x.com/i/article/2010…

Google has open-sourced the LangExtract Python library. It turns unstructured documents into structured data, with clear source references for each result. 100% open-source.

x.com/i/article/2012…

Fun visit to the Yorkshire #Amiga Group today 👾

Never thought a VR headset’s design could have me this obsessed but here we are. Concept by moonshake3d

OCR-tuned VLMs are getting really good AND cheap, really quickly This is a cool release from LightOn Good and cheap models are starting to saturate OlmOCR-bench. There's a lot of greenfield opportunity in curating an even more complex, diversified benchmark (lots of complex

AI-Induced Code Entropy: The Hidden Cost of Unchecked AI Acceleration In the rush to adopt AI coding tools, teams celebrate 5×–10× faster feature delivery and skyrocketing velocity metrics. Yet beneath the surface, a subtler but more dangerous phenomenon is emerging: AI-induced

x.com/i/article/2015…

x.com/i/article/2019…

x.com/i/article/2019…

Recursive Language Models (RLMs) let agents manage 10M+ tokens by delegating tasks recursively. This Google Cloud Community Article explains why ADK was the perfect choice for re-implementing the original RLM codebase in a more enterprise-ready format →goo.gle/4kjT12E

x.com/i/article/2020…

Péter Szilágyi The answer is to not let LLM touch secrets at all. The harness should keep it controlled and away from tools and inference engine and policy around secret <> domain. Working on this here - github.com/nearai/ironclaw

x.com/i/article/2021…

x.com/i/article/2021…

wait this is fucking AWESOME lol - dude fused an AI model *directly* into silicon which means this thing answers before you've even hit enter - processes 17,000 tokens PER SECOND 😂 for context, openai's quickest model does 1000 tokens per sec (i believe) - this is 17X faster -

I build AI for a living. I believe in what we're building. But this kind of rhetoric makes my work harder and more dangerous. Sam Altman, comparing human development to model training is tone-deaf, strategically reckless. People are losing jobs. They're getting angry. They're

built my own personal assistent device that runs OpenClaw. I was curious what the smallest form factor could be that fits in my pocket so I wanted to use the Pi Zero W. Works via Push to Talk->Transcribe->Sends to OpenClaw and streams the response back.