And how do you propose to get that medical diagnosis and treatment plan right when you lack the knowledge, education and experience to spot the mistakes? Right.

Good luck, you are going to need it. LLMs cannot replace experts.

Well, the belief in AI is a cult-like thing. No rationality involved. The last few AI hypes already showed that nicely.

Given that the current record for quantum-factorization is 35 (!), they will have to do some industrial-strength lying to get that across. Not that these two are unfamiliar with that approach.

Treat it as if they wrote it themselves, which means fraud, falsifying documents, lying under oath.. etc.

No? Why would you even think that except if you just wanted to throw crap?

Here in the real world, a perceptron is a very simple thing and an LLM very much is not. Are you unaware of that?

Namely that LLM-type AI has taken over and hence AI must be really great and valuable and the future and whatnot.

The truth looks a bit different. LLMs are still incapable as hell and will continue to hallucinate, because there is no way to fix that. Keeping LLMs updated gets harder and harder due to AI slop. Nobody is making any profits on LLMs, these things just burn money like crazy. And there are very few somewhat working use-cases and these come with really big caveats.

No, they cannot. LLMs cannot do logical reasoning and that is a requirement for any type of fact-checking or verification. All they can do is correlations and that is not enough.

While using non-LLM AI or non-AI tools is a possibility for fact-checking (in simple cases as the one here), you overlook that basically the only advantage of LLMs is that they are comparatively easy to create. Creating these "other tools" that would be needed here would be a lot more expensive and a lot more work than creating general LLMs. And hence, for most cases, these "other tools" do not exist.

I do not "think so". I am just referring to research results. You, on the other hand, have nothing. LLMs cannot fact-check, period.

Robotics just uses the fact that you can train whatever you use pretty well and you can use _other_ systems there to install guardrails and these other systems come with physics models and can do (limited) fact checking in there. These other systems are not LLMs though. LLMs cannot fact-check in robotics either.

I am not speaking for myself. I am speaking for solid results from criminology. You think you are smarter and know more than tons of researchers that have studied this for a long, long time?

Think about the logic you are applying here. It makes absolutely no sense.

What a stupid comment that entirely misses the point.

Indeed. Sure, as an experimental language, Rust is fine. But a specifically "secure" programming language must be held to a higher standard. It seems the IT field is still bumbling around under the illusion they know what they are doing.

Sure. This one is specifically aimed at creating secure software. A full spec is _mandatory_ for that purpose or you cannot prove properties. Seriously. Some insight required.

That summary at the top of this story is just way too long. I'll have a chatbot break it down and give me the gist.

Indeed. KISS is the basis of all solid engineering and complexity is the death of it. I do not think the people at the wheel are aware of that. Probably a lack of experience and insight, which would be bad on a whole other level.