🔴 Live show Thursday: AI Code Will Need More Observability 🤔
(remember to click the 🔔 on YouTube to get notified when we go live)
On my livestream this week...If AI causes us to ship more code, we’ll need more testing and QA. It’s unlikely that organizations will want to add staff just so they can use AI, so what’s the solution? My guest, Andrew Tunall, the President and Chief Product Officer at Embrace, joins me to discuss his prediction that we’ll all start shipping non-QA'd code (buggier code in production) and QA will need to be replaced with better observability.
🥳 Waitlist is now open for my next course
We've had a great response to this course announcement, and we're in the middle of recording lectures. People on the waitlist will get a chance to join early and provide course feedback as I build it!
Bret Fisher
🎧 Agentic CI/CD with Solomon Hykes of Dagger
👉 Podcast Audio or YouTube Video
In this latest podcast release, Nirmal and I discuss the intersection of AI agents and DevOps with Solomon Hykes, co-founder of Dagger and Docker. Solomon introduces Container Use, an open-source MCP server that gives coding agents isolated, containerized environments to work in, solving the critical problem of agents interfering with each other when running on the same system.
Also in this episode, we explore how AI agents are forcing a fundamental shift in CI/CD practices, with Solomon making the case that traditional CI/CD as we know it will disappear because agents develop 100x faster than humans and can't tolerate slow, unrepeatable build processes. Container Use enables developers to run multiple coding agents in parallel, each in their own sandboxed environment.
We examine the emerging MCP ecosystem and its potential trajectory, drawing parallels to how containers evolved from developer tools to infrastructure platforms. Solomon warns about the rapid "gatekeeping" phase already happening with MCP, where vendors rush to lock down and monetize what started as an open developer innovation space.
Other topics we touch on are: the four core components of agentic workflows, security challenges with untrusted text injection (the new "SQL injection" problem), token cost management, and the importance of balancing proprietary AI services with open-source infrastructure layers. I hope you enjoy!
👀 In case you missed a newsletter, read them at bret.news