It’s ALL online learning in nature — no pretraining and transfer.
RESEARCH
-
New fast-learning AIs: World Models trend in 2026, Tesla leads
By
–
Yup. And new AIs are on the way that learn faster. Which is why World Models will be a huge trend in 2026. Who has the best World Model? Tesla.
-

AI and Physical Intelligence: A Historical Perspective on Embodiment
By
–
The distinction between AI and physical intelligence is now widely recognized, but that wasn't always the case. https://
withoutwhy.substack.com/p/ai-embodimen
t-and-the-limits-of-simulation
… @BScotRousse -
Multi-Fingered Hand Assembly: Engineering vs Machine Learning Trade-offs
By
–
Precision assembly is common but OK! Doing it with a multi-fingered hand is interesting: how much of this demo is due to Model-Free Learning and how much is Good Old-Fashioned Engineering (GOFE)? I'll bet it's mostly the latter. Now use it to generate data for the former…
-
Segment Anything Models Advance Flood Monitoring and Disaster Response
By
–
Our Segment Anything Models are helping advance flood monitoring and disaster response.
— AI at Meta (@AIatMeta) 22 décembre 2025
See how @USRAedu and @USGS have fine-tuned SAM to automate a key bottleneck in real-time river mapping, enabling faster, scalable, and more cost-effective disaster preparedness:… pic.twitter.com/kSHkaOKlheOur Segment Anything Models are helping advance flood monitoring and disaster response. See how @USRAedu and @USGS have fine-tuned SAM to automate a key bottleneck in real-time river mapping, enabling faster, scalable, and more cost-effective disaster preparedness:
-

GLM-4.7 Model Benchmark Performance and Competitive Analysis
By
–
BREAKING : GLM-4.7 by @Zai_org scored 42% on the HLE benchmark, representing a 38% improvement over GLM-4.6 and approaching GPT-5.1 performance. New SOTA open source on SWE?
-
Technical limitations of LLMs in specific NLP tasks
By
–
Grave et c’est surtout un des rares sujets de NLP impossible à adresser actuellement par des LLM.
-
Brain General Intelligence vs Universal Intelligence Debate
By
–
Yann is just plain incorrect here, he’s confusing general intelligence with universal intelligence.
— Demis Hassabis (@demishassabis) 22 décembre 2025
Brains are the most exquisite and complex phenomena we know of in the universe (so far), and they are in fact extremely general.
Obviously one can’t circumvent the no free lunch… https://t.co/RjeqlaP7GOYann is just plain incorrect here, he’s confusing general intelligence with universal intelligence. Brains are the most exquisite and complex phenomena we know of in the universe (so far), and they are in fact extremely general. Obviously one can’t circumvent the no free lunch
-

Research Challenges Scaling Myths for AI Agent Tool Calling
By
–
This paper quietly destroys one of the most repeated myths in AI: that bigger models are always better for agents. The researchers took a 350M-parameter model and fine-tune it specifically for agentic tool calling. Not chatting. Not creativity. Just one job: pick the right tool,
-
AI Study Reduces Heart Disease Risk in Prediabetic Patients
By
–
A study published in The Lancet Diabetes & Endocrinology by researchers at King’s College London found that lowering blood sugar significantly reduces the risk of heart disease in patients with prediabetes. https://
bit.ly/3Nb4Njg
#diabetes #health #research #TheLancet