Hacker News Reader

Comment by thehardsphere | Hacker News Reader

I actually read some of the research that produced the 10x number.

The 0.1x developer, if they're actually competent enough to remain employed for more than a probationary period, probably isn't at 0.1x; you're probably underrating them because you're better than average and are thus overrating what average is. (This is part of the Dunning-Kruger effect).

-10x developers should be fired fairly quickly if they're actually that negative and the business is being managed competently. I've worked with a few; they were fired within months or weeks.

View on HN

I've always considered the Dunning-Kruger effect from the perspective of "too dumb to know I'm dumb" not "too smart to know I'm smart".

Interesting to think it goes in both directions!

It's a common misconception that the Dunning-Kruger effect is just about dumb people not knowing they're dumb. In fact, it's about something we all do - overestimating our knowledge and competence in areas where we have little of both. It's worth recognizing the distinction because everyone is subject to the Dunning-Kruger effect when we're out of our depth in some area and there are a lot of "unknown unknowns".

Yeah, most people know "too dumb to know I'm dumb." The "too smart to know I'm smart" is less well known but it has a lot of implications.

For instance, highly skilled people will often underestimate the difficulty of a task for average skilled people, because they complete it with relative ease. It's effectively an overestimation of what "average" is, because they think of themselves as average when they're really not.

"Relative" is the key word here. Your natural high skill performer may think a task is difficult, but they don't necessarily find the same sources of difficulty as average people.

I've always heard the "too smart to know I'm smart" as manifesting differently. Getting better improves two distinct sub-skills: excution, and perception of performance flaws. Dunning-Krugar effect, to me, just says that flaw perception grows faster with skill than execution does.

At a concrete level, I remember watching a lecture from a Jazz instructor by the name of Hal Galper, who said something like "Not liking the music you're playing is a sign that you're getting better, since it means you can now hear the mistakes you've always been making". It's really stuck with me through learning other creative pursuits (dance, in particular).

I have heard someone else say a similar thing about writing, and I believe it. They described it as "having or developing taste", and were encouraging people to keep writing through those feelings because as they got better it would help them know what to write and what not to write.

Dunno if I've seen it in software development, but I also don't have any real objective estimation of how good I am at software development.

It does, and has since its original discovery. Though the effect is not as pronounced as its popular conception would have you believe. The poor performers overrated their performance, but did not rate it higher than the high performers did. It is more like the grandparent observed that everyone tends to think of themselves as close to average.

No. The negative ones usually form crowds and alliances, and with a lot of backpadding and self management it leads to your typical failed projects. Many of them are very high-profile, and they would never have the insight that they are actually more destructive than useful. That's why forks do exist.

Reply on news.ycombinator.com