🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

You can laugh at xAI folks sleeping in office tents or grinding till 4:20 AM on weekends, but you have to admit: they’re the fastest-moving AI lab out there.

I don't really think benchmarks translate the well into real practical use cases, but boy looks like Grok 4 and xAI are making openAI feel that "first-mover disadvantage" I mean xAI has only been doing this for what < 2 years.

I'd hope so, never really thought painting a PhD as "super book smart" made sense ... sure some theory people are like that ... ask them a question they just know it ... but a PhD is all about asking questions and figuring out how to come to an understanding/meaning or answer.

So who at xAI is getting a fatty bonus?

i am beginning to suspect that Humanity’s Last Exam may not in fact be humanity’s last exam

"building self-driving labs powered by foundation models" This is my credo as well! I say something similar to my self every morning 🫡 Radical AI

Would make my PhD advisor proud ... he loved to program complex physics codes in a single FORTRAN file.

I'm coming to the realization that my research papers should be written first as web apps ... meaning rather than write LaTeX document for formal publishing ... write a web app structured with all the standard sections but with the additional functionality needed, then when you

Can it single shot code for Kohn-Sham DFT SCF using all electron LDA for BCC Li unit cell?

little embrassing but I still pay for $20 ChatGPT subscription, I've just been conditioned to use it but guess I should probably make the switch to some other provider ... don't really know who ... I think Gemini Pro 2.5 is really good but I don't really like outputs; it writes

well well after 2 years, 7 months, and 11 days OpenAI finally becomes Open!

Although not knowing the math and theoretical physics to appreciate things in detail ... I always thought Eric Weinstein's Geometric Unity was more an Idea than a theory. Its a loose guess of what explains our reality but its not a rigorously developed or substantiated by the

I'm very supportive of this in general, but headlines like this send the wrong message. These are AI-model based predictions with no synthesis aspect being explored, which is a major challenge, usually. Also, AI is trained on in-silico data only. Companies like Radical AI

Data availability (not 'by reasonable request'; you click a link in a paper, and there it is). My PhD advisor was talking about doing this with the group's data more than a decade ago. I was skeptical because 1) it could get misinterpreted by people who don't know the basics 2)

😬, didn't see that coming. But much ❤️ Janosh . You should connect with Matt on Ouro, maybe some synergy? Also glad someone is advancing in browser visualizations for materials and crystallography people. Let me know if you want my wish list (most would be from VESTA)

I think google chat is underrated

🤣

No way it gets better any time soon, AI/LLMs make it even easier to produce quality papers. I'm seeing very serious high-quality work in top-journals at unprecedented rates; just look at Aspuru-Guzik or K. Choudhary . The bigger issue is the question of value, is it valuable

Academia is perception Entrepreneurship is action