(Why) do big-company engineers like AI less?
Near the end of a piece that I recently blogged about, John Wang makes a brief but remarkable observation:
Organizations that bias towards speed over quality tend to see more IC adoption of AI (e.g. my network of engineers at startups are on the whole adopting AI and using it to speed quite a few things up, though not necessarily making things higher quality).
Interesting! I suspect it's true that the average engineer at a big company likes AI less than the average engineer at a startup. Why might that be?
- Big companies are more quality-focused, which means both that AI-assisted outputs are a worse substitute for non-AI-assisted ones (because the engineers are better)1 and that there's less organizational incentive to ship lower-quality outputs. (This is Wang's thesis.)
- Startups often recognize and reward quantity of output; big companies have a remarkably hard time doing this. This sounds like the same thesis repeated, but I'd emphasize that big companies are often "choosing quality over speed" only in the sense that they can't do otherwise. Quantity of output seems easy to recognize and reward, but at scale, it is not. (I'd like to say more in a future post about why this is true.)
- In immature projects, speed is a form of quality; Wang is presenting a false dichotomy. A buggy prototype is better than no prototype. Some observability, thoughtfully implemented, is better than no observability. And software development is irreducibly empirical: even if you do accept lower quality for a while, that brings quality benefits in the longer term, from the extra evidence and feedback you get from the thing's existing.
- Engineers at bigger companies feel more threatened by AI: they think, probably correctly, that their employers will react to huge efficiency gains by cutting at least a few engineering jobs. Many "speed-oriented" companies are less likely to do this: first, because firing even one engineer would be too much of their engineering staff, and second, because they are more likely to be limited by development speed in the first place. They can react to efficiency gains by simply having their engineers do more, which for a lot of reasons is harder to do at a big company.
- Engineers at smaller companies are often more mission-motivated. The more your professional motivation is coming from bringing a specific thing into existence, the better you'll feel about a thing that radically accelerates the process of doing so.
Here as before, it's important to distinguish someone's professional stance from their overall attitude. Many big-company engineers are lukewarm about AI on the job, then go home and joyfully fire up Claude Code in personal projects.
There are, of course, many engineers at smaller companies who can more than compete with big-company engineers. (I certainly hope this is true, because I am an engineer at a small company!) I take Wang's point to be that the best big tech companies tend to hire well-above-average engineers, which is basically true.↩