Ask HN: Why is it so hard to hire good engineers?

My beliefs that I'm in the process of validating further:

-> Poor resume filtering (based on keywords and proxy credentials) reduces the size of the candidate pool - biased to those that have worked at brand name institution, or went to brand name school, which leaves out many talented folks.

-> Technical interviews are low accuracy predictors of competence, which means you have to up-weight previous experience & credentials to reduce risk of bad hire. If you do choose to open up doors to interview candidates from non-traditional backgrounds, due to the greater variance in competence, you have to spend more time interviewing before you make a hire, driving up the costs.

-> Companies are biased towards cheaper, low-risk interviewing. They want to get the talent to meet the business needs, as long as they can do the job and fit the culture ==> paying more to get the talent that they know has greater chance to deliver is totally worth it.

A bigger company has the advantage of having more data about their hiring process, which could lead to more systematic experimentation to improve hiring. They also have a brand that would attract more people - bigger company often correlates with stability, and market rate pay, which are also drawing factors. They also have less risk --> a bad hire, that can be identified relatively quickly and fired, hurts a bigger company less than smaller ones, and they have more money to throw around.

Hiring at startups is tough: Your low-risk talent pool has other, great from a financial perspective, opportunities. Your higher risk talent pool is expensive to validate competence for because industry standard approaches are low-accuracy signals.

Fix interview process to give a higher signal, which allows you to reduce weight on credentials and experience, which allows you to increase size of candidate pool (because you've reduced the risk and cost to validate).

There's a lot of talent in the world that doesn't fall into the low-risk bucket that everyone is competing for.

Sure.

There's a name for it: The Bozo Event Horizon.

https://blogs.harvard.edu/waldo/2012/07/27/the-bozo-event-ho...

It takes a lot of sacrifice to make it to the top.

Should I be allowed to study while you enjoy your Valentines Day? Then I might just get ahead of you.

Luckily for us in tech you can make good money without being in the top 5-10%.

Dunno, his experience pretty much matches experience that I had in different parts of this industry in different roles all the way to being someone who only reports to a board of directors. It also matches the experience of lots of people that I know. It is the "mean" of all industries and even unicorns revert to the mean in time.

> Why would I do that? I believe if you want to get better at something, hang out with people who are already better. I get better if I work with better people and, as it happens, my bosses/shareholders are happy with that too because they end up with a more talented workforce.

That maybe true in some small companies. In fact, it probably is true for all small companies that survive. It is certainly not true in larger or large companies until those companies convert individual departments into entities judged by P&L and only P&L.

> It's no different to building a band: Paul, John, George and Ringo all knew the others were better at something than they were, and knew that was the secret to them all individually stepping up and beyond their previous abilities.

This is how the bands that do not make it out of a garage look like - the ones that are staffed by sad 40 year old men who think their music is awesome and they are just about to make it while working as waiters and bus boys in restaurants.

Successful bands are staffed by "rock stars" - the minimum acceptable level of performance is so high these people are not even in a market. There are no managers that are not rock stars at what they do and non-rock stars are definitely not telling rock stars what to do.

Real talk. Being able to define the term singleton or explain the difference between an interface and an abstract class is not much of a metric.