eugmill
Joined 13 years ago
Joined 13 years ago
I have been messing with vibe engineering on a solo project and I have such a hard time telling if there's an improvement. It's this feeling of "what's faster, one lead engineer coding or one lead engineer guiding 3 energetic but naive interns"?
I think the "GPT is amazing" vs "GPT is useless" debate is just going to get more confusing as more versions are released.
I still can't stand the bottom popping up and down and not being able to tell if I'm muted.
At the time I had facebook notifications send me sms messages, and if you reply, it would post the reply to facebook.
I had left a bunch of "I'm driving and can't talk right now" comments on random posts without realizing it until days later.
Also the offline support isn't great which is a dealbreaker in the subway where internet goes in and out.
Deadmau5 uses polymeters pretty extensively and rap/trap uses polyrhythms all the time.
Yes there is a ton of electronic music and yes a lot of it sounds similar but there are new things coming out if you can dig through all the volume.
A lot of the innovation is also in sound design and not necessarily composition.
I think if not for tesla, other car companies would not be scrambling to make better Evs.
In my limited experience, the ones that struggled most were the ones who simply didn't do the prework that was asked of them. Some people think education is something that will happen to them if they pay somebody enough, compared to those that went to a boot camp to accelerate the self-education they had already started.
Having said that, a PleaseDontDoAnythingStupidEval would probably slow down agentic coding quite a bit and make it less effective given how reliant agents are on making and recovering from mistakes. The solution is probably sandboxes and permission controls to not let them do something overly stupid, no different from an intern.