Tweet by zeeg

July 27, 2026

one of the biggest gripes i have about tech (and startups) right now is... everyones trying to build the-everything-solution what i mean is: many startups should be building "features" that can slot in, maybe break out one day. say take all the components you need to make a coding agent work: the IDE, verification layers, sandboxes/vms, security, etc. but instead they're not doing that, and thus no one is achieving any breakout success. they're all just me-toos that dont work that well, and certainly not better than the other persons idea i dont really blame them as you are taught to swing big, and you're somewhat naively convinced you can right now. except whats happening is everyones blindly confident that they've found _the secret sauce_ (100% placebo btw) and they barely ever scratch the surface on average I'd agree with the wisdom that you should aim big, avoid being the feature, but this is an odd domain where the problems - each component - is so uncommonly difficult and valuable that we SHOULD be building those. and thats just in software dev! you see the same thing in how indie developers operate, and in open source, where instead of trying to lego together things to (truly) solve a really big problem, everyones pretending they can vibe code some slop with the way its going we're like six months away from people thinking SQL is too complicated and having their agents invent new graph stores...

Author
zeeg
Date
July 27, 2026
Open on X