Simulating an 80-year-old with @MIT_AgeLab
's gear, @AmyDMarcus learns about the importance of balance and strength training https://
wsj.com/health/wellnes
s/mit-age-simulation-suit-678afa33?st=RSM3BW&reflink=desktopwebshare_permalink
…
RESEARCH
-
MIT Age Simulation Suit Highlights Balance and Strength Training
By
–
-
Agent Evals Offline Tests Real World Success Correlation
By
–
Fair, it's all definitions 🙂 But evals in the sense that most product builders think of them (static offline envs/test sets you run an agent version against) tend not to correlate super well to real-world success
-
Hallucinations explained: broken incentives, not broken models
By
–
This might be the cleanest explanation of hallucinations I’ve seen. It’s not about broken models.
It’s about broken incentives. -

Agentic AI Adoption Expected to Double by 2027 with Responsible Governance
By
–
Northeastern University: 25% of generative-AI users will pilot agentic AI in 2025; adoption will double by 2027; embedding responsible governance is crucial. #ResponsibleAI #AgenticAI #IoT #5G https://
ai.northeastern.edu/news/agentic-a
i-in-2025-a-responsible-ai-expert-cuts-through-the-chaos
… -
God Made Symmetries Differential Equations Human Invention
By
–
God made symmetries. Differential equations are a human invention.
-
Anthropic Hires Google Books Project Partnership Former Head
By
–
Anthropic actually hired the "former head of partnerships for Google's book-scanning project" for this
-
Flow Matching for Video Generation: Data Scale and Generalization Benefits
By
–
I loved this research paper on Flow Matching, the most popular approach for video gen. TLDR: More data means harder to fit any specific data (say image), better generalisation, greater coverage of learned concepts. Pretraining: Use Billions of data to learn many things and none
-

New Math and Reasoning Benchmark Released for AI Evaluation
By
–
New math and reasoning benchmark. Highly needed because most of the current ones are nearly or completely saturated.
-
Document value in model training data
By
–
One thing I've never understood is how much value an individual document has to training a model – are there single books that, if included in the training data, would
materially improve the resulting model? -
Fair Use and Book Scanning for AI Training: Legal Gray Areas
By
–
To clarify: it appears it is legal, at least in the USA, to buy a used copy of a physical book (used = the author gets nothing), chop the spine off, scan the pages, discard the paper copy and then train on the scanned content The transformation from paper to scan is "fair use"