Eric Lawton

“Can a chatbot be a co-author? AI helps crack a long-stalled gluon amplitude proof”

We don't ask if spreadsheets or other programs can be co-authors.

We are allowing AI™ vendors to twist our language for their marketing campaign to get us to accept their software as people.

https://phys.org/news/2026-02-chatbot-author-ai-stalled-gluon.html

#propaganda #AI #BigTech

Can a chatbot be a co-author? AI helps crack a long-stalled gluon amplitude proof

Like many scientists, theoretical physicist Andrew Strominger was unimpressed with early attempts at probing ChatGPT, receiving clever-sounding answers that didn't stand up to scrutiny. So he was skeptical when a talented former graduate student paused a promising academic career to take a job with OpenAI. Strominger told him physics needed him more than Silicon Valley.

Phys.org
Feb 21 at 4:53pmWeb
Show threadSzescstopniFeb 21
@EricLawton And once they have personhood by law, like companies, we won't be able to kill them.
Show threadWilliam WhitlowFeb 21

@Szescstopni @EricLawton

I suppose the logical consistency of this would then involve complex legal proceedings equivalent to withdrawing life support in order to close a computer program...

At that point would AI data centers be approaching the equivalence of hospitals from a critical infrastructure perspective?

Show threadBlahsterFeb 21
@wwhitlow @Szescstopni @EricLawton they are even above that just now - do you know how many diesel generators they have on their rooftops and how much fuel they store in their basements?
Show threadWilliam WhitlowFeb 21

@Blahster @Szescstopni @EricLawton

No, if you have a source I would be fascinated to read about it.

Show threadBlahsterFeb 21
@wwhitlow @Szescstopni @EricLawton the source is the company I work for - we build those emergency fuel supplies. I can't give you the details, but every data centre has about 20-30 emergency generators of their own, depending on their size.
Show threadAnthonyFeb 21
@[email protected] Betteridge's law of headlines dictates the answer is no, of course, but this is a tired and tiresome question. From 3 years ago: https://bucci.onl/notes/Another-AI-Hype-Cycle
Another AI Hype Cycle

Anthony Bucci's personal web site

Anthony Bucci
Show threadmike805Feb 21

@abucci @EricLawton This reminds me of the era when "UNIVAC predicted the Presidential election ahead of time!"

No it didn't. Someone wrote a pretty neat program, given that a computer had never been used to analyze an election before. Does anyone remember who that was?

Computers are back to being "giant brains" in the public mind. Now as then, it is basically a magic show where the manipulation is being done out of sight of the audience.

Show threadAnthonyFeb 22
@[email protected] @[email protected] That was definitely a UNIVAC, the first one apparently. Grace Hopper is known for being involved in creating it but I don't know if she was involved in creating the election predicting program they used. It'd been awhile since I revisited this and had forgotten that CBS didn't have the real computer in their studios. They had a dummy terminal that looked like the real deal but they were teletyping back and forth with the factory where the computer itself was running. Quite a stunt all around.
Show threadmike805Feb 22

@abucci @EricLawton Whenever the public was shown a computer in that era, there was a sort of ritual they performed. The reporter would be asked "Would you like to give UNIVAC a problem?" and he'd come up with something like "car A leaves at X mph, Y minutes later car B leaves the same point at Z mph, when do they meet?"

The techies would push some buttons and the machine would print the answer and everyone was duly impressed. What they did not tell you was, the reporter had given the 1/3

Show threadmike805Feb 22

@abucci @EricLawton problem ahead of time, the programmer did the algebra, and the machine merely did the arithmetic. So people were led to believe the machine was intelligent when it was not.

At the UK's largest computing center, people could request time to solve hard math problems. The staff were always having to explain to some relativity crank that the computer can only solve a problem if you know how to solve it with a calculator (even if that would take years.)

For a while with PCs 2/3

Show threadmike805Feb 22

@abucci @EricLawton people understood what a computer's limitations were.

With LLMs we are back to the public imagining computers to be giant brains that know everything.

It's still luck of the draw whether an LLM can correctly solve one of those linear algebra word questions. And if you don't know how to solve it you cannot check the LLM's answer.

Same old problem. 3/3