Daily Brief

Claude is capable of long horizon tasks. In this talk, we'll share lessons learned about building agent harnesses for reliable and secure long-horizon work. This include decoupling the brain and hands, self-verification, self-learning, and design for evolving agent harnesses. ### Lance Martin Member of Technical Staff · Anthropic [X/Twitter](https://x.com/RLanceMartin) · [LinkedIn](https://www.linkedin.com/in/lance-martin-64a33b5) · [Website](https://rlancemartin.github.io) Member of technical staff at Anthropic. Working on the Claude Platform, including Claude Managed Agents and the claude

To automate opening a bank account, your agent needs to verify identity, so a team wires it to the DMV and a passport service and ships it. Then the next team builds the next agent and rediscovers, from scratch, where its data lives, across a hundred databases plus Snowflake, Databricks, and S3, whether it can trust the version, and whether it is even allowed to touch it. Every agent repeats that wiring, nothing updates when a source moves without a manual rewire, and no agent is smarter tomorrow than today. Emil Eifrem's fix is to make the agents thin and put the intelligence in a shared subs

Stephen Chin gave two copies of the same agent the same facts about his home network, one storing them as a vector database, the other as a graph. He asked both what was running end of life software exposed to the internet. The vector agent said it could not find specific details. The graph agent traced the connections and flagged his daughter's Minecraft server running an out of date OS, then found real open management ports he quietly patched after the demo. Same data, and only one of them could actually answer. That gap is the whole talk. Most assistants, OpenClaw included, keep their memo

Not everyone needs the same tokens. Matan Grinberg of Factory on routing work across frontier models, Gemini Flash, and fine-tuned COBOL models — with policies written in natural language.

Yohei Nakajima was running a 500 question eval when his API key died at question 350. Normally that means restarting the whole long agent from scratch. Instead it rolled back one step and resumed at 353, because in ActiveGraph the log is the agent. Most people build agents around the LLM and bolt on memory and logging; Nakajima, the creator of BabyAGI, flips it and builds around an immutable event log. Every action and every change to the agent flattens into one typed log, which projects a graph that is the agent's state, so you get replays, rollbacks, and forks for free. On top of the log si

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

I had Kareem Amin all wrong. When Clay first crossed my radar, I thought he was too laid-back: no chip on his shoulder, not the hungry founder I expected to win. Clay has since exploded. Kareem builds from what he calls "a place of wholeness" rather than lack — nothing to prove, no enemy to beat. It makes him sound like a philosopher, but the operating instincts underneath are sharp. He spent roughly five years wandering before Clay clicked, and his takeaway is that the hard part of building isn't working hard — it's the courage to commit to one idea and stop listening to everyone else, cust

Human-agent collaboration is changing, becoming more visual. The agents most teams ship today still wait for us to type a paragraph to explain what we're looking at. They cannot see a screen, navigate a UI that changes, or recover when an application throws an unexpected modal. That is the architectural gap between agents that demo well and agents that work alongside real teams in real software. Perception agents close it: they see and use computers the way people do, reason about what they see, and act with clicks and keystrokes. Speaker: Antje Barth — Member of Technical Staff, Amazon AGI

In July 2025 Dex Horthy turned the lights off: an agent software factory where nobody read the code. It fell apart. An issue appeared that no amount of prompting could fix, the site was down, users were furious, and he was digging through a codebase he had stopped reading three months earlier. His claim is that this is not a skill or scale issue, and no harness or extra tokens fixes it, because it is a model training problem. Coding models are reinforced on one thing, did the test pass without breaking another, and nothing in that reward penalizes bad architecture, whose cost shows up months l

Every month it is the same trap. A reasoning model gets upgraded at the same price per token, then quietly burns three times the output tokens. Or the new version costs 40% more and deprecates its predecessor in four months. Are you growing 40%? Making three times the revenue? Sarah Sachs, who leads AI engineering at Notion and negotiates its model contracts, says no, and that is why she treats every vendor as a competitor. Buy tokens from a lab that also sells the first party product and you are paying a markup on a markup for something you cannot defend, with no exit if you lock yourself to

By declaring a task's inputs and outputs without initially considering model capability, you create the space needed to determine execution later. DSPy's promise is that AI engineering should happen above a particular prompt template or provider API shape: the Signature. That remains useful in a world of tools, RLMs, and Skills. Define a task strictly through its inputs and outputs, and the underlying implementation becomes flexible: experiment with models, settings, weights, templates, and output formats without touching the workflow. The talk covers DSPy 3.5 and previews DSPy 4.0, where mod

From spending $12 million building language models for code before the world cared to creating a Model Factory that can take a model from pre-training to release in eight weeks, Eiso Kant has spent more than a decade betting that code is the path to AGI. In this episode, the Poolside co-founder joins swyx and Vibhu to explain why ChatGPT felt like vindication, why Poolside embraced open weights and open research, and why he would rather live in a world with 100 foundation model companies than five even if Poolside were one of the five. We go deep on Poolside’s Model Factory: the engineering s