ICYMI — Part V of our rubric series looks ahead at how AI systems will be evaluated as they become more agentic, multi-turn, and tool-using.
Rubrics + evaluator ensembles + simulation environments = quality and reliability https://
snorkel.ai/blog/part-v-fu
ture-direction-and-emerging-trends/
…
#AI #Evaluation #AgenticAI
RESEARCH
-
Evaluating Agentic AI Systems: Rubrics, Ensembles, and Simulation
By
–
-
Eric Schmidt Warns About AI’s Transformative Power Within 5 Years
By
–
L’ancien PDG de Google, Eric Schmidt, lâche un avertissement glaçant sur l’avenir de l’IA.
— VISION IA (@vision_ia) 30 décembre 2025
« D’ici 5 ans, l’IA pourrait gérer un contexte infini, un raisonnement en chaîne avec des solutions en 1 000 étapes, et des millions d’agents travaillant ensemble.À terme, elles… pic.twitter.com/Jg7O1UtV7iThe former Google CEO, Eric Schmidt, issues a chilling warning about the future of AI. "Within 5 years, AI could handle an infinite context, chain-of-thought reasoning with solutions in 1,000 steps, and millions of agents working together. In the long run, they will develop
-
Humachineology Studies on Human Machine Evolution Advances
By
–
Humachineology studies provide you with advances, innovations, and trends in the evolution of humans and machines.
-
Robert Scoble interviews Leon Ming about Chinese AI competition
By
–
The AI competition from the Chinese side of the fence.
— Robert Scoble (@Scobleizer) 29 décembre 2025
Leon Ming, @LEON_0xx0, cofounder of @youware joins me for a long conversation about what's happening in China (he is one of the top entreprenurs there, and also was a cofounder of @capcutapp, among others).
YouWare is one… pic.twitter.com/A2dqyuBsU6The AI competition from the Chinese side of the fence. Leon Ming, @LEON_0xx0, cofounder of @youware joins me for a long conversation about what's happening in China (he is one of the top entreprenurs there, and also was a cofounder of @capcutapp
, among others). YouWare is one -
Donation Strategy for MIRI and Lightcone Based on Preference
By
–
If you like one of MIRI and Lightcone but not the other, donate to the one you like. If you like one at 8 of 10 and the other at 2 of 10 but still positive on both, split donations proportionally to emotion. (There are galaxy-brained arguments against this. I am going to note
-
Lightcone and MIRI both essential for human survival and community
By
–
On my read of things, both Lightcone and MIRI are necessary. If the world will be saved at all, it will be saved by people, and those people need community to stay alive and sane and in contact with each other.
-
MIRI’s Role in AI Alignment Research and Theory
By
–
MIRI is the organization that I founded in 2000. In the old days it did most of the deep theory on alignment research, if you filter out the wacky optimistic hopium. MIRI is approximately the only organized group on the planet where you can ask "So what goes wrong if you just
-
MIRI and Lightcone Final Fundraiser Push Recommendation
By
–
MIRI and Lightcone's SFF-matched fundraisers are coming down to the wire the next 3 days (as such often do). MIRI has $1M of $1.6M remaining on its match. Lightcone needs $1.6M of $2.6M. **If** you're equally motivated to donate to either, I'd say fill Lightcone's first.
-
Mixture of Experts Architecture Powering Frontier AI Models
By
–
Why are the world’s leading models built on mixture of experts?
— NVIDIA AI (@NVIDIAAI) 29 décembre 2025
Ian Buck, Vice President of Hyperscale and HPC at NVIDIA, unpacks the architecture powering today’s frontier AI and how extreme co-design is driving smarter models at lower cost.
🎧 Listen to the NVIDIA AI Podcast… pic.twitter.com/DBMFNDAIQ0Why are the world’s leading models built on mixture of experts? Ian Buck, Vice President of Hyperscale and HPC at NVIDIA, unpacks the architecture powering today’s frontier AI and how extreme co-design is driving smarter models at lower cost. Listen to the NVIDIA AI Podcast
-
VLMs for Robotics: The Challenge of Combining Vision and Manipulation
By
–
Also #3 strikes a chord — VLMs are fantastic and super useful for robotics but I don’t think we’ve figured out yet how to best combine VL and manipulation signals.