Post

Conversation

This is amazing. AI-assisted materials discovery boosted the number of new discovered materials by 44%, pushed up patent filings by 39%, and increased the number of new prototypes by 17%. AI is already transforming scientific discovery.
Image
Quote
Caleb Watney
@calebwatney
This is the best paper written so far about the impact of AI on scientific discovery
Image
David Watson 🥑
Post your reply

And maybe even reversing decline! Finding numbers were declining somewhat prior to AI adoption, but after it, the top researchers' productivity doubled, increasing inequality in productivity, but also pushing up general productivity massively.
🤖 Build Powerful AI Agents with Momen 🤖 No code, no limits—automate complex tasks with our newly launched AI feature! Build your full-stack AI apps today!
1:49
Build Your AI Agents with No Code
From momen.app
We’ll discover a bunch of amazing potential material, but won’t be able to make them until the cost of energy drops by a factor of 100. Most of these patents will have expired by the time any of these be made at scale.
Don't you think it's kindof strange to read through that entire paper, and there is no mention of a single material that was "discovered" or patented, there was no mention of a single applied problem that was solved with a new material? What exactly are they doing? The graphene
Show more
The study examines the impact of artificial intelligence (AI) on scientific discovery and innovation. It focuses on a large research and development (R&D) lab that specializes in materials science, a field that integrates insights from physics, chemistry, and engineering to
Show more
Image
It’s exciting to see tangible impacts of AI in scientific fields. This is proof that AI is more than just hype, it’s already reshaping industries.
I'll believe it when the metamaterials show in products. eg rare earths replaced by not-even-close-to-rare earths. or dirt cheap invisibility cloaks. or dirt-cheap perfect sound suppression wallpaper.
Missed the interesting part:
Quote
SHRIMPMONEYCAPITAL
@SHRIMPMONEYCAP
"AI-assisted researchers discover 44% more materials, resulting in a 39% increase in patent filings and a 17% rise in downstream product innovation." "... the bottom third of scientists see little benefit, the output of top researchers nearly doubles." "however, ...82% of x.com/calebwatney/st…
Show more
Incredible impact—AI-assisted materials discovery is reshaping research. A 44% increase in new materials, 39% boost in patent filings, and 17% rise in prototypes shows just how transformative AI can be for scientific discovery.
we’re nearing the singularity slowly but surely. if we take into account most available AI has an avg iq = 100 for all topics, just imagine what’ll happen when it hits 1000. nay, 1 million we are extremely early.
I’m confused. Isn’t using AI here just a supposition? They actually have to be able to make these compounds, have them be stable, be able to source some materials, and be able to do it for some reasonable price. Feels like the excitement’s a little bit ahead of itself.
I doubt it's sustainable. The result was because top researchers were able to use the AI as a "random idea generator" which they then evaluated. Basically it removed a time bottleneck from their own productivity /1
Are these new materials being used? Or is there just a space of all possible materials that is being explored more efficiently? I guess more materials discovered probably makes it easier to find use cases, but this seems inefficient?
I'm curious what specific AIs different fields use. I use LLMs to program, but what about everything else?
AI is revolutionizing materials science. With a 44% jump in new materials and a 39% surge in patent filings, it’s clear that AI is accelerating breakthroughs and paving the way for more innovation across industries.
Very interesting, thanks for sharing! Just a small remark btw - these numbers are related mostly to the "discovery" phase of materials research. Would be quite interesting to see, how many of these would be actually adopted by the industry.
AI works best where data contains the least bias and was not previously fully integrated nor previously systematically investigated. Materials sciences is such a niche. It works the worst when AI has substantial bias built into its algorithms such as in healthcare application.
Any guess on the number of patents that remain undisclosed to the public due to potential conflicts with the financial interests of influential entities.
then why are all the ai digital assistants FUCKING RETARDED? face it, AI seems like quantum computing and climate change. Forever a myth.
Show additional replies, including those that may contain offensive content

Discover more

Sourced from across X
I just read the paper in full; it is even more spectacular than I initially thought. A short thread about the results and their significance.
Quote
Caleb Watney
@calebwatney
This is the best paper written so far about the impact of AI on scientific discovery
Image
Some journalists seem alarmed that Elon has so much influence over Trump, but this is actually good news. Someone would have Trump's ear if Elon didn't, and it would surely be someone less smart and more politically extreme. Remember Steve Bannon?
I had a PM at Facebook in 2022 who was reported to HR by a new grad and reprimanded. His offense? He said "my WIFE" instead of "my partner." This is when I realized wokeness was out of control.
Quote
Paul Graham
@paulg
Prediction: Wokeness will recede significantly in 2024. There were always more people against it than there seemed, but many were afraid to say so. Now that it's safer to criticize it, more will.
Rate proposed Community Notes