Self-Maintaining APIs
The speaker argues that API communication is fundamentally broken across the industry, with breaking changes and new features going unnoticed. Since agentic coding tools have proven viable and developers accept external tool access to codebases, API providers should automatically apply changes to customer code rather than just announcing them—similar to how Dependabot works for dependencies.
Summary
The speaker presents a critical analysis of API communication failures based on experience working with over 50 API vendors, including early-stage startups and AWS. The core problem is that breaking changes ship with insufficient warning, useful features launch quietly without awareness, and changelogs go unread. This friction results in significant costs—the speaker cites that over 30% of service downtime at AWS was attributed to unnoticed external API and package changes.
The speaker argues this problem persisted because the infrastructure for automated code changes didn't exist. However, agentic coding tools like Claude, Davenin, and Gravile have fundamentally changed this landscape. Two years ago, developers would never grant external tools access to their codebases, but this is now standard practice, proving both developers and enterprises value these tools sufficiently to accept the security model.
The proposed solution is for API providers to move beyond passive communication to active code modification. When Stripe ships a breaking change or new feature, for example, an agent should automatically scan customer codebases, identify affected usages, and open pull requests with fixes. This could be implemented either as per-provider agents (like a Stripe Update Agent) or as a neutral third-party service tracking changes across multiple vendors—essentially Dependabot for APIs rather than just dependencies.
Key Insights
- Over 30% of service downtime at AWS was caused by external APIs and package changes going unnoticed, demonstrating the severity of the API communication problem even at large enterprises.
- Developers and enterprises have proven willing to grant codebase access to external agentic tools, a paradigm shift from two years ago when this was considered unthinkable.
- The infrastructure for automated code changes already exists; what is missing is the application layer connecting API providers to customer codebases.
- API providers should transition from announcing changes to actively applying fixes, with agents automatically scanning codebases and opening pull requests when changes are released.
- This solution could be implemented either as per-provider agents or as a neutral third-party service—essentially creating Dependabot-like functionality for APIs across vendors.
Topics
Transcript
[0:00] Over the past year, I've worked at over 50 API vendors, mostly early stage startups, and just one pattern is consistent. API communication is broken. Breaking changes ship with little warning. Useful features quietly launch and go unnoticed. Change logs don't get read. Heck, this is a problem even when I worked at AWS. Over 30% of a service downtime was due to external APIs and package [music] changes going unnoticed. This friction made sense before agentic coding tools existed. However, now it doesn't. Agentic coding tools like cloud code, davenin, [music] gravile, etc. [0:30] prove that developers and enterprises are willing to give codebase access to external tools provided they're valuable. Two years ago, this was unthinkable, but…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Y Combinator
What Actually Makes A Startup Durable
YC partners discuss what makes startups durable in the AI era, emphasizing that while AI reduces software development costs, founders must focus on harder problems with defensible moats, maintain direct customer contact, and prioritize co-founder relationships for emotional and strategic support. They argue that pure software businesses are no longer durable and that the pace of AI improvements means cost concerns will self-resolve within 6-12 months.
Proving You’re Human
Deepfakes and AI-generated voice clones have made digital fraud increasingly sophisticated and harder to detect, with a recent $25 million wire fraud involving deepfake video call participants exemplifying the problem. The traditional trust signals of seeing faces and hearing voices are no longer reliable, necessitating a fundamental rebuild of the internet's trust layer to verify that humans—not AI—are on the other end of communications.
What Big Tech Missed And How Startups Can Still Win
Alex Lebrun, CEO of Amilabs, discusses his journey building multiple AI startups and explains why world models—AI trained directly on sensory data rather than text—represent the next frontier beyond large language models. He argues that being early in technology requires specific founder strengths, and shares lessons on ambition, fundraising, and the challenge of managing a $1.2 billion seed round.
Why Physical AI Is the Next Platform Shift
Eric Landau, co-founder of Encord, discusses his career trajectory from particle physics to quantitative trading to founding a data infrastructure company for AI. He explains how Encord pivoted from general vision AI to focus on Physical AI, which he believes will be a massive market as robotics proliferates across manufacturing, logistics, and autonomous vehicles.
New Operating Systems for the Physical World
Traditional software for physical-world industries like construction and fleet operations hasn't evolved in 20 years, but AI agents, robots, and wearable-equipped humans now create a new operating system opportunity. These new systems will manage human and robot labor together, capturing end-to-end work data that incumbents lack, while operating in industries that spend 10-100x more on labor than software.