Comment by Ancalagon

Comment by Ancalagon 3 days ago

5 replies

I still feel like with all of these tools I as a senior engineer have to keep a close eye on what they're doing. Like an exuberant junior (myself 10 years ago), inevitably they still go off the rails and I need to reign them in. They still make the occasional security or performance flaw - often which can be resolved by pointing it out.

chasd00 3 days ago

I was experimenting this morning with claudecode standing up a basic web application (python backend, react+tailwindcss front end, auth0 integration, basic navigation, pages and user profile).

At one point it output "Excellent! the backend is working and the database is created." heh i remember being all wide eyed and bushy tailed about things like that. It definitely has the feel of a new hire ready to show their stuff.

btw, i was very impressed with the end result after a couple hours of basically just allowing claudecode to do what it wanted to do. Especially with front-end look/feel, something i always spend way too much time on.

englishspot 3 days ago

I keep hearing about how they're "really good" now, but my personal experience has been that I've always had to clear sessions and give them small "steps" to execute for them to work effectively. thankfully claude seems really good at creating "plans", though. so I just need claude code to walk through that plan in small chunks.

  • chrysoprace 3 days ago

    Setting small goals with quality gates is good. I'll usually write something like "once you've implemented this, I'll review before we continue".

    • englishspot 2 days ago

      I review even before they implement. My typical workflow for anything major is to ask for a plan and an overview of the steps to execute. This way I can read the plan, mull over it, make a few changes myself. And then when I'm ready I go through the steps with claude code, usually in fresh sessions.

kevin_thibedeau 3 days ago

I asked a niche technical question the other day and ChatGPT found fora posts that Google would never surface in a million years. It also 100% lied to me about another niche technical question by literally contradicting a factual assertion I made in my question to prime it with context. It suffers from lack of corpus material when probing poorly documented realms of human experience. The value for the human in the chain is knowing when to doubt the machine.