{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Chain of Thought | AI Agents, Infrastructure & Engineering","title":"For AI Agent Fleets, the Model Is Rarely the Bottleneck | Theory Ventures' Andy Triedman","html":"","width":"100%","height":180,"duration":2957,"description":"Andy Triedman is a partner at Theory Ventures, and he is building his own software factory, while assessing how startups are leveraging theirs. His read after running agent fleets: the models are frankly good enough for most of the work. The vast majority of things you'd ask an agent to do are not intelligence limited. They are context limited. We have infrastructure work to do - but we can be more ambitious!\nSo the hard part moves upstream. Deciding what the agents should build, what they need to know, and when they should stop and ask a person is now most of the job.\nWe cover:\nMoving from solving one problem at a time to building systems that solve whole classes of problems, like generating and testing integrations\nWhich tasks still need human approval, and which can be fully handed off\nWhy a short feature spec leaves too much unsaid, and how better questions up front cut review interruptions\nKeeping shared context and company decisions current, because an agent takes everything it reads as true\nInference ROI, and when well-scoped tasks can go to cheaper models\nSandboxes and simulated customer environments for agents\nFirst steps for teams: shared skills, cloud agents and approval gates\nProtecting customer data, and organizing agent memory from session traces to company priorities\nNew episodes, the ideas behind them and Conor's essays land in the Chain of Thought newsletter first. Subscribe: https://newsletter.chainofthought.show/\nChapters:\n(0:00) Agents are context limited, not intelligence limited\n(0:34) From coding agents to software factories\n(5:53) What still needs human approval\n(10:16) Agent loops that improve the product on their own\n(14:52) Where agent automation breaks down\n(17:59) The bitter lesson and engineering as experimentation\n(20:08) From writing code to directing agent fleets\n(24:18) Shared context and stale decisions\n(27:25) Inference ROI and cheaper models\n(30:23) Moving human time from code to customer intent\n(34:42) Sandboxes and simulated...","thumbnail_url":"https://img.transistorcdn.com/KnEL-9mfZGF096BNQ33xaE_zVyhLvBNV8rHifHuvYgc/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9hYWM2/NmYwMjc4MmJiYTYw/ZTI5NGZhNzRhNmZh/NTU0OC5qcGc.webp","thumbnail_width":300,"thumbnail_height":300}