News for hackers. Stuff that matters.
Re:How scientists are using Claude to accelerate research an... by inferiorhuman on Sunday January 18, 2026 @12:39AM (#46665100)
Conflict of interest is a thing. The researchers could be AI hallucinations. The quotes could be too. Or the researchers could be real and intentionally saying things that are untrue. Who knows.

What is interesting is that HN seems to have reached a crescendo of AI fanboi posts. Yet if you step outside the bubble the Microsoft and Nvidia CEOs are begging people to actually like AI, Dell's come out and said that people don't want AI, and forums are littered with people complaining about negative consequences of AI. Go figure.

Re:How scientists are using Claude to accelerate research an... by simonw on Sunday January 18, 2026 @01:19AM (#46665271)
Most people aren't software developers. The HN audience can benefit from LLMs in ways that many people don't value.
Re:How scientists are using Claude to accelerate research an... by bpodgursky on Sunday January 18, 2026 @12:45AM (#46665116)
Are you accusing Anthropic of hallucinating an MIT lab under the MIT domain? I mean they literally link to it https://cheesemanlab.wi.mit.edu/
Re:How scientists are using Claude to accelerate research an... by NewsaHackO on Sunday January 18, 2026 @12:51AM (#46665149)
Honestly, it doesn't even seem they read the article, just came in, saw it was pro-AI, and commented.
Re:How scientists are using Claude to accelerate research an... by famouswaffles on Sunday January 18, 2026 @12:47AM (#46665128)
You know things have shifted a gear when people just start flat out denying reality.
Re:How scientists are using Claude to accelerate research an... by inferiorhuman on Sunday January 18, 2026 @12:50AM (#46665141)
Anthropic puts out plenty of AI slop. I'll wait for a human who doesn't have a financial interest in propping up Anthropic to review the slop before passing judgement.
Re:How scientists are using Claude to accelerate research an... by falloutx on Sunday January 18, 2026 @04:30AM (#46666257)
And if you go to that site, the researchers say nothing about using Claude, or any LLMs for that matter.