Thorne 🌸 - Play Cozysters! There is a wide gulf between "every line is bespokely hand-written" and "yolo AI-generated code you don't understand into production" where the actual ideal exists
AI/ML Feed: Blend
by @smcgrath.phd
This feeds shows blended Machine Learning/AI posts (new & popular) from users on the ML Ranked Feed List created by @smcgrath.phd. The feed is whitelisted (let me know if I am missing anyone important). A RegEx and Natural Language node is in place to try and keep the posts relevant to the theme
Pull to refresh
Thorne 🌸 - Play Cozysters! 
Alessandro Crimi Did you Flexray already!!! "#AI Automatically Color-Codes Bones and Organs in X-rays" it Identifies 60 Anatomical Structures and Learns by Creating Simulated #Xray Images with Ground Truth from Existing #CT Scans Paper, code, etc. all here: flexray.csail.mit.edu
Julian Togelius New blog post: A.I.: Anti-Intellectualism togelius.blogspot.com/2026/10/ai-a...
Siobhán 1 / 6
My favorite parts of Bender & Koller (2020), in which they defend their claim — that LLMs can’t learn meaning — by explaining that it should not be expected to apply to actual LLM training processes.
JD Long TIL about LLM “abliteration”. You should too. tl;dr - post training that strips the weights that form guardrails. Django Unchained, if you will. abliteration.ai/blog/introdu...
Benjamin Riley These are indeed the right questions. We have empirical evidence that strongly suggests the immediate impact of AI is seriously harmful to learning, and since learning is cumulative, likely will have long-term harmful consequences. (That purple area is the direct measure of harm)
Pekka Lund Tao seems to be edging closer to the 'I don't want AI even if it cured cancer' level of argumentation.
Patrick Ohlson Looks promising Claude and Fable 5.1 as orchestrator Running a synthetic training of a system one model Target is Rhino No human involved so far