I don't think the "rockstar" act is a millennial thing. There have always been people like that in any field, or social group. Some people will always compensate for their flaws in ability with their strength of personality, and with our industry that just means painting the "rockstar coder" persona, because there are some people that seek that out.

That's why good hiring practices have to be used, even when you feel like it's a waste. Give everyone the same critical evaluation and don't allow yourself to skip questions in interviews because you like the person, etc. (I'm not accusing you, just saying as a rule)

I like using the word "rockstar" to describe coders, because the word has similar negative connotations as in the music context. An actual rock star often doesn't (but sometimes does) have: classical music training, knowledge of the music literature, vision beyond their area of expertise, humility in the face of their equals or betters, a desire to learn and ability to change with the times, the temperament to fit well with the business side, etc.

If that's what you want to hire, go hire your rockstar coders.

The hiring process spends all its time to evaluate if a candidate meets minimum viable functionality. Instead my recommendation is to evaluate all candidates as senior developers and set them up for anxiety failure and watch how the handle the result.

I can train somebody to be a good coder, but I will not be their parent.

Sad prediction: overly confident will have better careers and will be seen as geniuses by people who don't work directly with them. They will have more responsibility sooner and even as they will make mistakes inittially (and some entirely), they will he excused and seen as leaders more then their humble peers.

Incentives.

This only works in big or medium companies, fortunately being over-confident without actually delivering stuff gets caught pretty soon in small companies (because it directly and adversely influences the bottom-line). In other words, small-company bosses/owners don’t really care of how confident you are, they only care about actual work that makes the clients happy, so that the bills and the salaries can be paid at the end of the month. That’s one of the main reasons why I’ve worked only for small software shops in my career, I can’t stand the wannabe people who always brag about themselves without actually doing anything of value.

This works also great in startups where nobody really knows what he is doing and no one has firm grasp on what people produce/should produce except impressions. Knowing who really produces requires good management that has time to look at the details. That requires stability, non panic mode and sane politics.

Overworked manager will take overconfident person over humble cautious one, because the illusion feels good.

Those overconfident people rarely do nothing and are completely incapable. That is strawman situation, basically. Oftentimes they eventually succeed, altrough it takes more time, is overall somewhat less effective, their code is harder to maintain or something of that sort.

Marissa Mayer - https://en.wikipedia.org/wiki/Marissa_Mayer

By all accounts she seemed to be a great product manager, but product management isn't organizational leadership.

Likewise there is Hugh Jones - https://www.sabre.com/insights/releases/hugh-jones-to-head-t...

Hugh Jones was known for tenacious leadership in large volume business operations, but experience in retail and product was limited or absent.

There are many, many companies outside of SV where a decent fresh grad (maybe with some co-op experience in SV) will be better than the "senior" developers at those companies.

Not that the fresh grads are awesome in any absolute sense, but there are so many awful developers out there who made it to "senior" level at their clueless company just because they stuck around for more than a year or two (really, because they were unable to find better work).

I wouldn't use the incompetence of others as a safe validation for one's own incompetence starting out.

That is taking a horrid situation, lowering the baseline to compensate, and then filling to the lowered baseline. The result will be worse than the incompetence you already have. The positive resolution is to fire (or move) the incompetent people and change the culture of the team.

Let me give an example if you don't believe me: I once argued with a senior dev for an hour that his code had a race condition and when he finally saw it he merged it anyways because it was "unlikely to happen". He was also "too busy" to write tests.

He probably thought I was an arrogant millennial too.

Dunning-Kruger Effect : https://en.wikipedia.org/wiki/Dunning–Kruger_effect

First thing I thought of too - the only category where students ranked themselves as less proficient than they were perceived to be is technology - the one area they are probably actually proficient in.

Did they goto private (public in the uk sense) schools? those type of schools inculcate there students with a master of the universe vibe.