The Gemini Pro models do not seem to be iterating anywhere near as quickly as Claude or GPT (last release was 3.1 Pro in February). Its causing a growing performance gap between Google and the other two labs, and the Gemini 3.5 Flash model, good as it is, doesn't close it much.
RESEARCH
-
Yann LeCun clarifies his role: ran FAIR, not involved in Llama
By
–
I didn't "run Meta's AI initiatives."
I ran FAIR (the fundamental AI research lab) from 2014 to 2018.
Then, I became an IC, and returned to research. I was nobody's boss.
Iwas never involved in LLM research, and made zero technical contributions to Llama. I merely cheered from -
AutoScientists: open-source AI team doing real science
By
–
AI agents just started doing real science, not just answering questions.
— AlphaSignal AI (@AlphaSignalAI) 6 juin 2026
Most AI research agents follow one path or take orders from a central planner.
They forget what failed and stop improving once they plateau.
AutoScientists works differently.
It is an open-source team… pic.twitter.com/Ou6x2ONkq1AI agents just started doing real science, not just answering questions. Most AI research agents follow one path or take orders from a central planner. They forget what failed and stop improving once they plateau. AutoScientists works differently. It is an open-source team
-
Lack of useful tasks for 330,000 H100s questions scaling hypothesis
By
–
fair point that the contracts aren’t long term, but it still doesn’t speak well for their confidence in their original scaling hypothesis if they can’t find something useful to do with 330,000 H100s.
-
Claude 5 Mythos release contingent on GPT-5.6, next week expected
By
–
Under no circumstances will Claude 5 Mythos be released without GPT-5.6 being released in the same week. I am now firmly convinced that next week will be release week.
-
AI-detected artery inflammation dramatically increases cardiac mortality risk
By
–
The study cited in the article that I considered a wake-up call found a 13-fold risk of cardiac mortality with 1 inflamed artery, which rose to nearly 30-fold when all 3 arteries were inflamed, as determined by AI of non-invasive CT angio imaging https://
thelancet.com/article/S0140-
6736(24)00596-8/fulltext
… -
AlphaProof vs LeanMarathon: proof search vs fidelity maintenance
By
–
The important distinction: AlphaProof-style systems show how AI can search for formal proofs. LeanMarathon asks how an AI system can preserve target fidelity across an entire research-level Lean development. That is a different problem: not just proving, but maintaining a
-
LeanMarathon: Building reliable AI co-mathematicians via long proofs
By
–
AI co-mathematicians will not be built by making one prover cleverer. They will be built by making long proofs survive time. A new paper by Yuanhe Zhang, Yuekai Sun, Taiji Suzuki, Jason D. Lee, and Fanghui Liu introduces LeanMarathon: LeanMarathon: Toward Reliable AI
-
Test of distilled and optimized models for this infrastructure with Nemotron
By
–
Afterwards I'll only take distilled and optimized ones for this infrastructure! I'm going to test with Nemotron.
-
New Claude Mythos 5 Model Slug Spotted; New Model Class Coming?
By
–
BREAKING : A new Claude Mythos 5 model slug has been spotted via Dev Mode. Claude Mythos is planned to be released as its own model class, besides Haiku, Sonnet and Opus model families. Soon?