It is generally frowned upon to have LLMs precisely regurgitate part of their training set, but it is an interesting question how you could use LLM training to nearly losslesly compress a huge corpus like the entirety of the Internet Archive. The Hutter Prize is for perfect
RESEARCH
-
GE-Sim 2.0: Advanced World Simulator for Robotic Manipulation
By
–
From my sponsor:#AGIBOTAIWeek Day4:
— Robert Scoble (@Scobleizer) 15 avril 2026
Introducing GE-Sim 2.0 (Genie Envisioner World Simulator 2.0), a comprehensive embodied world simulator for robotic manipulation.
By integrating future video generation, proprioceptive state estimation, and reward-based policy assessment,… pic.twitter.com/krBCH4laR2From my sponsor: #AGIBOTAIWeek Day4: Introducing GE-Sim 2.0 (Genie Envisioner World Simulator 2.0), a comprehensive embodied world simulator for robotic manipulation. By integrating future video generation, proprioceptive state estimation, and reward-based policy assessment,
-
Understanding vs Results: The Economics of AI Research
By
–
Scientists think it’s understanding. Everyone else thinks it’s getting the right answer. The latter pay the former.
-
Social Sciences Versus Natural Sciences in AI Research
By
–
The social sciences are quicksand. The natural sciences are bedrock.
-
Switching to Hermes LLM from Nous Research
By
–
We switched over to using Hermes from @NousResearch
. Talked about that on the space last week. Here: https://
x.com/Scobleizer/sta
tus/2042009419094106350?s=20
… Well worth listening to. -
Pioneer Who Transformed Neural Processing into Bayesian Framework
By
–
The man who turned NIPS into BIPS (Neural Information Processing Systems into Bayesian ones). Off-the-charts smart. https://t.co/kb7u5cyEuq
— Pedro Domingos (@pmddomingos) 15 avril 2026The man who turned NIPS into BIPS (Neural Information Processing Systems into Bayesian ones). Off-the-charts smart.
-
Future Guilds and AI Knowledge Protection Dilemmas
By
–
The 21st century will be dominated by guilds that murder any member who discloses their knowledge to AI.
-
Xerox PARC: Similar Stories of Untapped Innovation
By
–
Yeah! I have so many similar stories. Like Xerox PARC.
-
Jagged Intelligence and the Implications for Frontier Model Prompting
By
–
AI model that wins gold at the Math Olympiad can't read a clock. Stanford has a name for this: "jagged intelligence." And their 2026 AI Index proves it changes everything about how you should prompt. Frontier models now score above PhD-level on science benchmarks and dominate