c/technology · by beep@piefed.world · 10d

Real Science Is Harder Than Benchmarks: University students find that AI isn’t the cure-all in drug discovery, as AI isn’t as capable as human researchers for drug discovery and development

https://news.northeastern.edu/2026/08/26/ai-drug-discovery-study/

cross-posted from: https://piefed.world/c/tech/p/1363454/real-science-is-harder-than-benchmarks-uni-students-find-that-ai-isnt-the-cure-all-in-dr

142 points · 9 comments · view on lemmy.world

9 Comments

yeahiknow3@lemmy.dbzer0.com · 30 pts · 10d

This is absolutely shocking. It’s almost as if those theorems from the early 20th century establishing that digital algorithms can’t replace epistemology are true or something. Who could have guessed? Other than Gödel, Tarski, and Turing.

CosmoNova@lemmy.world · 16 pts · 10d (1 reply)

Mind you this is about the currently heavily used machine learning architecture and not even LLMs which are another product of this architecture and have even less potential in my opinion.

Machine learning is a useful tool for research but it‘s just a tool that needs to be handled by competent people.

I‘m mean we‘re seeing what happens when you let AI loose on incompetent individuals.

ranzispa@mander.xyz · 3 pts · 10d

This is very much about LLMs. All the methodologies they tested are based on LLMs.

bugbear@lemmy.sdf.org · 9 pts · 10d (5 replies)

Oh good, so AI proponents can finally shut up about it.

wildncrazyguy138@fedia.io · 5 pts · 10d (2 replies)

But it wasn’t all bad news. In some respects, AI can be a useful tool in completing repetitive tasks, statistical support and stress testing ideas, Sinitskiy said the research revealed.

“The point is AI can do really important things, but you need to know what kind of things it can do, and you need to run it properly,” he said. “That’s what I’m teaching students in my courses to do.”

GreenBeard@lemmy.ca · 1 pts · 10d (1 reply)

What I'm hearing you say is AI is really only useful for doing the busy-work you already know well, but needs to be done anyway. No one should be using it to do things that you can't do yourself... which is a use, but not what they've been advertising.

On a different level, what I'm hearing you say is the secret sauce in AI is us, it always has been, there are billions of people effectively operating a giant Mechanical Turk against their will and as far as the executives are concerned, all the real people making this stuff possible are invisible... which is a problem.

wildncrazyguy138@fedia.io · 1 pts · 10d

I fear that you, much like AI on occasion, seem to be hallucinating.

Get those ears checked out.

ranzispa@mander.xyz · 4 pts · 10d (1 reply)

Machine learning is widely applied and useful in drug discovery.

A couple years ago the Nobel prize was given out for that.

The approaches they tested, which use LLMs, were not satisfactory for their objectives.

bugbear@lemmy.sdf.org · 2 pts · 10d

I'm aware, yes. That was my point