Andrew Zigler’s Post

In this week’s Friday Deploy (yes, Dev Interrupted has a dedicated Friday news segment -- get into ittttt!), Ben Lloyd Pearson and I are joined by Warp founder/CEO Zach Lloyd (a past guest and friend of the pod) to talk about the infrastructure gap blocking most teams from capturing the value from agents -- and why he launched Oz to fix it (btw don't miss the great launch demo from Ben Holmes). But first we cover the news...! And yeah, it's been a week: 🔥🧯 On Monday, GitHub had service issues impacting Actions, PRs, notifications, and Copilot, including notification delays on the order of ~50 minutes before things recovered later that day. (Did anyone else's agents get cranky about this?!) -- its a signal that many are moving from single-agent productivity to multi-agent throughput. 🧛♂️ 💸 We also dig into Steve Yegge “AI Vampire” framing: the productivity boost is real BUT the fatigue curve is real-er. What are we going to do about all this new economic pressure to convert 10x tooling into “10x output”... and who really captures the value of that output? And we shout out Ai2 + the work of Tim Dettmers that explores a methodology for training repo-specialized coding agents on brownfield codebases, using a small number of insights to leverage massive gains with minimal fine-tuning. So tune in now to catch Zach's deep dive into Oz, Warp’s new orchestration platform for running agents in the cloud with sharing, scheduling, artifacts, and auditability. Because folks, we're not in Kansas anymore. 🌪️ 🎧 Breaking GitHub, AI vampires & the great Oz: https://lnkd.in/gdREvbXa