Hacker News Reader

wow really? Nov 2025 I was still fighting blatant hallucinations in helm charts. Now I'm two prompts away from a new build published on the app store and 5 video reels in the queue complete with captions and everything

View on HN

The thing is I'm not spinning up a new app every day, I'm just working on a large existing system with enough constraints. From that perspective tools + the general rigor I already try to apply means that I can get some value out of LLMs without waiting for it to do something "good".

"I can one-shot a new thing" is interesting but definitionally a thing is new only once. Maybe if I worked in a space where we had way more throwaway things to work on.

At the end of the day I want to understand the output anyways, so it one shotting a thing vs me going back and forth with it in smaller increments is a mixed bag anyways.

Very true, creating a lot of content and using the algorithms as a testbed is much different from curating a performant codebase. One example of how the models have improved in the latter regard is how Opus really seems to go the extra mile, on things I might skip purely from a time/effort savings perspective. There are a lot of optional fields that have different defaults in different environments, and I wouldn't typically explicitly define them myself until there was a problem, but Opus will now go through and set those values without me asking. I still make some modifications (as it's easier to spot an error rather than set everything from scratch), but it's very helpful.

Reply on news.ycombinator.com