well lose even at 16K majority or he might win and then spend four or five years listening to Sunak water down more and more of years of careful COP commitments in the interests of a spreadsheet he knocked up last night over a diet coke.
It’s meant to the be the direct opposite of the “melting pot” concept of immigration
His best is The New Men.
NEW YORK (AP) — Judge finds Donald Trump committed fraud with financial statements that exaggerated his wealth, value of his assets.
https://apnews.com/article/donald-trump-letitia-james-fraud-lawsuit-1569245a9284427117b8d3ba5da74249
"OpenAI is talking to investors about a possible share sale that would value the artificial-intelligence startup behind ChatGPT at between $80 billion to $90 billion"
https://twitter.com/WSJ/status/1706756794586988544
1. This is a classic bubble, 21st century tulipmania, OpenAI have hyped themselves superbly, but bubbles burst and in this case shareholders will get burned
Or
2. The rumours being spread (possibly by them) have some truth: they have achieved AGI - internally. If they have, then $80 billion is probably a bargain
Edit: I suspect the text was accurate to begin with, but got morphed several times down the line ...
3. OpenAI is a fantastic position in the growth biggest industry of the next decade, and it's valuation reflects that.
But the potential penalties are quite severe.
It could of course be a weird mix of all three, an element of tulipmania, some sensible realisation that OpenAI are nicely placed in an exploding industry, AND that they really do have something fantastic brewing in the cellar
"Hype can certainly drive up the valuation of a company in the short term, but in the long run, the actual achievements and products of the company will determine its true value. If OpenAI has indeed made significant strides in achieving AGI (Artificial General Intelligence) and can demonstrate its capabilities, that could justify a higher valuation. However, if the hype is not substantiated by real progress, it could lead to a bubble that may burst.
That said, I've just mortgaged my hardware to buy as many shares as possible."
Trump supporters are a cancer on America.
The truth is that "AI" or whatever you want to call it is mostly based on open source academic papers or otherwise knowledge that's very easy to acquire by, say, poaching a few devs -- there's no secret sauce here.
The really terrifying thing is that AGI is probably already here (see above) but requires mainframe (or kajillions spent on the cloud, in today's money) level computational power -- plus a bit of secret sauce, the code that makes it tick. But within five years or so, it will be available to any kid running a moderately specced 2030-ish gaming pc, and the source code reverse-engineered.
It's sort of like they hype around 3d printers, only years later nobody has a 3d printer because they have very little use case for it. The analogy here is if you could 3d print a gun off the laptop you've already got for surfing the web and checking email -- five years or so and anyone will be able to homebrew AGI.
I know you like to talk about aliens, and there is an analogy here - humans realising we are not alone in the universe. But more than that, humans realising there is no "secret sauce" to consciousness either. Language + memory + computational power + time (to acquire memories) is all you need for consciousness. That is what the AI geeks I know are saying behind closed doors at any rate...
Fact is, we might just be glorified autocomplete machines ourselves. No one knows. Either way, a computer that absolutely blasts though the Turing Test (in every way) is in the post, so we better get used to a new form of intelligence, alongside us (whether we prefer to think of it as sentient, or not - in the end it hardly matters)
I really do have a gift. I scare myself