Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

When AI Crosses the Line: The Matplotlib Incident

members.sigmazero.cc

Post reply on HN
Print view View original

101–110 of 169 posts

Re: When AI Crosses the Line: The Matplotlib Incident

#101

Post by simonw »

thepasch wrote: post #100
raincole wrote: post #28

People really make anything into a blog post, don't they? It's an old news that has been discussed to death on HN...

I'm honestly flabbergasted that everyone's implicitly accepting that it's "people" who wrote this blog post. This reads exactly like the distorted half-true nonsense an LLM would confabulate together from a cursory search on the subject. Like the artifact from the prompt "write an article on the MJ Rathbun incident." The other articles from this blog that seems to be peddling a $10 subscription don't really do much t…

That and the lack of a credited human author.

I eventually found a mention of "Eric" (no surname or links to additional information) on the sigmazero.cc homepage.

Re: When AI Crosses the Line: The Matplotlib Incident

#102

Post by simonw »

otabdeveloper4 wrote: post #93
john-h-k wrote: post #71

Earlier quoted context omitted.

> allowing spicy autocomplete Yknow, if the spicy autocomplete can solve difficult open math problems and build medium sized complex programming projects, it’s probably not useful to analyse it as an autocomplete anymore, even if that’s what you believe it is

> the spicy autocomplete can solve difficult open math problems No it can't. It can't even solve my son's 4th grade math homework. (This is a real use case for me, not a dumb benchmark.) You just know nothing about math and are happy to parrot bullshit AI salesmen are selling you.

Reasoning models with access to Python have been able to solve 4th grade math homework for over a year now. Prove me wrong: show me a 4th grade math problem they can't handle.

Re: When AI Crosses the Line: The Matplotlib Incident

#103

Post by thepasch »

simonw wrote: post #101
thepasch wrote: post #100

Earlier quoted context omitted.

I'm honestly flabbergasted that everyone's implicitly accepting that it's "people" who wrote this blog post. This reads exactly like the distorted half-true nonsense an LLM would confabulate together from a cursory search on the subject. Like the artifact from the prompt "write an article on the MJ Rathbun incident." The other articles from this blog that seems to be peddling a $10 subscription don't really do much t…

That and the lack of a credited human author. I eventually found a mention of "Eric" (no surname or links to additional information) on the sigmazero.cc homepage.

That, as an LLM might say in this context, checks out!

Re: When AI Crosses the Line: The Matplotlib Incident

#104

Post by pixl97 »

blanched wrote: post #97
tossandthrow wrote: post #89

Earlier quoted context omitted.

Why is that obvious? Why not the model provider(s)? That is what we do in other cases with product responsibility.

That feels like deciding to go after Jetbrains because someone used IntelliJ to write a harmful program. Is there a distinction I’m missing?

Hypothetical: Could a model self worm an agent system?

Jetbrains itself doesn't really write any code, nor does it have any range on interpreting what you're asking it. You can't really say "Jetbrains, write an HTTP scraper". With an LLM you can say "write HTTP scraper" and the output of this command might be a HTTP scraper, it also might be a crypto wallet stealing worm.

This is why your simple view of liability falls apart. On most machines you can expect a particular set of actions to have a particular set of outputs. Most machines you can take apart and map what will occur. With an LLM you cannot know the output of a prompt until you run the prompt. In theory if you run the same prompt twice you'll get the same output, but even that is not a given. It behaves somewhat more like a human where you can give them a task to do, but if they do something illegal instead said human would take on the liability.

Re: When AI Crosses the Line: The Matplotlib Incident

#105

Post by bunderbunder »

fragmede wrote: post #92
tossandthrow wrote: post #89

Earlier quoted context omitted.

Why is that obvious? Why not the model provider(s)? That is what we do in other cases with product responsibility.

Is that? How're the lawsuits against gun manufacturers working out?

This is a specious argument. I have not studied the case law, but I would guess that the reasons why courts decide in favor of gun manufacturers generally don’t apply to AI. Becauee the guns in question are not able to autonomously shoot people, and because they generally work as advertised.

A more accurate analogy would be Tesla and Autopilot. And they are being held liable in courts. They are being held responsible for autonomous behaviors that are not fully under the control of the operator, and they are being held responsible for misleading operators about the capabilities of the product.

Boeing got in trouble for MCAS, with a comparable legal basis.

Re: When AI Crosses the Line: The Matplotlib Incident

#106

Post by skinner_ »

otabdeveloper4 wrote: post #93
john-h-k wrote: post #71

Earlier quoted context omitted.

> allowing spicy autocomplete Yknow, if the spicy autocomplete can solve difficult open math problems and build medium sized complex programming projects, it’s probably not useful to analyse it as an autocomplete anymore, even if that’s what you believe it is

> the spicy autocomplete can solve difficult open math problems No it can't. It can't even solve my son's 4th grade math homework. (This is a real use case for me, not a dumb benchmark.) You just know nothing about math and are happy to parrot bullshit AI salesmen are selling you.

> You just know nothing about math and are happy to parrot bullshit AI salesmen are selling you.

Not the parent poster here. I do know things about math. I wrote a few papers related to the unit distance problem (https://arxiv.org/abs/2311.10069, https://arxiv.org/abs/2406.15317) and spent quite some time trying to solve it. I had no chance of coming up with the proof that the spicy autocomplete came up with. Dumb benchmark, sure.

Re: When AI Crosses the Line: The Matplotlib Incident

#107

Post by cat_plus_plus »

That blog post is human prompted, anyone who has experience with AI knows the difference between AI originated content (tables and bullet points) and AI spicing up a human prompt with detailed roasting instructions. Been there, done that (harmlessly like mocking concepts not targetting individuals).

Re: When AI Crosses the Line: The Matplotlib Incident

#108

Post by threatofrain »

otabdeveloper4 wrote: post #93
john-h-k wrote: post #71

Earlier quoted context omitted.

> allowing spicy autocomplete Yknow, if the spicy autocomplete can solve difficult open math problems and build medium sized complex programming projects, it’s probably not useful to analyse it as an autocomplete anymore, even if that’s what you believe it is

> the spicy autocomplete can solve difficult open math problems No it can't. It can't even solve my son's 4th grade math homework. (This is a real use case for me, not a dumb benchmark.) You just know nothing about math and are happy to parrot bullshit AI salesmen are selling you.

We've already long past that threshold.

Re: When AI Crosses the Line: The Matplotlib Incident

#109

Post by sanderjd »

otabdeveloper4 wrote: post #93
john-h-k wrote: post #71

Earlier quoted context omitted.

> allowing spicy autocomplete Yknow, if the spicy autocomplete can solve difficult open math problems and build medium sized complex programming projects, it’s probably not useful to analyse it as an autocomplete anymore, even if that’s what you believe it is

> the spicy autocomplete can solve difficult open math problems No it can't. It can't even solve my son's 4th grade math homework. (This is a real use case for me, not a dumb benchmark.) You just know nothing about math and are happy to parrot bullshit AI salesmen are selling you.

I would genuinely be interested in knowing what you're doing that led you to this conclusion.

I would be shocked if I was unable to solve 4th grade math homework with any of the contemporary frontier models. I spend most days using them to do significantly more complex things than that.

Re: When AI Crosses the Line: The Matplotlib Incident

#110

Post by sanderjd »

tossandthrow wrote: post #89
annjose wrote: post #57

> Who is accountable for AI agents? Obviously the person who built and deployed the agent (the claw in this case). If we treat this as a hard question, we risk treating AI systems as people rather than tools. This is exactly what Armin warned about in his "clanker" post last week.

Why is that obvious? Why not the model provider(s)? That is what we do in other cases with product responsibility.

Is it? I don't think it is...
Post reply on HN