[ 66774540.000 ] posts.x: Full talk on the S3 metadata bottleneck, GPU failure recovery, and the infrastructure behind a 12 trillion token synthetic data run:  |   [ 66774540.000 ] posts.x: Pre-training runs are hitting a wall: the scaling laws demand exponentially more data for linear gains, and the open web only holds so many…  |   [ 66770700.000 ] posts.x: 🫡  |   [ 66765360.000 ] posts.x: Full talk on why a 99% per-step success rate still produces a 36% success rate at scale, and how to design a system around that:  |   [ 66765360.000 ] posts.x: A browser agent that succeeds 99% of the time per step still looks production ready, until you chain 100 steps together: the math alone puts the…  |   [ 66750960.000 ] posts.x: Full talk on why automated code review is spreading fast across GitHub and Cursor, and where it still gets fooled:  |   [ 66750960.000 ] posts.x: Developers using autonomous coding agents wrote 741% more code in a study tracking over 100,000 GitHub developers, but the software they actually…  |   [ 66672720.000 ] posts.x: Full talk on what building a real software factory takes, from model routing to validation to context engineering:  |   [ 66672720.000 ] posts.x: Writing code turns out to be the easy part of building software. In one customer mission that ran sixteen hours straight, the validators spent forty…  |   [ 66663960.000 ] posts.x: Full talk on Warp's open source factory loop, the public dashboard, and skill loops that self-correct:  |   [ 66663960.000 ] posts.x: If it's nearly free to build software now, it's nearly free to clone too. @zachlloydtweets, founder and CEO of Warp, works through what that does to…  |   [ 66616680.000 ] posts.x: Full talk on Antigravity's sub-agent swarms, sidecars, and generative UI, including the OS kernel built by 93 sub-agents:  |   [ 66616680.000 ] posts.x: Give Messi the ball and get out of the way. That's the coaching logic @kevinhou22, engineering lead for Antigravity at Google DeepMind, applies to AI…  |   [ 66607680.000 ] posts.x: Full talk on treating a fleet of coding agents like units in an RTS, and what that changes about who burns out:  |   [ 66607680.000 ] posts.x: Scaling up coding agents doesn't remove the bottleneck, it just moves it onto you: more agents means more steering, more reviewing, more approving…  |  

[ 66774540.000 ] posts.x: Full talk on the S3 metadata bottleneck, GPU failure recovery, and the infrastructure behind a 12 trillion token synthetic data run:  |   [ 66774540.000 ] posts.x: Pre-training runs are hitting a wall: the scaling laws demand exponentially more data for linear gains, and the open web only holds so many…  |   [ 66770700.000 ] posts.x: 🫡  |   [ 66765360.000 ] posts.x: Full talk on why a 99% per-step success rate still produces a 36% success rate at scale, and how to design a system around that:  |   [ 66765360.000 ] posts.x: A browser agent that succeeds 99% of the time per step still looks production ready, until you chain 100 steps together: the math alone puts the…  |   [ 66750960.000 ] posts.x: Full talk on why automated code review is spreading fast across GitHub and Cursor, and where it still gets fooled:  |   [ 66750960.000 ] posts.x: Developers using autonomous coding agents wrote 741% more code in a study tracking over 100,000 GitHub developers, but the software they actually…  |   [ 66672720.000 ] posts.x: Full talk on what building a real software factory takes, from model routing to validation to context engineering:  |   [ 66672720.000 ] posts.x: Writing code turns out to be the easy part of building software. In one customer mission that ran sixteen hours straight, the validators spent forty…  |   [ 66663960.000 ] posts.x: Full talk on Warp's open source factory loop, the public dashboard, and skill loops that self-correct:  |   [ 66663960.000 ] posts.x: If it's nearly free to build software now, it's nearly free to clone too. @zachlloydtweets, founder and CEO of Warp, works through what that does to…  |   [ 66616680.000 ] posts.x: Full talk on Antigravity's sub-agent swarms, sidecars, and generative UI, including the OS kernel built by 93 sub-agents:  |   [ 66616680.000 ] posts.x: Give Messi the ball and get out of the way. That's the coaching logic @kevinhou22, engineering lead for Antigravity at Google DeepMind, applies to AI…  |   [ 66607680.000 ] posts.x: Full talk on treating a fleet of coding agents like units in an RTS, and what that changes about who burns out:  |   [ 66607680.000 ] posts.x: Scaling up coding agents doesn't remove the bottleneck, it just moves it onto you: more agents means more steering, more reviewing, more approving…  |