Top KOLs
Top 20 research KOLs by dynamic heat, with categories and recent high-signal tweets.
- 01
Jerry Liu@jerryjliu01147 tweets11K engagementfounderI built a Claude Code skill that allows it to generate a deep research report over any collection of complex docs (PDFs, Word, Pptx)….and generate word-level citations and bounding boxes directly b...This is a great tutorial (credits @itsclelia + @lancedb) on how to build a practical retrieval pipeline that integrates directly with your agent harness. 1. Ingest a massive pile of docs with litep...Hosted filesystems for agents are the new RAG - 02
andrew chen@andrewchen1076 tweets8.1K engagementinvestornoticing a trend of startups replacing standard resumes/interviews with week-long (or at least 3-day weekend) in-office trials. Makes sense in a world of AI-generated resumes and interview response...hosting zoom office hours this week for a16z speedrun you'll be able to apply to a16z speedrun in just a few weeks, so so come AMA. first come, first served drop a comment + like and I'll DM you a ...prev startup advice: do things that don't scale 2026 update: do things that don't scale, then build AI agents to scale them - 03
Dario Amodei@darioamodei1043 tweets23K engagementfounderI’m proud that so many of the world’s leading companies have joined us for Project Glasswing to confront the cyber threat posed by increasingly capable AI systems head-on. https://x.com/AnthropicAI...The dangers of getting this wrong are obvious, but if we get it right, there is a real opportunity to create a fundamentally more secure internet and world than we had before the advent of AI-power...Cyber is the first clear and present danger from frontier AI models, but it won’t be the last. If we are able to collectively rise to the challenge and confront this risk, it could serve as a bluep... - 04
Simon Willison@simonw1025 tweets7.9K engagementbuilderVibe check on Gemma 4 now that we've had a few days to play with it - how does it hold up against Qwen 3.5?754B parameters, 1.51TB on Hugging FaceWrote up some thoughts on Anthropic's Project Glassing, where their latest Opus-beating model is available to partnered security research organizations only Given recent alarm bells raised by credi... - 05
Unsloth AI@UnslothAI1022 tweets29K engagementbuilderYou can now fine-tune Gemma 4 with our free notebooks! 🔥 You just need 8GB VRAM to train Gemma 4 locally! Unsloth trains Gemma4 1.5x faster with 50% less VRAM. GitHub: https://github.com/unslothai...GLM-5.1 can now be run locally!🔥 GLM-5.1 is a new open model for SOTA agentic coding & chat. We shrank the 744B model from 1.65TB to 220GB (-86%) via Dynamic 2-bit. Runs on a 256GB Mac or RAM/... - 06
- 09
Kevin Weil 🇺🇸@kevinweil934 tweets4.3K engagementfounder💥 New in Prism today: Paper Review, an AI workflow for reviewing technical and scientific papers. This is the opposite of AI slop: we're using AI to improve scientific rigor, correctness, and repr...Btw: @hemal built this as a simple skill in a few hours, since Prism is powered by Codex. More to come. And if you love the idea of building AI-native products to accelerate science, we're hiring a...Try it out at http://prism.openai.com today and let me know what you think! - 10
- 13
François Chollet@fchollet862 tweets4.9K engagementresearcher - 14
Romain Huet@romainhuet865 tweets1.2K engagementbuilderReally loved this conversation with @ashebytes. Ashe is one of the most original and creative builders in the community. We talked about building in public, relational intelligence, second brains, ...Big milestone for Codex: 3 million weekly builders! Grateful for everyone building with us, pushing us to make it better, and making this community so special. We’re celebrating the Codex way: refr...Really enjoyed joining @petergyang with @embirico to talk about how we’ve been building Codex at OpenAI. We show a live demo of the Codex app and go behind the scenes. What’s been striking: role li... - 15
Julien Chaumond@julien_c841 tweets5.8K engagementfounder - 16
Mustafa Suleyman@mustafasuleyman834 tweets1.4K engagementfounderAnother SOTA model drop! This time from the @Bing team: meet Harrier, a new open-source embedding model with state-of-the-art performance and the #1 spot on the industry standard multilingual MTEB-...If building humanist AI with unlimited tokens sounds up your alley, we're hiring at MAI. All the benefits of startup speed + agility w/the resources of one of the world's leading tech companies. If...Bing's web grounding already powers almost every major AI chatbot today. With Harrier, it just got a big upgrade for the agentic era. Better embeddings lead to better retrieval, often more accurate... - 17
Bindu Reddy@bindureddy834 tweets1.3K engagementbuilderThe best way to make cheap models work is to have big models direct them Have an expensive model like GPT 5.4 or Opus write up a derailed spec Use Kimi or GLM 5 to implement it. We are observing so...Agent Swarms - A Master Agent will create, manage and modify worker agents to get huge projects done Entire SaaS apps with all the bells and whistles can be built with Agent Swarms 🔥 Insanely exci...GLM 5.1 Is Out - Open Source Is On A Roll 🚀🚀 We have seen adoption of small open source models EXPLODE....GLM 5.1 helps accelerate this even more Coming to LiveBench today https://t.co/maivd3farC - 18
Felix Rieseberg@felixrieseberg823 tweets1.8K engagementbuilderLet's talk about Claude Mythos Preview. It's pretty hard to overstate what a step function change this model has been inside Anthropic. Its ability to identify security vulnerabilities feels like a...This chart shows an evaluation testing models’ ability to successfully exploit vulnerabilities in Firefox 147. There are three grade levels: 0 for no progress, 0.5 for partial control (controlled c...https://t.co/dwn46QK4Yj - 19
Bret Taylor@btaylor782 tweets2K engagementfounderHere’s a demo of a healthcare payment phone call with the new PCI compliant payments functionality in Sierra https://t.co/R1NX49LTwFIn a first for the conversational AI industry, Sierra agents are now Level 1 PCI compliant, which means your agent can collect card and ACH payments securely, directly in the conversation. This is ... - 20