Ron Arel

Feel the rain on your skin No one else can feel it for you Only you can let it in No one else, no one else Can speak the words on your lips

Genuinely horrific way to experience life and the people around you. People don’t try to make new friends because it’s “what you are supposed to do”. Life is given meaning by the wonderful people you meet along the way and choose to spend time with. You don’t have to be friends with everyone, but to shut off the possibility of new human connection due to a lack of calculated “ROI” is fundamentally misunderstanding of the human experience.

Goodfire casually doing the coolest things ever

I've joined @IntologyAI! I'm excited to push the boundaries of AI-accelerated scientific discovery with an incredibly driven and talented team. Looking forward to dive deep into research on AI-driven automation and creativity!

Locus achieves a new WR on the nanogpt speedrun by developing a fused kernel:

New NanoGPT Speedrun WR at 105.9s (-1.0s) from @soren_dunn_ , with a triton kernel to fuse the logit softcap and multi-token prediction cross entropy calc. Interestingly, Soren mentioned that their autonomous system Locus at Intology discovered and implemented the improvement. github.com/KellerJordan/m…

You'll still be standing next to me You could be my luck Even if we're six feet underground I know that we'll be safe and sound

Cause you're hot, then you're cold You're yes, then you're no You're in, then you're out You're up, then you're down

Programmer is out Shoggoth pilot is in

@akuhayum Grok propaganda

The insane evolution of Al over the years.

Awesome results by the @poetiq_ai team!

@Rainmaker1973 guys... its AI yall

This video briefly compares bite force across dog breeds, showing how size and jaw structure affect potential impact.

@Zoom So Zoom is just a frontier lab now?

Zoom achieved a new state-of-the-art (SOTA) result on Humanity’s Last Exam (HLE): 48.1% — outperforming other AI models with a 2.3% jump over the previous SOTA. ✨ HLE is one of the most rigorous tests in AI, built to measure real expert-level knowledge and deep reasoning across complex problems. What that means for you: ✅ More accurate summaries ✅ Better reasoning ✅ More powerful automation in AI Companion 3.0 Click the link to learn more. 🔗 zm.me/3MxVbyS

@NeurIPSConf Admitting the 13 1-star reviews is crazy

We have been made aware of several fake apps pretending to be the NeurIPS official app. To clarify, NeurIPS is using atconf. We advise attendees to carefully check thet they are downloading the correct app.

@TMZ I'm sorry, did TMZ buy an AI-generated video thinking it was real and posted it???

While TMZ has obtained several real images of Diddy at Fort Dix federal prison ... this video and the other crystal clear selfies you might have seen, are NOT one of them. tmz.me/dj0oWUX

@ns123abc @ssi What Ilya is on

Ilya Sustkever explains @SSI business model: “the answer will reveal itself”

@niklassheth @IntologyAI We did! We also brought in a 3rd party to verify the most exciting ones! You can check them out on the evals repo: github.com/IntologyAI/loc…

@IntologyAI Have you manually validated that the generated kernels are correct?

Introducing Locus: the first AI system to outperform human experts at AI R&D Locus conducts research autonomously over multiple days and achieves superhuman results on RE-Bench given the same resources as humans, as well as SOTA performance on GPU kernel & ML engineering tasks. RE-Bench is a collection of several frontier AI research tasks that typically take human experts (e.g., top ML PhDs and frontier lab researchers) several days. By scaling experimentation to far longer time horizons than previous systems, Locus represents a step change in AI scientist capabilities. 🧵