👨🏻💻 Lead Data Scientist ➿ code junkie ➿ breaking stuff with AI ➿ all chaos lives on my github👇github.com/genieincodebot… p(wₜ | w₁:ₜ₋₁) Joined February 2024
Log probability is one of the most underutilised parts of an LLM response.
When used properly, it can improve results & provide interpretability in otherwise black box, closed source LLMs.
Unfortunately, not all major models expose it, Claude doesn’t (likely for their safety &…
Our new Threat Intelligence report details how we’ve identified and disrupted sophisticated attempts to use Claude for cybercrime.
We describe a fraudulent employment scheme from North Korea, the sale of AI-created ransomware by someone with only basic coding skills, and more.
We’ve developed Claude for Chrome, where Claude works directly in your browser and takes actions on your behalf.
We’re releasing it at first as a research preview to 1,000 users, so we can gather real-world insights on how it’s used.
Pretty cool that they open sourced the actual full-sized production model.
Here’s the Grok 2.5 architecture overview next to a roughly similarly sized Qwen3 model.
The MoE residual is quite interesting. Kind of like a shared expert. I don't think I've seen this setup before.
Pretty cool that they open sourced the actual full-sized production model.
Here’s the Grok 2.5 architecture overview next to a roughly similarly sized Qwen3 model.
The MoE residual is quite interesting. Kind of like a shared expert. I don't think I've seen this setup before. https://t.co/3t3c2buu3A
The @xai Grok 2.5 model, which was our best model last year, is now open source.
Grok 3 will be made open source in about 6 months.
huggingface.co/xai-org/grok-2
New Anthropic research: filtering out dangerous information at pretraining.
We’re experimenting with ways to remove information about chemical, biological, radiological and nuclear (CBRN) weapons from our models’ training data without affecting performance on harmless tasks.
Introducing Gemma 3 270M 🔥
🤏A tiny model! Just 270 million parameters
🧠 Very strong instruction following
🤖 Fine-tune in just a few minutes, with a large vocabulary to serve as a high-quality foundation
developers.googleblog.com/en/introducing…
Pro tip for Sam:
Never show an Asian face on the OpenAI livestream.
Meta’s running a face detection on every frame. If it spots an Asian, 9-figure offer is instantly sent from Zuck's personal email.
1/n I’m thrilled to share that our @OpenAI reasoning system scored high enough to achieve gold 🥇🥇 in one of the world’s top programming competitions - the 2025 International Olympiad in Informatics (IOI) - placing first among AI participants! 👨💻👨💻
1K Followers 3K FollowingNews/Updates that can Impact YOU
NO Left. NO Right. FULLY Unbiased NEWS on X
Politics, Wars, Tech, World Affairs... EVERYTHING
3K Followers 2K Followingfounder @hyperspell (YC F25) | context and memory for AI agents | hacker house life @mission__ctrl | @JoinODF | investing @weekendfund @aforevc | AuDHDer | DJ
24K Followers 6K FollowingCreating tools to help people work smarter. Built yfinance (3M downloads/mo). CTO as a Service @automazeio
My newsletter 👉 https://t.co/rIRlqaXIcZ
9 Followers 60 FollowingBe your own sun, light yourself up and warm others.I came here to find more like-minded friends to share and communicate together!
4K Followers 2K Followingconnecting police & catching criminals in 196 countries for more than century, tracking of scammers to get back lost funds. https://t.co/Ym8M2LGjhe
148 Followers 5K FollowingWhite House Tech Support 💻👮♂️| D.O.G.E🚀🇺🇸| Making the world great again with innovations like never before… Tesla 🚘 Cybertruck📐
3K Followers 2K Followingfounder @hyperspell (YC F25) | context and memory for AI agents | hacker house life @mission__ctrl | @JoinODF | investing @weekendfund @aforevc | AuDHDer | DJ
20K Followers 1K FollowingResearcher @MSFTResearch, AI Frontiers Lab; Prof @UWMadison (on leave); learning in context; thinking about reasoning; babas of Inez Lily.
24K Followers 6K FollowingCreating tools to help people work smarter. Built yfinance (3M downloads/mo). CTO as a Service @automazeio
My newsletter 👉 https://t.co/rIRlqaXIcZ
109K Followers 1 FollowingClaude is an AI assistant built by @anthropicai to be safe, accurate, and secure. Talk to Claude on https://t.co/ZhTwG8dz3D or download the app.
109 Followers 285 FollowingBuilding @AskYourVideoPro | Love shipping cool stuff and learning about backend, llm, ai agent, rag, mcp, fine-tuning and more
489K Followers 146 FollowingNobel Laureate. Co-Founder & CEO @GoogleDeepMind - working on AGI. Solving disease @IsomorphicLabs. Trying to understand the fundamental nature of reality.
86K Followers 1K Followingi help make https://t.co/jZh799yNH4, the best AI for self-improvement, introspection, and emotional processing. https://t.co/ac0cp4UZ9h
25 Followers 219 FollowingJava & Spring Boot Dev | Creating AI tools for better mental health | Therapy AI — where tech meets empathy | 🌱 Coding with purpose | BTech’26
15K Followers 1K FollowingCo-founder and CEO @Hyperbolic_Labs. ex-@avax & ex-@citsecurities. Finished Math PhD in 2yrs @UCBerkeley. Math Olympiad Gold Medalist. Highest honor @PKU1898
30K Followers 123 FollowingMechanistic Interpretability lead DeepMind. Formerly @AnthropicAI, independent. In this to reduce AI X-risk. Neural networks can be understood, let's go do it!
50K Followers 883 FollowingAssistant professor (of mathematics) at the University of Toronto. Algebraic geometry, number theory, forever distracted and confused, etc. He/him.