🔥Telegram İfşa

Happy "The NetHack Learning Environment is still completely unsolved" day for those of you who are celebrating it. We released The NetHack Learning Environment (arxiv.org/abs/2006.13760) on this day five years ago. Current frontier models achieve only ~1.7% progression (see balrogai.com).

I am hugely excited to announce that ThinkSono has raised $6Mn in our oversubscribed seed two funding round ($13Mn raised to date)! This was led by Id4 ventures with participation from Cur8 Capital, Brandenburg Kapital, clinical leaders and the founders!

A little shot of hope and optimism today from the United Nations. I got to speak about AI at their annual Internet Governance Forum (my first time at a UN thing 🤩) and it really does feel good to see so many people from so many places around the world convening in good faith to

In Silicon Valley, old-school spreadsheet measures are out; vibe valuing is in. We explain how AI startups are raising billions before even making a dollar econ.st/3Ghqco1

The LLM bias to positivity is interesting.

Microsoft claims their new AI framework diagnoses 4x better than doctors. I'm a medical doctor and I actually read the paper. Here's my perspective on why this is both impressive AND misleading ... 🧵

Many top academic AI researchers are GPU poor. The Sovereign AI Unit wants to change this story. We are seeking ambitious AI research proposals for UK academics aiming to take their research to the next level. Read below to find out more and apply 🧵👇

77 years ago this week Claude Shannon ushered in the field of information theory with his paper "A Mathematical Theory of Communication," which has been cited over 100,000 times: bit.ly/44rfByX

We are now closer to the year 2100 than to 1950. Also closer to 2050 than to 2000. Time to start acting like it.

Today is the day of the pre-ICML event at UCL! Come check out the exciting work from academics, industry researchers, postdocs and PhD students from around London: sites.google.com/view/pre-icml-… UCL Statistical Science UCL CSML

The automation of innovation is within reach! Delighted that my RAAIS talk is now available for anyone to watch, alongside an excellent blogpost summary by the inimitable Nathan Benaich.

Grok 4 (Thinking) achieves new SOTA on ARC-AGI-2 with 15.9% This nearly doubles the previous commercial SOTA and tops the current Kaggle competition SOTA

I’m excited about this: 12 month fellowship for AI engineers to work on some of the biggest challenges in government👇

I am starting to think sycophancy is going to be a bigger problem than pure hallucination as LLMs improve. Models that won’t tell you directly when you are wrong (and justify your correctness) are ultimately more dangerous to decision-making than models that are sometimes wrong.

If you are building advanced technology in Europe check out Andreas Klinger 🦾’s PROTOTYPE prototypecap.com (now on Fund III) 🇪🇺🦾

Humanity has prevailed (for now!) I'm completely exhausted. I figured, I had 10h of sleep in the last 3 days and I'm barely alive. I'll post more about the contest when I get some rest. (To be clear, those are provisional results, but my lead should be big enough)

Today we're releasing a developer preview of our next-gen benchmark, ARC-AGI-3. The goal of this preview, leading up to the full version launch in early 2026, is to collaborate with the community. We invite you to provide feedback to help us build the most robust and effective

What's included in the developer preview: • Three test games: Go play them yourself – they're fun! • An agent API: Start building and testing your agents against the games. • Sprint competition: Submit the best-performing agent in the next 4 weeks and win a $10,000 prize.

Do you have a PhD (or equivalent) or will have one in the coming months (i.e. 2-3 months away from graduating)? Do you want to help build open-ended agents that help humans do humans things better, rather than replace them? We're hiring 1-2 Research Scientists! Check the 🧵👇

if you’re interested in pushing new records on fusion with AI, we’ve just launched a kaggle-like competition with Proxima Fusion (the startup that spun out of Wendelstein 7-X discussed in this HN news) All the details at huggingface.co/blog/cgeorgiaw…