July 29, 2026
Pull request or pull the plug?
GitHub is the wrong shape for this new world
Coders are asking if GitHub still fits a world where AI writes the software too
TLDR: A new essay says GitHub’s human-paced workflow may not work in a world where AI can write software fast. Commenters were split: some said AI already fits into today’s setup just fine, while others argued the real shake-up is that teams — not GitHub — may be the thing getting replaced.
The big claim in this post is pure workplace drama: the old way of building software may no longer fit the new AI-fueled chaos. The writer argues that GitHub — the site many teams use to store code, review changes, and approve updates — was built for humans moving at human speed. Now that AI tools can spit out features in minutes, the old routine of “make a change, open a review, wait for tests, wait for approvals” is starting to look less like a smooth assembly line and more like a traffic jam.
But the comments? That’s where the real fight breaks out. One camp basically shrugged and said, “What crisis?” User philips stole the show by saying they made a bot account and treated it like a turbocharged teammate — and it was… fine. Not scary, not awkward, just another contributor sending in changes. That’s a spicy plot twist for anyone predicting total AI disaster.
Then the skeptics arrived. cortesoft flat-out questioned the whole premise, asking what exactly is wrong besides slow testing. In other words: is GitHub broken, or are people just impatient? Meanwhile gerdesj went full old-school rebel, arguing GitHub already made coding too centralized and polished, joking that it turned messy, freewheeling software culture into something more like a corporate garden with rules. And ChicagoDave brought the coldest take of all: maybe the real change isn’t GitHub — maybe it’s that big coding teams themselves are becoming optional. So yes, the article asked whether GitHub is the wrong shape. The crowd answered with a louder question: is the tool the problem, or is the whole job changing underneath it?
Key Points
- •The article argues that GitHub-style workflows are mismatched with how software is increasingly built today.
- •It says longstanding bottlenecks such as source control, CI/CD, code review, and deployments now constrain higher development velocity.
- •The article states that code contribution has expanded beyond engineering teams to functions such as sales, support, and marketing.
- •It identifies unresolved questions around code ownership, trust, review, interoperability, and the cost-quality tradeoff of token usage.
- •It describes GitHub, GitLab, and Bitbucket as collaboration tools built around branches and pull requests, and says newer AI agents are pushing beyond the assumptions of that model.