<TechBlogs/>
GitHubAI-Native DevelopmentPierre ComputerReliability Issues

Is GitHub Still the Best Option for AI-Native Development?

Is GitHub Still the Best Option for AI-Native Development?

Introduction

The reliability of GitHub, the largest Git host in the world, has been a concern lately. With an increase in load from AI agents generating more code and pull requests, GitHub's infrastructure has struggled to keep up. In this article, we will explore the current state of GitHub, its reliability issues, and whether it is still the best option for AI-native development.

Reliability Issues

GitHub's reliability has been down to one nine in the past month, with issues on 3 days out of every 30 days, or issues/degradations for 2.5 hours daily. This is unacceptable for a platform that is expected to be highly reliable, with four-nines of availability (99.99%, meaning about 52 minutes of downtime per year). The main cause of these issues seems to be the massive increase in infra load from agents, which GitHub's infrastructure is unable to handle.

CTO's Response

GitHub's CTO, Vladimir Fedorov, addressed the availability issues in a blog post, covering three major incidents:

  • 2 February: security policies unintentionally blocked access to virtual machine metadata
  • 9 February: a database cluster got overloaded
  • 5 March: writes failed on a Redis cluster

A helpful analysis of these outages and the CTO's response was done by software engineer Lori Hochstein, who noted that:

  • Saturation: the database cluster incident was a case of the database getting saturated due to higher-than-expected usage.
  • Failover + telemetry gap: the 2 February incident was a combination of an infra issue in one region failing over to a healthy region, and making things worse with a telemetry gap.
  • Failover + configuration issue: the 5 March incident was similar, with a configuration issue blocking writes on a Redis cluster after a failover.

Pierre Computer: A New Player in the Field

While GitHub struggles to keep up with the increase in load from AI agents, a new startup called Pierre Computer claims to have built an 'AI-native' solution for AI agents pushing code, which scales far beyond what GitHub can do. Pierre was founded by Jacob Thornton, formerly an engineer at Coinbase, Medium, and Twitter, and also the creator of the once-very popular Bootstrap CSS library.

Pierre supports features that GitHub does not, such as handling a sustained peak of > 15,000 repos per minute for 3 hours, and creating > 9M repos in the last 30 days. These numbers are incredible, and something that GitHub clearly cannot get close to, at least not today.

Has GitHub Lost Focus and Purpose?

GitHub's reliability issues are acute enough that, if it keeps up, teams will start giving alternatives like small startups such as Pierre a try, or perhaps even consider self-hosting Git. But how did the largest Git host in the world neglect its customers, and fail to prepare its infra for an increase in code commits and pull requests?

What's Next for GitHub?

Mitchell Hashimoto, founder of Ghostty, and a heavy user of GitHub himself, had advice on what he would do if he was in charge of GitHub. He suggested establishing a North Star plan around being critical infrastructure for agentic code lifecycles, firing everyone who works on or advocates for Copilot, buying Pierre, and re-evaluating all product lines and initiatives against the new North Star.

My sense is that GitHub has three concurrent problems:

1. GitHub and Copilot are entangled with Microsoft's internal politics.

2. GitHub has no leader, seemingly by design.

3. GitHub has no focus, and is stuck chasing Copilot as a revenue source.

Conclusion

GitHub's reliability issues and lack of focus on AI-native development have raised concerns about its ability to support the increasing demands of AI agents. While GitHub is still the largest Git host in the world, its current state and lack of vision for the future may lead to a decline in its popularity. Pierre Computer, on the other hand, seems to be a promising alternative, with its 'AI-native' solution for AI agents pushing code. As the demand for AI-native development continues to grow, it will be interesting to see how GitHub responds to these challenges and whether it can regain its position as the best option for AI-native development.