we need more advanced evals to push forward the field!
RESEARCH
-
Reasoning Less Competitive Moat Than Expected in 2025
By
–
2025's biggest surprise so far: Reasoning is less of a moat than anyone thought.
-
LLMs overhype: breakthrough claims versus realistic applications
By
–
You mean it’s not an AMAZING BREAKTHROUGH and here are 50 USE CASES THAT WILL BLOW YOUR MIND? (LLMs have a lot to answer for.)
-

Rigorous AI Benchmarks Advance Field Standards
By
–
Awesome work from @hendrycks @alexandr_wang and teams, the field urgently needs much harder and scientifically rigorous benchmarks like this – congrats, and we look forward to testing on it!
-
Scaling AGI and Latest Breakthroughs Discussion
By
–
Thanks @Kantrowitz for the great chat. Really enjoyed covering some of my favorite topics… scaling, AGI, and some of our latest breakthroughs.
-
Best ArXiv Papers Read Throughout 2024
By
–
Ha, yeah, it's basically the "best of" of the arxiv slice I read in ~365 days in 2024 😛
-
The AI Literacy Project: Addressing Critical Knowledge Gaps
By
–
Introducing The AI Literacy Project We have an AI literacy problem. AI is changing everything about how we communicate, sell, learn, teach, design products, start and grow businesses, hire and develop talent, manage tech stacks, build strategies, allocate resources, and drive
-
PhD Research on Attention Architecture and NLP Benchmarks
By
–
This gave me a good chance to reflect on my phd years, it was a pretty fun time. I mostly did tons of architecture work, played around a lot on attention blocks, recurrence etc. Felt like a professional lego builder. I was playing on the SNLI, MNLI, Squad leaderboards etc,
-
Quantum Computing Researcher Pushes Beyond Classical Capabilities
By
–
Meet Research Scientist Alexis Morvan, PhD, who is part of the team that pushes the boundaries of quantum computing by achieving computations beyond classical capabilities, & ensures the validity of these quantum results. More about his work →https://t.co/V6Sd3s0yEW #QuantumAI pic.twitter.com/zhyPxbHczk
— Google AI (@GoogleAI) 24 janvier 2025Meet Research Scientist Alexis Morvan, PhD, who is part of the team that pushes the boundaries of quantum computing by achieving computations beyond classical capabilities, & ensures the validity of these quantum results. More about his work →
https://
goo.gle/4g8UiG7 #QuantumAI -

PhD Thesis Critique Using Gemini 2.0 Flash Thinking
By
–
Trying out this fun critique trend on my PhD thesis ("neural architectures for natural language understanding",) using gemini 2.0 flash thinking. It thinks most of my thesis is empirical and has no theoretical contribution (yes, true), no strong overarching narrative cohesion