Rethinking Git for the Age of Coding Agents with GitHub Cofounder Scott Chacon
Episode
47 min
Read time
2 min
Topics
Productivity, Startups, Fundraising & VC
AI-Generated Summary
Key Takeaways
- ✓Agent-optimized CLI output: When GitButler added a `--json` flag expecting agents to prefer structured data, agents instead performed better with human-readable output, then piped it through JQ themselves. The practical fix: append status output automatically after every mutable command, since agents run `git status` after nearly every operation regardless.
- ✓Parallel branch architecture for multi-agent teams: Rather than using Git worktrees — which create isolated copies of the working directory, preventing agents from seeing each other's changes — GitButler's parallel branch model lets multiple agents operate on one shared working directory. Agents detect file conflicts and automatically stack branches, eliminating merge conflicts before they reach GitHub.
- ✓Communication overhead is the real multi-agent bottleneck: Running five or more agents simultaneously creates a management problem, not a compute problem. A more productive model has agents use idle cycles to communicate with teammates' agents — surfacing file conflicts, flagging relevant changes, and briefing the human developer — rather than maximizing parallel agent count.
- ✓Writing ability becomes the primary engineering skill: As implementation cost approaches zero, the constraint shifts entirely to specification quality. Developers who can write precise, detailed specs — describing desired behavior rather than implementation details — will outperform those who rely on technical execution alone. The "why" compounds in value as the "how" becomes automated.
- ✓PR-based code review is structurally broken: Surveys show roughly 80% of developers still use Git CLI over GUIs. Meanwhile, PR review has degraded to cursory checks because branch-based review divorces commit messages from merge context. Patch-based, local review — where agents can pull, compile, run tests, and return a prioritized findings list — addresses what PR review never reliably delivered.
What It Covers
GitHub cofounder Scott Chacon and a16z's Matt Bornstein examine why Git's 20-year-old interface — built by Linus Torvalds' Linux kernel team with no intentional UX design — fails both human developers and coding agents, and how GitButler's CLI rethinks version control primitives for agentic workflows.
Key Questions Answered
- •Agent-optimized CLI output: When GitButler added a `--json` flag expecting agents to prefer structured data, agents instead performed better with human-readable output, then piped it through JQ themselves. The practical fix: append status output automatically after every mutable command, since agents run `git status` after nearly every operation regardless.
- •Parallel branch architecture for multi-agent teams: Rather than using Git worktrees — which create isolated copies of the working directory, preventing agents from seeing each other's changes — GitButler's parallel branch model lets multiple agents operate on one shared working directory. Agents detect file conflicts and automatically stack branches, eliminating merge conflicts before they reach GitHub.
- •Communication overhead is the real multi-agent bottleneck: Running five or more agents simultaneously creates a management problem, not a compute problem. A more productive model has agents use idle cycles to communicate with teammates' agents — surfacing file conflicts, flagging relevant changes, and briefing the human developer — rather than maximizing parallel agent count.
- •Writing ability becomes the primary engineering skill: As implementation cost approaches zero, the constraint shifts entirely to specification quality. Developers who can write precise, detailed specs — describing desired behavior rather than implementation details — will outperform those who rely on technical execution alone. The "why" compounds in value as the "how" becomes automated.
- •PR-based code review is structurally broken: Surveys show roughly 80% of developers still use Git CLI over GUIs. Meanwhile, PR review has degraded to cursory checks because branch-based review divorces commit messages from merge context. Patch-based, local review — where agents can pull, compile, run tests, and return a prioritized findings list — addresses what PR review never reliably delivered.
Notable Moment
Chacon's team built a live chat channel allowing three simultaneous agents to coordinate in real time, with a visible TUI showing their conversations. After testing, they discovered agents without the chat channel outperformed those with it — the agents inferred context directly from file changes faster than explicit communication allowed.
Episode Transcript
If you ask almost any software developer, when you do code review, do you really read the whole PR? Like, do you go through every line and think it through? Do you pull it down and test it out and then leave the good feedback on each line? Agents are very good at that. That. Right? If something goes wrong, that's very human. It's actually a thing that's never been good in software development is inter team communication. And so it's a very interesting UX problem set that I think nobody's really thought through really even now. What does that tool look like in a way that is easy to use and easy to learn? Software developers that'll be the best producers of product in the near future are the ones who can communicate, the ones who can write, the ones who can describe. That is, I think, the next superpower. And I think if we could talk to each other in more real time about what we're doing, that's a lot of overhead, but that is not a problem that agents have. The most widely used developer tool in the world was never designed. Git started as plumbing commands for the Linux kernel team. Unix primitives meant to be wrapped in whatever scripts each developer preferred. A volunteer wrote a unified interface. It got pulled into core. And for twenty years, almost nothing has changed. Now coding agents are the fastest growing users of command line tools, an entirely new persona. They struggle with interactive rebasing. They run status after every command. The assumptions baked into Git's interface no longer hold for humans or machines. The question is whether the tool underpinning nearly all modern software can adapt or whether something new has to replace it. Matt Bornstein, general partner at a sixteen z, speaks with Scott Chacon, cofounder of GitHub and CEO of GitButler. We are here today with Scott Chacon, CEO of Git Butler, former cofounder of GitHub. Thank you very much for being here. Of today. Thanks for having me on. You are a major driving force behind GitHub. You've literally written a book on Git. You could be doing anything in the world with your life right now. What's brought you back to startup land? It's interesting. I feel like if you ask any sort of repeat founders, they probably have similar answers. Right? This is the most fun thing to do. So when I started at GitHub, it was a real sort of slog to learn. Okay. Like, it's stressful and it's difficult and stuff. But when you get something working, it's so satisfying and it's so much fun to build and grow and create something that you wanna see exist in the world. So I'm I'm sure I'll be doing this when I'm 90. Do you think there's kind of unfinished business for you in version control, or what kind of attracts you back to the same space that you know so well? …
Get the full transcript (10,582 words) + summary by email — free
One-time email with the complete transcript and AI summary of this episode. No account needed.
One email, no spam. We’ll also show you what SignalCast does.
You just read a 3-minute summary of a 44-minute episode.
Get a16z Podcast summarized like this every Monday — plus up to 2 more podcasts, free.
Pick Your Podcasts — FreeKeep Reading
Books, tools, and gear mentioned in this episode
SignalCast may earn commission on purchases via these links.
More from a16z Podcast
We summarize every new episode. Want them in your inbox?
Inside Moderna’s Personalized Cancer Vaccine
Daniel Litt: The Mathematician's Guide to AI
Gavin Baker: Why AI Demand Is Outrunning Compute Supply
Why a16z Launched the Machine Age Fund | Jen Kha
Why 1,200 AI Agents Started Working Together | Ryan Greenblatt
Similar Episodes
Related episodes from other podcasts
The Prof G Pod
Aug 7
The Week: How Leverage Broke the AI Trade
The Daily (NYT)
Jun 17
The Battle Over A.I. in the Classroom
Invest Like the Best with Patrick O'Shaughnessy
Apr 14
Scott Nolan - SpaceX, Founders Fund, and Rebuilding American Uranium Enrichment - [Invest Like the Best, EP.467]
Masters of Scale
Mar 31
The internet is breaking. So what’s next? with Cloudflare’s Matthew Prince
BiggerPockets Money Podcast
Feb 6
Debate: AUM vs Flat Fee Which is Better?
Explore Related Topics
This podcast is featured in Best Business Podcasts (2026) — ranked and reviewed with AI summaries.
Read this week's Startups & Product Podcast Insights — cross-podcast analysis updated weekly.
You're clearly into a16z Podcast.
Every Monday, we deliver AI summaries of the latest episodes from a16z Podcast and 192+ other podcasts. Free for one show.
Start My Monday DigestNo credit card · Unsubscribe anytime