hollerboy

if anything i feel like using agent to code is a great opportunity to sharpen your critical thinking if you take the time tear apart the agent's assumptions and actively find flaws, areas of improvements etc. i don't think llms have removed any opportunity to practice good thinking (yet), it's just that you have more room to reason in design space rather than implementation space. TL;DR just keep poking at whatever your agents claim to be correct

speaking from personal experience only, i think it pretty much boils down to owning the code, i.e always having a deep and up to date understanding of the abstractions and the underneath assumptions. ironically this isn't new, it's tempting to just hand-over ownership to the AI so llms kind of just made the trap easier to fall into

0. Developer 1. Lenovo L340 2. I often find that ~5 years is about the right time to change so i can stay somewhat in touch with the current tech. I often get lower-spec devices as i mostly work via remote machines though