Awesome blogpost! RL Scaling Laws for LLMs How scaling laws have evolved from pretraining to reinforcement learning… https://
cameronrwolfe.substack.com/p/rl-scaling-l
aws
…
RESEARCH
-
RL Scaling Laws for LLMs: From Pretraining to RL
By
–
-
Presentation of the paper ‘Agents’ Last Exam’ on arXiv
By
–
—
Agents' Last Exam Paper: https://
arxiv.org/abs/2606.05405
— -
Agents’ Last Exam: Benchmarking AI on real-world professional tasks for economy
By
–
Are AI benchmarks really measuring what matters for the economy? Enter Agents’ Last Exam (ALE) — a benchmark that tests AI agents on long, real-world professional tasks, not just puzzles. It covers 1,000+ tasks across 55 fields mapped to U.S. job classifications. The
-
Claude Fable 5 claims #1 on DeepSWE Bench with 66% Pass@1
By
–
Claude Fable 5 has debuted on DeepSWE Bench with a 66% Pass@1, claiming the #1 spot and edging out GPT-5.5. The result reinforces a broader trend across recent coding benchmarks: strong raw performance combined with consistent reliability and efficiency in real-world software
-
Gemini for Science Could Accelerate Next Big Breakthrough
By
–
Gemini for Science Could Accelerate the Next Big Scientific Breakthrough
— Ronald van Loon (@Ronald_vanLoon) 12 juin 2026
by @GoogleDeepMind#ArtificialIntelligence #MachineLearning #EmergingTech #FutureTech pic.twitter.com/mTaw3V7quUGemini for Science Could Accelerate the Next Big Scientific Breakthrough
by @GoogleDeepMind #ArtificialIntelligence #MachineLearning #EmergingTech #FutureTech -
Anthropic needs certification department for powerful model users
By
–
I think Anthropic needs to build a certification department that audits and approves users of powerful models. Computer security companies, biotech researchers, academic labs, doctors, government institutions need access to the best AI we can build.
-
AI’s productivity paradox in science and creative solutions
By
–
What happens when AI makes science too productive? @StanfordHAI Senior Fellow @SuryaGanguli talks about creative solutions for addressing AI's productivity paradox. Read more insights from the AI+Science conference: https://t.co/X9kKtL1oKo pic.twitter.com/c0CZL5sYjw
— Stanford HAI (@StanfordHAI) 12 juin 2026What happens when AI makes science too productive? @StanfordHAI Senior Fellow @SuryaGanguli talks about creative solutions for addressing AI's productivity paradox. Read more insights from the AI+Science conference: https://
hai.stanford.edu/news/how-ai-is
-transforming-scientific-discovery-while-keeping-humans-at-the-center
… -
LCLMs start with compression, not retrieval, for agent memory
By
–
Most agent systems start with retrieval. LCLMs start with compression. Compress everything once. Reason globally. Expand locally when needed. That's a very different way to think about agent memory.
-
Skepticism over fuss about 6-point benchmark difference
By
–
obviously different benchmarks will yield different results and maybe saturation is an issue and maybe this and maybe that …. but … all this fuss over … 6 points?
-
Fast KV Compaction reduces LLM memory by 50x without loss
By
–
What if you could reduce the memory of a language model by 50x in seconds without losing performance? Researchers from MIT present Fast KV Compaction via Attention Matching. They build compact key-value caches in the latent space that preserve