throwaway12345t

He is very talented but is maybe the least interested person in AI generating/reviewing code and borderline dismissive. I agreed a year ago but things have changed.

He is not championing the effort on the team and as a result the rest of the team is either mostly not using it or just not disclosing it.

He is more senior and in general the senior engineers seem the most existentially stressed or indifferent.

It feels like we’re moving very slowly compared to the competition now and I’m unsure how to convince him to reevaluate LLM’s use in our development work.

It’s not a post to deny that the field is moving and is very interesting to follow but it’s just become too much.

Every week there’s several new things, most examples are cherry picked, benchmarks are blown out etc.

I run a software company that has not added any AI features, mostly because the usefulness seems to fall below the bar for value that most of our features try to maintain.

I use ChatGPT personally for random searches and Cursor + whatever model is deemed best at the time but even with that it takes a lot of work to get something valuable out of it in day to day.

I feel like I’m losing my mind when I see startups posting 10m ARR numbers in just 6 months or whatever.

I’m hearing from VC’s churn is 15-30% in many of these companies and they’re far from profitable but the growth is just wild.

Do I succumb and just add yet another text generation that maps to an object like fill-in-the-blank app of the week?

It feels disingenuous but even companies I know and respect are adding questionable “agent” features that rarely work.

Anyway, how are you feeling?