The fact that there is absurd AI hype right now doesn't mean that we should let equally absurd bullshit pass on the other side of the spectrum. Having a reasonable and accurate discussion about the benefits, drawbacks, side effects, etc. is WAY more important right now than being flagrantly incorrect in either direction.

Meanwhile this entire comment thread is about what appears to be, as fumi2026 points out in their comment, a predatory marketing play by a startup hoping to capitalize on the exact sort of anti AI sentiment that you seem to think is important... just because there is pro AI sentiment?

Naming and shaming everyday researchers based on the idea that they have let hallucinations slip into their paper all because your own AI model has decided thatit was AI so you can signal boost your product seems pretty shitty and exploitative to me, and is only viable as a product and marketing strategy because of the visceral anti AI sentiment in some places.

“anti-ai sentiment”

No that’s a straw man, sorry. Skepticism is not the same thing as irrational rejection. It means that I don’t believe you until you’ve proven with evidence that what you’re saying is true.

The efficacy and reliability of LLMs requires proof. Ai companies are pouring extraordinary, unprecedented amounts of money into promoting the idea that their products are intelligent and trustworthy. That marketing push absolutely dwarfs the skeptical voices and that’s what makes those voices more important at the moment. If the researchers named have claims made against them that aren’t true, that should be a pretty easy thing for them to refute.

The cat is out of the bag tho. AI does have provably crazy value. Certainly not the agi hype marketing spews and who knows how economically viable it would be without vc.

However, i think any one who is still skeptical of the real efficacy is willfully ignorant. This is not a moral endorsement on how it was made or if it is moral to use but god damn it is a game changer across vast domains.

There were a number of studies already shared reporting on the impression of increased efficiency without the actual increase in efficiency.

Which means that it's still not a given, though there are obviously cases where individual cases seem to be good proof of it.

There was a front page post just a couple of days ago where the article claimed LLMs have not improved in any way in over a year - an obviously absurd statement. A year before Opus 4.5, I couldn't get models to spit out a one shot Tampermonkey script to add chapter turns to my arrow keys. Now I can one small personal projects in claude code.

If you are saying that people are not making irrational and intellectually dishonest arguments about AI, I can't believe that we're reading the same articles and same comments.

Isn’t that the whole point of publishing? This happened plenty before AI too, and the claims are easily verified by checking the claimed hallucinations. Don’t publish things that aren’t verified and you won’t have a problem, same as before but perhaps now it’s easier to verify, which is a good thing. We see this problem in many areas, last week it was a criminal case where a made up law was referenced, luckily the judge knew to call it out. We can’t just blindly trust things in this era, and calling it out is the only way to bring it up to the surface.

> Isn’t that the whole point of publishing?

No, obviously not. You're confusing a marketing post by people with a product to sell with an actual review of the work by the relevant community, or even review by interested laypeople.

This is a marketing post where they provide no evidence that any of these are hallucinations beyond their own AI tool telling them so - and how do we know it isn't hallucinating? Are there hallucinations in there? Almost certainly. Would the authors deserve being called out by people reviewing their work? Sure.

But what people don't deserve is an unrelated VC funded tech company jumping in and claiming all of their errors are LLM hallucinations when they have no actual proof, painting them all a certain way so they can sell their product.

> Don’t publish things that aren’t verified and you won’t have a problem

If we were holding this company to the same standard, this blog wouldn't be posted either. They have not and can not verify their claims - they can't even say that their claims are based on their own investigations.

Most research is funded by someone with a product to sell, not all but a frightening amount of it. VC to sell, VC to review. The burden of proof is always on the one publishing and it can be a very frustrating experience, but that is how it is, the one making the claim needs to defend themselves, from people (who can be a very big hit or miss) or machines alike. The good thing is that if this product is crap then it will quickly disappear.

That's still different from a bunch of researchers being specifically put in a negative light purely to sell a product. They weren't criticized so that they could do better, be it in their own error checking if it was a human-induced issue, or not relying on LLMs to do the work they should have been. They were put on blast to sell a product.

That's quite a bit different than a study being funded by someone with a product to sell.