Except that you can teach them things.

It's hard to teach an unfriendly person anything, and it's impossible to teach an arrogant asshole anything at all, because they think they're better than anyone else.

There are exceptions of course, but friendly and unteachable and unfriendly but very teachable are both very common scenarios. Teaching the unfriendly person might not be fun but it still works.

OTOH, for software, the most important skill is resourcefulness, and that seems like more of an intrinsic property of someone's personality.

I would argue that arrogance ebbs with experiencing or observing failure just enough times.

Failures and their lessons usually correlate with age. However, it is not unusual to find people with enough age and nous yet lacking the graciousness you expect from experts/mentors.

So I'm personally not like that though in most situations. I think it's counterproductive with most people and often rude. I think the above is a bad strategy in more than 60% of cases (scientific number; totally not pulling that out of my ass), because most people take direct criticism personally or become defensive.

However in my closer professional circle I very often prefer uncomfortable no-bullshit type of style. It's simply more _effective_ and clear.

I've heard stories of "brilliant assholes" but once I've peeled back the layers of these stories I always develop a strong impression that the brilliance is a facade.

And the other component is that assholes often times end up dominating rooms and looking brilliant, because they effectively bully others into being silent. For most people, being silent is better then risking the asshole will target you.

Competence is not necessarily linked with a lack of humility.

In my experience, a brilliant asshole is worse. He's less likely to be terminated, and causes problems for everyone. Incompetent people are self aware of their incompetence. Brilliant assholes rarely acknowledge their problem.

The one time I dealt with a brilliant asshole - oh wow. He would be right 90% of the time, but for the remaining 10% there would be no way on Earth you could convince him he was wrong. You could bring evidence, mathematical proof, anything: He just wouldn't listen. It got to me being very careful that he not be around when I'd ask for help - because he often misunderstood my problem and would then insist I implement his solution, and there was no way I could convince him that he misunderstood the problem statement. If I ignored him and implemented a different solution, he would throw a loud tantrum. And he had no stake in my work - we were working on different projects.

I spent two years in that team and every time he acted up I started documenting it.

I never complained (it was clear the manager didn't want to deal with people problems), and on the outside I didn't let my frustration show. I now hear that another member of that team is really complaining to the manager about him. I reached out to him and let him know that if he wants to escalate with HR, I have plenty of material to provide.

The one nice thing with incompetent people is you at least look better when it comes to reviews. I know in one job I had I ended up slacking quite a bit, but I knew it wouldn't hurt me because they had quite a few people at my grade level who were just plain incompetent. Management isn't going to give the whole team a poor review.

Except of course for the perfectly (if narrowly) competent but decidedly jerkface and/or outright asshole teammate. Who is even more likely to get a pass for being such a "rainmaker". Or because "it's crunchtime and we need all the firepower we can get our hands on".

Which was basically the article's point.

What about a very competent but toxic personality who ends up preventing contributions from other team members, because when they contribute they get belittled or bullied?

At worst, the former is a neutral in his overall contributions, whereas the latter will be a net negative.

No, with an extra teammate the expectations on your team increases, so you have to work hard enough to pay for that guys salary as well.

Lets extend this a bit, would you prefer to work in a team with 5 incompetent but nice persons who don't contribute anything, so your work has to be enough to pay all their salaries, or would you want to work with 5 assholes who work hard enough that you can slack all day because they pick it up? I'm sure a majority would prefer the second scenario, the first scenario would lead to burnout really quickly and soon they will call you the toxic asshole genius.

I've worked with people that refuse to do anything in git other than commit, which means:

    - new copy of the repo every time origin changes
    - spending a week manually rebasing a dozen commits (obviously with many unique mistakes each time)

and refuses to upgrade their monitor from 1366x766 even after being given the money to do it, because "it's not necessary"

This reminds of people who try to defend someone accused of incompetence: "but s/he's so nice!"

If you could translate "friendliness" and "competence" into equivalent units, I think competence would be devalued compared to friendliness, at least once you get into negative values (e.g. for every extra point of unfriendliness, you need 5-10 competence points to make up for it).

The easier you make doing my job (eg, I can get my work done without you adding bullshit code that I need to make my way around), the more I’m willing to forgive almost anything.

Just be invisible to my work plans for the day, and we’re all good.

Did you met unfriendly and incompetent? Cause those really sux.

The worst is "jerk, average competent, but people who don't work directly with him assume he is genius because he is jerk".

This is the wrong way of looking at things. You shouldn't compromise on friendliness and you shouldn't compromise on competence. If you have to compromise, you compromise a little bit on both but not too much on either. An unfriendly teammate is very bad.

If they ask for help and won't just go missing for a week or something like that, then its not a big deal if they need some hand holding on a more complex task. Reliably mediocre is better than unreliable, any day.

Anecdata: colleague of mine at least twice a week reboots production servers during the day, sometimes by themselves, sometimes along with their ESX-hosts...

His response? "Oops."

Does he ever learn from it? Doesn't look like it...

I mean, i'm not exactly working at a hospital or something - i.e. no peoples lives on the line, but try to explain for the n-th time to someone whose last hours work was lost because someone couldn't be bothered to check if he can reboot that server now or not?

Not to mention that the "research paper" industry is often very manipulated, inaccurate, and sometimes downright fraudulent. Ironically, also discussed on HN once a blue moon.

What I am saying is, do site the "research papers", but don't use that to shut down an argument. It's a clue, not a fact.