Volume is pretty high when you give agents autonomy to find and fix or improve things. I have now (finally?) set up my VC-era laptop to just run improvement loops for everything. It finds bugs and errors and uses my own analytics and PostHog to find improvements autonomously.
It also does regular browser checks, which seems to be a very optimal way for agents to find bugs and UI/UX improvements (instead of just reading code).
I was kind of hesitant to give it this much autonomy, but with some iteration it actually seems to work very, very well. The future of code development is definitely this, that agents just take care of everything autonomously, and it is most likely the only way to keep up with agentic attacks too.
1
1
247















