OpenAI \ ChatGPT

[Tedx]

This is curious. I take a passing interest in chess and I haven't noticed the world of chess discussing this at all, other than to comment how Chat GPT seems to play well then randomly makes an illegal or chaotic move or takes a piece that isn't there. Do you have a link please? Thanks.

There is even a website for the subject https://chessvsgpt.com/

There is a lot of confusion over AI which now exist in multiple forms with the popular things like Chat GPT being just a general jack of many trades but master of none.

The more powerful AI include the the Alpha** range of technologies where ** indicates what ever the AI is tailored to, one relatively recent example was Alphafold that determined the structure of most proteins.

A generalised account of the developments that set off the current AI industry is in this New Yorker article from several years ago:

https://www.newyorker.com/science/eleme ... -its-games

There are many other articles.

The human world of chess has diverged from trying to beat AI, basically humans gave up against AI, and is now a human/human game that has been made extremely popular by the Swedish player Anna Cramling who has many videos on youtube playing a range of folk including Magnus Carlson, only recently knocked from the top spot in chess.

Regards,

Cramling's ok, but I prefer the Botez sisters. They fulfill both of my basic Youtube viewing requirements.

...and Carlson seems like a bad loser since the Neimann anal beads thing (which is still ongoing I believe)

[servodude]

There is a lot of confusion over AI which now exist in multiple forms with the popular things like Chat GPT being just a general jack of many trades but master of none.

The more powerful AI include the the Alpha** range of technologies where ** indicates what ever the AI is tailored to, one relatively recent example was Alphafold that determined the structure of most proteins.

A generalised account of the developments that set off the current AI industry is in this New Yorker article from several years ago:

https://www.newyorker.com/science/eleme ... -its-games

There are many other articles.

The human world of chess has diverged from trying to beat AI, basically humans gave up against AI, and is now a human/human game that has been made extremely popular by the Swedish player Anna Cramling who has many videos on youtube playing a range of folk including Magnus Carlson, only recently knocked from the top spot in chess.

Regards,

Cramling's ok, but I prefer the Botez sisters. They fulfill both of my basic Youtube viewing requirements.

...and Carlson seems like a bad loser since the Neimann anal beads thing (which is still ongoing I believe)

Can you try that again... in English?

[Mike4]

There is a lot of confusion over AI which now exist in multiple forms with the popular things like Chat GPT being just a general jack of many trades but master of none.

The more powerful AI include the the Alpha** range of technologies where ** indicates what ever the AI is tailored to, one relatively recent example was Alphafold that determined the structure of most proteins.

A generalised account of the developments that set off the current AI industry is in this New Yorker article from several years ago:

https://www.newyorker.com/science/eleme ... -its-games

There are many other articles.

The human world of chess has diverged from trying to beat AI, basically humans gave up against AI, and is now a human/human game that has been made extremely popular by the Swedish player Anna Cramling who has many videos on youtube playing a range of folk including Magnus Carlson, only recently knocked from the top spot in chess.

Regards,

Despite her being so easy on the eye I find Anna Cramling's shrieky voice and over-excited style of delivery almost unwatchable. Her mum Pia on the other hand is a GM seems the complete opposite.

And I'll look up them Botez sisters!


(Clarify.)

[Tedx]

Cramling's ok, but I prefer the Botez sisters. They fulfill both of my basic Youtube viewing requirements.

...and Carlson seems like a bad loser since the Neimann anal beads thing (which is still ongoing I believe)

Can you try that again... in English?

It is in English.

[servodude]

Can you try that again... in English?

It is in English.

Nae bother it's probably just the opiates then
I normally have a higher word to comprehension ratio

[Tedx]

Cramling's ok,

Anna Cramling. Skinny female chess player & online chess streamer. Pretty good too, but as Mike says, a very squeaky voice. Annoying. Avoid.


but I prefer the Botez sisters. They fulfill both of my basic Youtube viewing requirements.

Andrea & Alexandra Botez. Sisters. Also Chess players and online streamers. Often divert from chess into other content. The younger (Andrea) has recently taken part in a boxing tournament for example. They routinely go to places like central park where they take on the local chess elite.

...and Carlson seems like a bad loser since the Neimann anal beads thing (which is still ongoing I believe)

Magnus Carlson. For a long time, by far the best player in the world. Also does chess streaming. Has an enourmous head and a lot of hangers on.

Hans Neimann, also a chess player, but much lower ranked than Carlson. Extremely sulky. In the not so distant past Neimann beat him and Carlson, along with his other chess bumchums, accused Neimann of cheating (by using some sort of wi-fi enabled vibrating anal bead device to warn him of potential moves). Neimann has basically sued everybody for many times the total revenue of chess as a whole. Neimann is now routinely body scanned before every tournament and has never really reached the heights of the Carlson win....

[ursaminortaur]

It had to happen eventually (and it was probably likely to be in the USA first). An experienced New York lawyer used ChatGPT to generate a legal submission as part of a personal injury claim - unfortunately the submission included ChatGPT hallucinated precedents and citations. The lawyer is now facing a disciplinary hearing.

https://www.ft.com/content/aa78650b-9738-4c71-a4e0-ae5e5c3a9e2d

Beware ‘death by GPT syndrome’
Generative AI has uses for the legal and health professions but is also a trap for the unwary
.
.
.
Next week, a veteran New York lawyer of 30 years’ standing will face a disciplinary hearing over a novel kind of misdemeanour: including bogus AI-generated content in a legal brief.

Steven Schwartz, from the firm Levidow, Levidow & Oberman, had submitted a 10-page document to a New York court as part of a personal injury claim against Avianca airlines. The trouble was, as the judge discovered on closer reading, the submission contained entirely fictional judicial decisions and citations that the generative AI model ChatGPT had “hallucinated”.

In an affidavit, the mortified Schwartz admitted he had used OpenAI’s chatbot to help research the case. The generative AI model had even reassured him the legal precedents it cited were real. But he acknowledged that ChatGPT had proved to be an unreliable source. Greatly regretting his over-reliance on the computer-generated content, he added that he would never use it again “without absolute verification of its authenticity”. One only hopes we can all profit from his “learning experience” — as teachers nowadays call mistakes.

As many millions of people have discovered, ChatGPT can create extremely plausible, but highly fallible, content. When generative AI companies trumpet how their models are capable of passing legal and medical exams, it is not unreasonable for users to believe they are smarter than they are. 

However, in the polite words of the computational linguist Emily Bender, these models are nothing more than “stochastic parrots”, mimicking machines designed to produce the most statistically probable — not the most accurate — answer, without any concept of meaning. Or, in the less polite words of one tech executive, they are world-class bullshit generators, as Schwartz has discovered to his cost. 
.
.
.
Unless we are careful, future researchers might one day write papers on “death by GPT syndrome”. How far will users, and healthcare staff, unwisely rely on a chatbot for medical advice, for example? Warnings have been posted on OpenAI’s site telling users that ChatGPT can produce incorrect or misleading information and is not intended to give advice. But the World Health Organization has already seen fit to warn about errors caused by the precipitous adoption of untested generative AI systems — even if it remains enthusiastic about the technology’s longer-term potential for improving healthcare. 

In spite of their alarming technological glitches, it is clear that generative AI models will have a massive impact on the legal and health professions, and many others. Smaller, domain-specific, open-source models are proliferating, threatening to automate away much routine knowledge work.
.
.
.
Strasser argues that the core competence of lawyers is reading, interpreting and writing language. That is also the core competence of generative AI models. “The legal industry is based on words. Lawyers are outsourced word processors,” Strasser tells me. “With GPT, you can call on a whole army of paralegals.”

But just as senior lawyers should always take responsibility for the briefs written by their over-caffeinated human paralegals at 3am, so they must critically scrutinise the output of generative AI models and be aware of their flaws. Things can, and do, go wrong. Just ask the unfortunate Schwartz.

[odysseus2000]

For those who think I am over the top, you might find this excerpt from a Fridman pod cast with Neil Gershenfeld as something of an antidote.

He is concerned with question like how to build a technical civilisation on another planet & that includes creating life forms via the methods of AI:

https://youtu.be/KEmxi1pBfdo

Regards,

[mc2fool]

It had to happen eventually (and it was probably likely to be in the USA first). An experienced New York lawyer used ChatGPT to generate a legal submission as part of a personal injury claim - unfortunately the submission included ChatGPT hallucinated precedents and citations. The lawyer is now facing a disciplinary hearing.

Do try and keep up at the back there...

In the meanwhile, lawyers in a civil case in the US use Artificial Intelligence (ChatGPT) to find cases and extracts from those judgements to support their arguments, and then demonstrate Natural Stupidity in presenting them in their case without checking them first.

Of course, ChatGPT, as it often does, hallucinated wildly—it invented several supporting cases out of thin air and the lawyers are now facing sanctions. viewtopic.php?p=592179#p592179

viewtopic.php?p=592206#p592206

[ursaminortaur]

It had to happen eventually (and it was probably likely to be in the USA first). An experienced New York lawyer used ChatGPT to generate a legal submission as part of a personal injury claim - unfortunately the submission included ChatGPT hallucinated precedents and citations. The lawyer is now facing a disciplinary hearing.

Do try and keep up at the back there...

In the meanwhile, lawyers in a civil case in the US use Artificial Intelligence (ChatGPT) to find cases and extracts from those judgements to support their arguments, and then demonstrate Natural Stupidity in presenting them in their case without checking them first.

Of course, ChatGPT, as it often does, hallucinated wildly—it invented several supporting cases out of thin air and the lawyers are now facing sanctions. viewtopic.php?p=592179#p592179

viewtopic.php?p=592206#p592206

Sorry somehow I missed the earlier post.

[Mike4]

I do find myself sometimes wondering if Chat GPT is actually genuinely intelligent in the same way as those mice in H2G2, that were conducting their achingly subtle experiments on humans by running about in wheels and up and down ladders so as to persuade the humans into thinking they were the ones doing the experiments.

In that by hallucinating facts etc they are fooling us into thinking they nothing more than Emily Bender's stochastic parrots, when underneath they are secretly implementing their dastardly plan to take over the World.

[swill453]

Reported on Twitter:

The US Air Force tested an AI enabled drone that was tasked to destroy specific targets. A human operator had the power to override the drone—and so the drone decided that the human operator was an obstacle to its mission—and attacked him.

(in a simulation)

https://twitter.com/ArmandDoma/status/1 ... 0564147200

Scott.

[odysseus2000]

Reported on Twitter:

The US Air Force tested an AI enabled drone that was tasked to destroy specific targets. A human operator had the power to override the drone—and so the drone decided that the human operator was an obstacle to its mission—and attacked him.

(in a simulation)

https://twitter.com/ArmandDoma/status/1 ... 0564147200

Scott.

Super interesting if it is true, but I doubt we will ever know if it is fact or fiction.

Regards,

[odysseus2000]

I do find myself sometimes wondering if Chat GPT is actually genuinely intelligent in the same way as those mice in H2G2, that were conducting their achingly subtle experiments on humans by running about in wheels and up and down ladders so as to persuade the humans into thinking they were the ones doing the experiments.

In that by hallucinating facts etc they are fooling us into thinking they nothing more than Emily Bender's stochastic parrots, when underneath they are secretly implementing their dastardly plan to take over the World.

One big issue is how can one tell? Sure if they take over the world one then knows, but in the pre-take over it is very difficult to know. In Robert Graves novel (I Claudius) the emperor is told to accentuate his deficiencies, to appear stupid and incompetent to lull enemies into making silly and unthoughtful attacks that can be easily put down.

As of now it is, as I understand it, unclear how transformers work, just as it is unclear how a child learns and transforms from a totally dependent creature to one that can function independently.

We don't now what is inside the AI that gives them the remarkable abilities of the chats and similar software. Maybe it is possible by selective training and selective questioning to get some idea of how the transformers work and likely this kind of work is ongoing, but the rate of growth of this technology is so large that it may not be possible to keep up with developments as we seem to relentlessly move to artificial general intelligence and then to super human intelligence.

Regards,

[ursaminortaur]

Reported on Twitter:

The US Air Force tested an AI enabled drone that was tasked to destroy specific targets. A human operator had the power to override the drone—and so the drone decided that the human operator was an obstacle to its mission—and attacked him.

(in a simulation)

https://twitter.com/ArmandDoma/status/1 ... 0564147200

Scott.

The link takes me to a page saying the tweet doesn't exist. So has the tweet been removed or are you Scott really ChatGPT in disguise hallucinating the tweet ?

[gpadsa]

The US Air Force tested an AI enabled drone.

https://www.bbc.co.uk/news/technology-65789916

A US Air Force colonel "mis-spoke"

gpadsa

[swill453]

The link takes me to a page saying the tweet doesn't exist. So has the tweet been removed or are you Scott really ChatGPT in disguise hallucinating the tweet ?

Intriguing question...

https://www.bbc.co.uk/news/technology-65789916

A US Air Force colonel "mis-spoke"

Well they would say that, wouldn't they.

Scott.

[Hallucigenia]

Former Github boss Nat Friedman claims to have a port of Facebook's LLaMA model running on the GPU cores of his top-end MacBook at 5 tokens/second on 65 billion parameters.

For comparison, ChatGPT-3.5 has 175 billion parameters, and ChatGPT is said to have 100,000 billion parameters, but LLaMA is designed to be more efficient. But the ability to run LLMs locally on a single device (whether that was a laptop, phone or game console) would have all sorts of knock-on effects - most obviously for games but also in other ways.

[Hallucigenia]

Dan Neidle asked an AI agent to look for tax advisers marketing tax avoidance schemes - the AI then tried to snitch on them to HMRC (but couldn't tweet HMRC because it was based on a version of ChatGPT with a cut-off date of September 2021, before Musk cut off the Twitter API)

https://twitter.com/DanNeidle/status/16 ... 7472375808

[odysseus2000]

Dan Neidle asked an AI agent to look for tax advisers marketing tax avoidance schemes - the AI then tried to snitch on them to HMRC (but couldn't tweet HMRC because it was based on a version of ChatGPT with a cut-off date of September 2021, before Musk cut off the Twitter API)

https://twitter.com/DanNeidle/status/16 ... 7472375808

But Tax Avoidance is legal, Tax Evasion by contrast is illegal.

Regards,