Software creation is exploding. But who tests it?
Tools like Cursor and Devin are 100x-ing developer speed. This creates a massive new QA bottleneck. If code is written 100x faster, who tests it?
If you work in tech, you've felt the ground shifting.
Tools like Cursor or Devin are changing the speed in which software is created. But this software creation acceleration results in an immediate, massive, and largely un-discussed bottleneck: Quality Assurance.
SaaS companies are focusing on shipping faster, but we need to ask a critical question: If software is shipped 10-100x faster, who tests it?
Massive - and growing - speed mismatch
For the last decade, QA has struggled to keep up with agile development and CI/CD, often becoming the bottleneck before shipping. But what we are facing now is something else entirely.
- Old mismatch: Human developers (working in 2-week sprints) vs. Human testers (writing scripts that take days).
- New mismatch: 10-100x AI-assisted developers (working in few-hour cycles) vs. a Human tester (still writing scripts that take days).
In order to stay relevant, companies need to be shipping software faster than ever. And yet, the "QA bottleneck" is growing rapidly. This is forcing companies to take a high-risk gamble, shipping products with a much greater chance of critical bugs.
The question for high-shipping-velocity teams is no longer "How can we speed up our testing?" but "How does our quality strategy scale when our development speed just 10x-ed?"
Questions on the new QA bottleneck
How do we test the quality of our product if we ship 10x faster?
This is the critical question. Shipping 10x faster means the volume of change is 10x higher. Old methods, like manual testing or script-based automation, don't scale with this speed. They break down, forcing teams to choose between slowing down or shipping untested code.
Why can't traditional test automation keep up with AI developers?
Traditional automation relies on brittle, hand-coded scripts. An AI developer will refactor and change the UI and code logic far too rapidly for a human team to maintain these scripts. The maintenance that already plagues QA teams will become impossible to manage.
What's the real business risk of this new QA bottleneck?
The risk is twofold. First, an increase in critical, customer-facing bugs that damage reputation and revenue. Second, teams will be forced to slow down their AI-powered devs to let human QA catch up, completely wasting their investment in development speed and losing their competitive edge.
What happens to the human QA team when AI writes the code?
The role of human QA must evolve. Instead of spending 90% of their time writing or fixing brittle scripts, their focus must shift. The new role will likely be more strategic, focusing on defining test goals and validating the user experience, rather than manually testing every button.