Post

My AI enthusiasm ebbs and flows. Today I’m realizing that some of my perceived productivity increase is an illusion. I merely moved some of the effort from writing code to having to micromanage a coder.

Curious: How do you start a feature? Do you just describe the feature and tell it to go? Do you have it interview you? Do you write down a spec or plan? Something else?

We usually have a long discussion, plan mode, then execute. But the discussion can feel like micromanagement too because Infind I have to spell a lot more out to it to get a good plan than I would get from an experienced colleague.

How do the topics in discussion get raised? Do you decide or does the agent decide? If the agent decides, how did you prompt it to choose?

I’ve had success asking the agent to raise any open questions it has, points of ambiguity, or gaps where we might be misaligned.

asking for open questions and ambiguities up front is the best cheap trick i've found too. it surfaces the spots where the agent quietly filled a gap with a guess. i'd rather see the guess flagged than buried in the diff later.

I know there’s a lot of content out there about orchestrating many parallel autonomous agents. I have not achieved that nor have I seen nearly enough smooth successes to be in a hurry to get there. I need to have my attention on one thing at a time still. Skill issue? We’ll see.

And also even if I could drive that well, I can’t afford it while I’m self-funded.

And also, I’ve been getting this sort of content for months and months. If everyone is so good at this for real, where is the orders of magnitude more and better software?