On building product: Startups get more from AI than large companies

I've noticed indie hackers and startups are getting far more from AI than larger companies.

Developers often point to codebase size - it's easier for an LLM to grasp 10K lines of code (LOCs) than millions. That's true to an extent, but it misses the point entirely.

In my experience, the real difference is how smaller companies organize their work. AI gives huge autonomy to individuals. Someone with no art skills can participate meaningfully in design reviews. Someone who couldn't write a decent Slack response can now communicate more clearly with stakeholders. A developer who has no idea of writing a good copy can now write one for the initial iterations. 

To benefit from this, larger companies need to rethink their processes, and it is a mindset shift rather than a productivity problem. If you needed five specialists to ship something, maybe now it's just three or two people who can complete the work. This also means that experts can be involved in foundational work and monitor the rest, rather than being involved in every detail.

This brings the focus from making each specialist more efficient to restructuring the work around fewer handoffs, more individual ownership, and faster decisions to get things done.

One potential problem I'm seeing in this shift is that many of us like credentialism - someone doing our work without having that particular credential (e.g., an engineer does part of the design).

The companies that figure this out won't just be more productive - they'll operate in a completely different way.