Hacker News Reader

Comment by seanmcdirmid | Hacker News Reader

Many weird nerds are on the spectrum and require very good people managers above them to mitigate their quirks, and should never go into people management themselves.

The industry has really changed though. It is much easier to hire social well adapted people who can code at all than to hire a bunch of people with social problems who can code really good. The former is scalable at least, the latter breaks down as no one can get along.

View on HN

Any non-trivial engineering or research is a team effort. It's good to have a mix of perspectives and temperaments in a team to keep poeple creative and honest. But the team ultimately has to cohere in order to deliver. That's really the core job of a manager: to create an environment where people can work together meaningfully and direct it to productive ends.

> Any non-trivial engineering or research is a team effort

There's plenty of counterexamples in the software space, to the point where most great software was originally written by a single person - C, Unix, Linux, Git, TeX, Python, Ruby, Perl, SQLite, QEMU...

Most of these now have huge teams maintaining them of course, but the initial research and engineering that was needed to come up with the golden egg is usually a one man show. C designed by committee is COBOL.

This isn’t true, most of the best software is written by individuals. Even inside large tech companies, it still happens that major value is created by one person.

I totally empathize with this, but there are hardly many people managers who can pull that off. And often those that can get promoted up quickly so that they aren’t managing teams directly anymore. It is much easier to produce with mediocre people managers with a team of socially well adjusted mediocre programmers, than a unicorn people manager with a diverse team of highly skilled programmers who aren’t necessarily great at social skills.

Individual efforts are completely different. A programmer like Notch can produce Minecraft on his own and then just get help to push it into a maintained product. This happens all the time, but you can’t produce that easily (and definitely not consistently) at the big corp level.

> the former is scalable at least,

I’d argue that it’s not due to communication costs. The more you attempt to parallelise a problem the harder communication gets.

> The industry has really changed though. It is much easier to hire social well adapted people who can code at all than to hire a bunch of people with social problems who can code really good

If you have no financial constraints? In what world do you spend massive amounts of money to mitigate people being "weird"?

The shift to high level languages like COBOL, BASIC, FORTRAN from machine language? The shift to GC languages? The shift to vastly inefficient but easy to use scripting languages like Python? There's a long history of trading off CPU/RAM/storage efficiency (i.e. massive amounts of money) to make computing accessible to less and less nerdy people.

Those also save time for all people? Which also saves money? I don't think anyone was like, "we need a language to save us from having to hire all these weird people!"

Reply on news.ycombinator.com