Fyi, “AI” has been used in medicine research for decades. GenAI is the one that’s wonky. I’d be surprised and sceptical of any researcher that would suggest genAI as the star tool when there are so many predictive ML models that already work so well…
AI may soon make Nobel-level discovery, scientists predict
Submitted 17 hours ago by return2ozma@lemmy.world to technology@lemmy.world
https://www.semafor.com/article/10/07/2025/ai-may-soon-make-nobel-level-discovery-scientists-predict
Comments
fushuan@lemmy.blahaj.zone 10 hours ago
CosmoNova@lemmy.world 16 hours ago
Technically machines make most discoveries possible these days but I have yet to see an electric microscope receive the prize. I don‘t see how this is any different.
SnoringEarthworm@sh.itjust.works 11 hours ago
This is the first of three waves of AI in science, says Sam Rodriques, chief executive of FutureHouse — a research lab in San Francisco, California, that debuted an LLM designed to do chemistry tasks earlier this year.
The “one scientist” seems to have switched tracks from doing science to doing capitalism.
Bronzebeard@lemmy.zip 14 hours ago
We’ve been shoving large amounts of data into machine learning algorithms for ages now. Still need people to interpret the outputs and actually test that the results are accurate.
nyan@lemmy.cafe 13 hours ago
I’m pretty sure that you can find one researcher, somewhere, who will agree with anything you say, including that the weather is being affected by a war between Martians and the people living inside the hollow earth. Especially if you’re offering a large bribe to said researcher to make a statement about something outside their field while they’re somewhat drunk, and then mutilating their remark out of context via the process fondly known as journalism.
In other words, “one researcher” predicting something is pretty much worthless.
Fyrnyx@kbin.melroy.org 16 hours ago
Well, until AI finds a cure for cancer, solves climate issues and fixes the economy for everyone, it is still shit.
Xanthobilly@lemmy.world 16 hours ago
Solves climate issues by turning itself off.
Fyrnyx@kbin.melroy.org 16 hours ago
After it does the other two things, yes.
lunatique@lemmy.ml 13 hours ago
Can you find a cure to cancer? If you don’t are you shit?
Fyrnyx@kbin.melroy.org 12 hours ago
You're just mad because you suck at posting.
Buffalox@lemmy.world 14 hours ago
“eventually” is a cheap cop out. Because I have no doubt AI will eventually surpass us, it’s simply the nature of the speed of development of technology over evolution. But we are not there yet.
phdepressed@sh.itjust.works 14 hours ago
Eventually we’ll make agi instead of this llm bullshit. Assuming we don’t destroy ourselves first.
frezik@lemmy.blahaj.zone 11 hours ago
This one probably will happen.
The reason is that there are certain fields where you have to sift through massive amounts of data to find the thing you’re looking for. This is an ideal task for machine learning. It’s not going to replace real scientists, and it sure as hell shouldn’t replace peer review. It’s a tool with a use.
As one example, the longest known black hole jet was recently discovered using ML techniques: caltech.edu/…/gargantuan-black-hole-jets-are-bigg…
baggachipz@sh.itjust.works 14 hours ago
Any day now….
bufalo1973@piefed.social 15 hours ago
May… or maybe not.
6nk06@sh.itjust.works 17 hours ago
Nice astrology bro.
brathoven@feddit.org 17 hours ago
Nobel price in astrology, hallucinated.