"Move 37" is the word-of-day - it's when an AI, trained via the trial-and-error process of reinforcement learning, discovers actions that are new, surprising, and secretly brilliant even to expert humans. It is a magical, just slightly unnerving, emergent phenomenon only

This is wild - UC Berkeley shows that a tiny 1.5B model beats o1-preview on math by RL! They applied simple RL to Deepseek-R1-Distilled-Qwen-1.5B on 40K math problems, trained at 8K context, then scaled to 16K & 24K. 3,800 A100 hours ($4,500) to beat o1-preview in math! Best

Deepseek R1 inference in pure JAX! Currently on TPU, with GPU and distilled models in-progress. Features MLA-style attention, expert/tensor parallelism & int8 quantization. Contributions welcome!

Becks made me fall in love with football & Manchester United in 1996, which led me to play fantasy football and then create Dream11. Life came full circle - now we co-own a football team Salford City FC along with Gary Neville, Scholes, Giggs, etc. If I'm 'Dream'ing, don't ever wake me up!

i work at meta. hr systems. mostly comp processing. quiet job. stable. sometimes i daydream about retiring in portugal. today a package hit my queue. base + bonus + equity. looks normal at first glance. then i open the details. $1,000,000,000 over four years. plus signing. plus

AI games are going to be amazing (sound on)

This is going to be bad for me 😓

openai.com/index/inside-o…

This article resonates with I've been going through since Opus 4.6 I've been testing end-to-end app features just with prompts. Inflection point in Knowledge based skills industry. x.com/i/status/20212…

My biggest takeaways from Sherwin Wu: 1. AI is writing virtually all code at OpenAI. 95% of the engineers use Codex, and engineers who embrace these tools open 70% more pull requests than their peers, and that gap is widening over time. 2. The role of a software engineer is

Shocking new MRI study on 3–5 year olds: Just 2 hours of interactive screen time per day is linked to measurable loss of white matter in the brain. Professor Mike Nagel (University of the Sunshine Coast): “White matter is myelin — it insulates axons like plastic on a wire.

Working with Codex 5.3 & Opus 4.6 - noticed Opus is better but has an interesting quirk: it rewrites full modules vs Codex's incremental approach. Token usage difference is real. Same experience or just me? 🤔

Had claude analyze my entire Downloads folder (1,156 files, 8+ years) and it found patterns i didn't see myself like how my triathlon and my day job in Sports AI are basically the same thing now! if you have years of files piled up, try! this is wild!

Our latest Claude Code hackathon is officially a wrap. 500 builders spent a week exploring what they could do with Opus 4.6 and Claude Code. Meet the winners:

Free throw shooting isn't just physical — it's mental. 🏀 This study on simulated mental practice techniques shows significant accuracy gains in highly skilled players. Worth a read for coaches & athletes. researchgate.net/publication/33…

anthropic.com/research/labor… A picture is worth a 1000 words.

Based on everything explored in the source code, here's the full technical recipe behind Claude Code's memory architecture: [shared by claude code] Claude Code’s memory system is actually insanely well-designed. It isn't like “store everything” but constrained, structured and

I reverse-engineered Claude Code's leaked source against billions of tokens of my own agent logs. Turns out Anthropic is aware of CC hallucination/laziness, and the fixes are gated to employees only. Here's the report and CLAUDE.md you need to bypass employee verification:👇