Hacker News Reader

Comment by lores | Hacker News Reader

People on HN skew young, smart (in one domain), and tend to live in a bubble of similar people. If you know you're smart, the smart people you talk to validate your smartness (in one domain), society validates it some more by paying you massive amounts, and you're not experienced enough to know better, you're bound to overestimate your abilities and knowledge.

It needn't be most, or even many on HN, and people of all kinds vastly overestimate their abilities. It's just that on HN it's overestimating with great ambition.

(I say this very confidently, don't I?)

View on HN

So funny to jump to the "they're just kids" explanation for this when we are literally talking on a forum hosted by a VC incubator.

Is it not Occam's razor that people are like this because this world of startups, "cutting edge tech", "move fast and break things", etc. gives quite clear incentives to be like this? The entire of financial world of tech is quite significantly propped up by the inertia of unearned confidence!

> If you know you're smart, the smart people you talk to validate your smartness (in one domain), society validates it some more by paying you massive amounts, and you're not experienced enough to know better, you're bound to overestimate your abilities and knowledge.

And then you become the richest man in the world and buy Twitter and show everyone that you're kind of just clueless outside of your area of expertise, but putting up with you is profitable enough that people just go with it.

Also, lets not discount the fact that people can have a lot of success stepping out of their core domain.

People can do this repeatedly with positive feedback and increasing scope until eventually it doesn't work.

You sound like an expert in psychology.

Eh, I'm nothing of the sort, I'm only advancing in years and have made it a point to exist in as many segments of society as I could. I was that cocky engineer once, my words are only anecdote from first-hand experience and observation. I never expect to be right, only hopefully more right than wrong.

How would discourse change to eliminate this problem? Should we only speak about topics we are employed in? Lead each comment with a summary of our qualifications, or a proclamation of humility where we signal how little we know?

I mean, hypothetically MS or Meta could already automatically do this...

I know you jest, but I think it wouldn't be a bad idea at all. There are languages whose grammar forces the speaker to explicitly clarify the source of information; Eastern Pomo, for example, has different verb forms for whether it's something you know first hand, saw, are repeating, or deducing. I imagine it's not only useful for the listener, it also helps the speaker realise if maybe they are building a shaky argument to make a point. I, for one, would be interested to see that system in English, it could lead to interesting developments.

See, you put the caveat at the bottom, but I think you are just having a normal discussion. You aren't speaking "very confidently," you are just making an argument.

What I think happens is people who are very knowledgeable about a subject are hyper-sensitive to slightly incorrect information. And to boost their egos they like to diminish the people making the incorrect statements as not just incorrect, but confidently incorrect, a la Dunning Kruger.

See how confidently I made the exaggerative statement above? I don't necessarily mean it to be completely true, but I am making an argument. I think an assessment of confidence requires more than seeing no mollifying qualifiers like "I think" or "it might be". There's no verbal tone on the web.

It was a little meta-joke, but I think the world could use a lot more expressions of doubt. Very few things are certain or universally true, and those that do tend to have Greek letters in them. I find highly confident people highly suspicious, and a culture that rewards overconfidence and punishes doubt both exhausting and dangerous.

Reply on news.ycombinator.com