Response to "Startups need to hire more people from the humanities"

In Science and Tech, you generally need to be Reductionist to get something to work at all. By reductionist I mean that one tries to remove as many aspects of a problem as possible, before attempting to solve it. An example from physics might be to ignore friction and wind resistance to make a model of mechanics. This introduces a limit of the applicability, but it helps us learn something important. If you are Newton or Galileo, then this is an approximation worth making. An example from every-day design is to assume that all people are right handed. If you are one of the first to mass produce scissors or golf clubs, then making this approximation is worth it if it helps you get to market first. 

Startups make new things, or bring old things to new markets. Being first to market counts for a lot. Sometimes, when there are network effects, all that counts is to be the first to market with an okay product. Approximations can hopefully be corrected afterwards, when you have already beaten most of your competitors and try not to succumb to other businesses and institutions attacking you. 

Software in particular involves making 100s of these little assumptions and approximations. Project managers and back seat drivers want to enumerate and catalog these assumptions by interviewing and monitoring the developers. This is the bureaucratic approach to taming the chaos. It makes the whole thing take ten times longer, and it misses the unknown unknowns anyway. Since the top-down approach doesn't work, the number of necessary assumptions need to be kept to a minimum. The solution is to work on a narrower problem, where huge parts of the design space has been sliced away with big, explicit assumptions. That is why reductionism is critical for startups and researchers.

People are hard to make sense of with a reductionist approach, because they put up a fight. I take this as an a priori model, that people will over time develop a psychological resistance to any model that could be used for manipulating them. The only lasting model of human behavior that can be made then, is one that is so useless that it can never be used for exploitation. Advertising, propaganda, predatory economic systems; nothing works forever, they all have expiration dates.

So are 'reductionist tech' and 'irreducible people' mixed? Surely, if tech is supposed to be used by people in the end, surely we need at least one person who is less reductionist on our team? Well, that is not how user-friendliness is solved in the startup world. Successful startups just start with some intuit about what people want, that can be solved with scalable tech, and then work on that problem. The ones who get the intuition wrong simply die. The founders then make a new startup, or go back to working for the Man. Consequently, the "human" part is not solved on the individual startup level, but on the market level.