on product mindset.
it's finally cheap enough to do all the crazy stuff that was too boring or tedious to do before.
←
back to thoughts
may 12
the cycle time to iterate on ideas has collapsed. what used to only be possible in quarters now can be done in hours.
may 12
code has way less value on its own now, specially in startups. the value is in the speed at which you can learn from the users using it. anyone can just copy the features, but they won't be able to copy the system behind improving it. that's not visible or obvious. and that usually affects the smaller things.
may 12
users don't know what they need. they know they have pain, and they communicate it however they can. your job is to take that and turn it into the actual solution. to figure out what they mean when they say they need X, because most of the time there is much more to what they say.
may 12
feature work is actually bug-fixing in disguise. you have your hypotheses, you test them, you see which one fixes the issue. just more abstract and with a less clear definition of fixed. but then you just build systems that make those definitions possible and easy to work with.
may 12
for an mvp, lower-quality code is fine but only if you're going to scrap it, or you have a plan to clean it up before it turns into a huge fucking mess. if you wanna take your mvp and scale it to unicorn you'll make your future engineers wanna kill themselves. i've lived through it. trust me. don't.
may 12
taste and judgment are the real moat now. everyone can create code. but tase and judgment they come from actually having built things. llms are trash at that. for now. who knows in the future.
may 12
the goal isn't to kill fast. it's to learn fast. the killing will come naturally once you figure out what the users actually need, and you can only do that through learning.
−
100%
+
⊙