When Nvidia said they were sending me a DGX A100 for machine learning, they didn't say I had to build a room for it! Say hello to Zubi & Oompa Loompa 🐈 btw. these are Maine Coons, they are a bit smaller than tigers. I am not prone to exaggeration. How big was the Cray-1 🤔

I really pulled out all the stops for the new version of beads_viewer (bv). The original version of bv was made in a single day and was just under 7k lines of Golang. This new version is… 80k lines. I added an insane number of great features. Your agents will love it (you, too).

I’m excited to share PROTON, a pre-trained relational transformer for neurology. 🧠 PROTON generates testable neurological hypotheses, which we evaluated across: 🧬 Parkinson’s disease: genome-wide in vitro and in vivo screens 🧫 Bipolar disorder: patient-derived brain

Well, this is a little wild. There are now over 11,000 of you here 🙊 When I started posting on X, I was completely new to AI and developer growth, trying to figure things out in real-time. I'm honestly blown away that so many of you have decided to follow along, so THANK YOU

Europe should be building one new Spotify every year. A global category leader. Built from Europe. From day one. But too many founders still build for their local market first. We’re bringing together the ones who actually think and compete global: • 3-month full lock-in in

Multi-agent AI systems are eating single-agent architectures. But most teams have no idea how to build them. As agent systems get more complex, we're seeing a clear shift from single-agent architectures (one agent handles all tasks) to multi-agent architectures that distribute

Java developers: your Weaviate workflows are about to get way cleaner ✨ After months of beta testing and community feedback, we've shipped a ground-up redesign that transforms how Java developers interact with Weaviate. Key improvements include: • 𝗧𝘂𝗰𝗸𝗲𝗱 𝗕𝘂𝗶𝗹𝗱𝗲𝗿

More details in our free context engineering eBook:

We spend hours building the perfect retrieval system, optimizing embeddings, and tuning retrieval. Then users type something vague like "how do i make this work when my api call keeps failing?" and our beautiful system breaks. The problem isn't retrieval. It's what happens

"Friends, Romans, countrymen, lend me your compute" - Shakespeare

Now imagine that you're shipping tweets from an apartment in San Francisco's Tenderloin district thanks to tech and AI

Annual reminder to make sure you pass your RFC-25-12 audit. Weaviate is ready: https://t.co/bSKu0sDBQQ

𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝘄𝗶𝗻𝗱𝗼𝘄𝘀 𝗰𝗮𝗻 𝗵𝗮𝗻𝗱𝗹𝗲 𝟭𝗠+ 𝘁𝗼𝗸𝗲𝗻𝘀. Yet agents are 𝘭𝘦𝘴𝘴 reliable than ever. We just published a deep-dive blog on context engineering for AI agents, covering everything from theory to hands-on implementation 🔽 All LLMs are constrained by

Ink is cheap. LLM tokens are expensive. So @femke_plantinga, @victorialslocum, and I wrote a long blog on Context Engineering to help reduce the token burn. It also includes a working example of an agent built using Elysia (the open source agentic RAG framework) to put the

sick of the yapping? the AI? everyone shilling random shit down your throat on CT? well, I scraped 70k users from 7 infofi platforms and built a way for you to mute them all forever Introducing: shhhh (a prototype) ↓

Let's build 𝗮𝗴𝗲𝗻𝘁 𝗺𝗲𝗺𝗼𝗿𝘆 with LlamaIndex + Weaviate + Gemini 👀 This recipe notebook walks through building persistent memory for AI agents using three components: @llama_index (orchestration), @weaviate_io (vector storage), and Gemini (LLM + embeddings). See it here:

Has Claude sped up your biology research or enabled new capabilities for your lab? Is it underperforming on specific tasks? Do you have ideas to improve its usefulness in a specific domain of biology? Share feedback here:

Delighted to share our work on “Graph AI generates neurological hypotheses validated in molecular, organoid, and clinical systems,” tomorrow at @Stanford_AI_Bio. Learn more about PROTON, a Pre-trained RelatiOnal TransfOrmer for Neurology. 📍Gates 415 or Zoom 🗓️ Tuesday,

Just released: Multi2Vec CLIP inference container 1.5.0 🎉 This release contains: - Support for facebook MetaClip2 models - Support for ModernVBERT/modernvbert-embed model - Added support for running inference container on NVIDIA Jetson devices Check out the docs to spin it

Finally a causal formulation of spike-wave duality. Over the past couple of years I have been constantly asked about causality of brain waves. After two years of spontaneously thinking about this, I dared to put this "baby" out. Shout out to my talented collaborator

this is a nice / simple pattern for agent memory. reflect over session logs, distill preferences / feedback from actual use to update memory. been doing this w/ Claude Code for ~1-2 months and very effective. write up + code: https://t.co/CTSLKgM581