







I think a lot of confusion about AI comes from relatively few people having worked in a founding context.
The difference between the 80% vibe-coded solution and the 99% real product is everything. In that 20% is your ability to actually charge customers and to retain them, the ability to compete, and all areas of differentiation.
That 20% represents a million tough decisions.
These decisions are hard, because there's often no way to know the objectively correct choice. You make decisions with a mix of noisy data, intuition, forecasts on where a market or competitor may be going, and tons of annoying constraints.
This hard work hasn't disappeared.
LLMs can assume some decisions (though even these smaller "automated choices" are not without risk). But I don't think anyone can credibly claim that an executive should or can offload the really important, fuzzy decisions that make a product real and a company successful to an LLM.
Likely, we'll all increasingly feel the burden of decisions as AI implicitly promotes people into quasi-executive roles that command agents.


















