In the world of humachinekind, you have the opportunity to live and work the way you want.
MACHINE LEARNING
-
AI Lucie and Albert: Millions of Public Money Wasted, Unused
By
–
And we, for public servants, have the AIs Lucie and Albert, paid largely with public money (several million euros) that were just thrown in the trash because no one uses them. COCORICO!!!
-
RL Scaling Progress: Engineering Breakthrough in Model Training
By
–
i am GIDDILY following the open progress in RL scaling notable that it takes such a herculean engineering effort to train a model for 3000 steps. but these results are undeniable, esp with shortened context also it's only been two months since v1… hard to keep up…
-
Explaining self-supervised learning for AI education
By
–
The way you explained self supervised learning should be in every AI 101 course. It makes the concept click instantly.
-
Fastest Growing Jobs by 2030: Big Data, AI, and FinTech Lead
By
–
The fastest growing jobs by 2030, with Big Data, AI, and FinTech roles leading the future of work https://
visualcapitalist.com/charted-the-wo
rlds-fastest-growing-jobs-2025-2030/
… via @visualcap #FutureofWork @HaroldSinnott @Nicochan33 @TerenceLeungSF @enricomolinari @AndrewinContact @anand_narang @segundoatdell @jblefevre60 -
DeepMind SIMA Agent Operates Inside Genie 3 World Model
By
–
Demis Hassabis décrit une configuration surréaliste : une IA jouant à l’intérieur de l’esprit d’une autre.
— VISION IA (@vision_ia) 12 août 2025
On est vraiment en train de recréer la simulation, et matrix…
L’agent SIMA de DeepMind peut désormais opérer à l’intérieur de Genie 3, et prenant des actions dans un… pic.twitter.com/rDSMuiWOQHDemis Hassabis décrit une configuration surréaliste : une IA jouant à l’intérieur de l’esprit d’une autre. On est vraiment en train de recréer la simulation, et matrix… L’agent SIMA de DeepMind peut désormais opérer à l’intérieur de Genie 3, et prenant des actions dans un
-
TorchTitan MoE Implementation with FSDP and Torch Requirements
By
–
torchtitan has an MoE impl that supports grouped mm and composes with FSDP: https://
github.com/pytorch/torcht
itan/blob/main/torchtitan/models/moe.py
… needs the latest torch version though (2.8) which flash-attn doesnt have a wheel for yet 🙁 -
Distributed Training with torchrun and FSDP Made Simple
By
–
It's pretty easy with torchrun, fsdp, etc.
-
ICML Criticized as Band-Aid Solutions for Large Language Models
By
–
ICML should be renamed ICBL (International Conference on Band-Aids for LLMs).
-
Pixel Reasoner: Open-Source VLM Reasoning Framework Breakthrough
By
–
🚀 New Breakthrough in Vision-Language Reasoning 🧠🖼️
— Dr. Debashis Dutta (@debashis_dutta) 12 août 2025
Originally shared by @WenhuChen — Pixel Reasoner is the first open-source framework enabling Vision-Language Models (VLMs) to “think in pixel space” through curiosity-driven reinforcement learning.
💡 The Challenge
Most… pic.twitter.com/IXCFVuNMeMNew Breakthrough in Vision-Language Reasoning Originally shared by @WenhuChen — Pixel Reasoner is the first open-source framework enabling Vision-Language Models (VLMs) to “think in pixel space” through curiosity-driven reinforcement learning. The Challenge Most