Btw some of you might find this space interesting
LLMS
-
Explaining Why Sora is Not a Proper World Model
By
–
Search my tweets (or better yet my substack) w sora and world models for endless explanations for why Sora was NOT a proper world model.
-
Custom Entity Recognition Accelerated with Claude and Amazon Bedrock
By
–
Accelerating Custom Entity Recognition with Claude Tool use in Amazon Bedrock! #BigData #Analytics #DataScience #AI #MachineLearning #NLProc #LLM #IoT #IIoT #PyTorch #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux
-
Google Launches WebSim-Like Tool Powered by Gemini Flash
By
–
Google just made their own version of WebSim using Gemini 3.1 Flash-Lite (but without the actually memory). https://t.co/IOETZk9RBj
— Matt Wolfe (@mreflow) 25 mars 2026Google just made their own version of WebSim using Gemini 3.1 Flash-Lite (but without the actually memory).
-
Software Review in the Era of AI: What OpenSci Tests
By
–
Software Review in the Era of AI: What We Are Testing at rOpenSci https://
buff.ly/cBRP7hp
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
Commentary on AI-generated text content
By
–
Please read up on history and Muslim rule in Europe from 711 to 1492 that spanned 781 years or almost a millenium
-
Hyperagents: New Research Paper on AI Agents
By
–
Hyperagents Zhang et al.: https://
arxiv.org/abs/2603.19461 #ArtificialIntelligence #AIAgents -
ChatGPT Pro and Codex introduction opportunity
By
–
Chat, who can intro me to them? would love to hook them up w/ ChatGPT Pro sub & introduce them to Codex too!
-
Large Scale Training Infrastructure Kernels and Distributed RL Scaling
By
–
We go into detail about the infrastructure behind large scale training including the kernels we developed and open-sourced for the project. We also discuss distributed training and environment scaling for RL.
-
Reinforcement Learning Phase Critical for Model Performance
By
–
The reinforcement learning phase is critical for final performance. We discuss the algorithms we apply for this stage. We find that simple approaches often work best, and improve performance broadly.