How to be a sane programmer

So, when I was growing up, where the family had to drive 50 miles to visit other Asians, I started to realize there was a pattern to how bumf#cks would come circling in for what was in their minds the justified pleasure of humiliating a non-white person. Most people were decent, but I and my family members stuck out.

I notice the same sort of circling at some programmer meet-ups here in the bay area, though I don't think it's my race that prompts it. Programmers make good money, are highly sought after, and in many companies are positively coddled. Just about every group that found itself in such a position has developed some small sub population of arrogant, entitled jerks. In retrospect, I may have been guilty of being a little bit of that myself.

One might like to think that good programmers are also good at thinking rationally in general. That's not how it works. Being good at programming just trains us to think rationally in a very specific context. The rest of the time, we're primates vying for a place in the social hierarchy. The best we can hope for are social structures and patterns that co-opt such tendencies for the uncovering of truth and the good of the group. Seeking social contexts that humanely succeed at this should underly one's seeking for a place of work.

It's hard to do this, and most of the time, people find a situation that achieves group cohesion and effectiveness through an "us vs them" dynamic.

That being said, I have some hand wavy opinions about the specialization/not specialization question being "at too high a level of abstraction", and that the focus should be on finding some way to impart "drive" (motivation?) into people, regardless of field and focus. I'd like to think on this more but I am currently late for work, so this'll have to do :(

I want to hire creatives who don't need to be delegated to, and can own full cycle of tasks on their own. I have been doing tons of research in the last two months and a lot of soul searching trying to figure out precisely what makes me tick.

The truth is, every startup WANTS to hire entrepreneurial developers, but rarely let them be entrepreneurs.

In going through this process on my own, for the second time, it is clear that if I want to hire people like me, I have to treat them like I treat myself.

Trust them, give them freedom, and be willing to ask for help when necessary.

First: software is nearly always a dicey proposition. There are so many variables that it's just impossible to predict much about it with any regularity. The best approaches that I've seen so far accept this, and say, "Trial and error? So be it: we will do trial and error as best as it can possibly be done." YC and the lean startup model (aim low, accept that most things will fail) are good examples.

The other approach is the "tried, no error" method, which cares intensely about the past. Hiring a web developer? Portfolio first, examples of use of specific technologies, examples of specific features. Find no one who meets your exact qualifications, and proceed to either a)raise salaries a lot, and constantly (banks, hedge funds), or b) complain to Congress about a "talent shortage" (everyone else).

I think YC is the better model, and is a sort of company-less version of michaelochurch's "open allocation." The key insight is: if you want 10x engineers to ride up on their white horses and save you, you can't presume to tell them what to do. They're cars, not faster horses. Beating a car with a whip won't make it go faster---and this is a feature, not a bug.

To get the best out of developers, you have to accept that they bring risk, and manage around that, rather than blindly trying to reduce it. That's why I think Apple is really fucking smart to have a huge cash cushion like they do: with that cushion., it frees you to go for long shots. Apple is nearly riskless in its financials, and so can afford to be very risky in its product development process.

If I were starting a software company, and knew exactly what I wanted to build, I'd hire people with proven track records, pay them a bazillion dollars each, and give them zero autonomy. Or, I'd hire a bunch of college dropouts who've obssessed over some kernel of good ideas (Tufte, Hickey, FRP, ML), pay them nothing, and let them run wild (this is basically YC's business model).

If you're really serious about this---and I think it's worth it to be---I'd read everything on michaelochurch's blog. He exaggerates, has a chip on his shoulder, uses weird analogies, and has a better understanding of these issues than anyone I can think of.