Daily Brief

If you run Claude and Codex side by side, one planning and one reviewing, you're doing the job of a network router, passing messages between two stateful agents by hand. Loop engineering hands that job to a script. Neither one lets agents actually work together. Vlad Luzin (Co-founder & CTO, Band) explains why connecting agents is a distributed systems problem. Bigger context windows don't fix the single-agent bottleneck. Messaging platforms like Slack, Discord and WhatsApp are built for humans and block bot-to-bot traffic. Protocols like MCP and A2A are too low-level: stateless calls, one-wa

Developers using autonomous agents wrote 741% more code but shipped only 30% more software. The bottleneck is human review, and "just review harder" doesn't scale. Reviewer effectiveness collapses past about 400 lines, and agents now open 10,000-line PRs. Laurie Voss (Head of Developer Relations at Arize AI, co-founder of npm) reviews what the industry is actually doing about it. The evidence covers OpenAI's zero-human-code product, METR's finding that about half of SWE-bench-passing PRs wouldn't be merged, and Cognition's FrontierCode (88% on SWE-bench Pro vs. 29% on real mergeability). She

The Chief AI Officer is one of the fastest-growing roles in tech, and one of the least defined. AI is in everything, so the job can easily turn into everything. Rania Khalaf (Chief AI Officer, WSO2; previously 20 years at IBM Research, where she ran about a third of the global AI research organization) shares a framework from two years in the role. It splits the job into three focus areas, Scientist, Architect and Coach, each a slider whose setting depends on your company type, its AI maturity and your own skills. She draws on her time at an ag-biotech unicorn, where simple blob detection bea

Justin Reock (Deputy CTO, DX) shares five trends from DX's data on about 200,000 engineers. Deployment frequency is rising, but change failure rate has become far more volatile. Maintainability is up while change confidence is down. PRs have grown from about 44 to 72 lines. Juniors use AI the most, but staff+ engineers save as much time while using fewer tokens. Median gains in PR throughput are around 7.7%, and even top performers didn't reach 2x, because code generation was never the bottleneck. Justin walks through the DX AI Measurement Framework (utilization, impact, cost), how to assess y

Full episode: https://www.youtube.com/watch?v=f6cjAEr08qk Me on twitter: https://x.com/dwarkesh_sp

Full episode: https://www.youtube.com/watch?v=f6cjAEr08qk Me on twitter: https://x.com/dwarkesh_sp

From giving AI agents their own cloud computers to making Computer Use faster than humans at real-world tasks, OpenAI is pushing agents much closer to actually operating software end-to-end. In this episode, recorded immediately after OpenAI DevDay, Ari Weinstein, who leads product and engineering for Computer Use, joins swyx and Vibhu to unpack Dots, GPT-6.1, the Agents API, App Shots, and the rapid improvements that have made Computer Use dramatically faster, cheaper, and more reliable. Ari explains why Computer Use is now “180 degrees different” from where it was months ago, how agents are

The defense industrial base of the future will be built by small, fast-moving teams working on some of the hardest problems in national security. Live from Washington, D.C., Y Combinator brings together founders, senior government officials, and military leaders for a day of conversations about rebuilding America’s defense industrial base, working with the government, and closing critical capability gaps across autonomy, munitions, space, secure communications, manufacturing, and more. Full event agenda here: https://docs.google.com/document/d/1RktUN1vOL4uzPYhwJcdsUvc1molLQhFaQApB_FIMzOk/edi

Halluminate (YC S25) is building reinforcement learning environments and benchmarks to help AI models do knowledge work beyond coding, starting with finance. With a team of fewer than 10 people, the company works with four of the top five closed-source US AI labs and recently raised a $30 million Series A led by Oak HC/FT. In this Fireside, co-founders Jerry Wu and Wyatt Marshall sit down with YC Partner Jon Xu to share how testing browser agents led them to building simulations where models practice tasks like completing financial spreadsheets. They explain why accurate scoring and expert re

本篇为神奇组织之1972苏联粮食大劫案。