It's a big assumption that the competitive interview process is sufficient to identify only the top talent. Every other week News@YC has posts complaining that every type of interview process doesn't identify talent well.
I didn’t go into AWS as a developer. I worked there in Professional Services. I shadowed a few coding interviews and conducted a few system design interviews of software devs while there.
You can see the results of this in Amazon's consistently poor companywide performance in every arena. They're even managing to lose ground to Microsoft in cloud computing. It's pathetic.
I surmise large number of engineers there are just collecting bounty on Amazon monopoly on cloud infra. Their skill and work is not more challenging or interesting than hundreds of thousand engineers at 20-30% income of Amazon in boring enterprise IT shop.
Another former coworker who was a project manager (engagement manager) got maybe 60% of their compensation and is doing more work.
Myself personally, I was an L5 at AWS and now I’m doing at least L6 level work and I’m still making 15% less than I was making at Amazon.
It’s going to take a couple of years at the company I’m at now and building out an entire specialty practice - L7 level work - to get to my L5 compensation.
I have a roadmap I designed with my manager to do it and it’s a straightforward process, it’s just going to take time. I’m excited about doing it. It’s the scope/impact level of work I did to get into AWS on a much smaller scale at a startup.
Generally a mix of people who exhausted other paths to higher compensation by their mid 30s - mid 40s (senior management track / trader track) and this was the next big thing.
We're they from a top tier IB, top tier HF or Prop trading company?
Or from lower tier IB/HF/Trading company or worse one of the 1000's of fintechs.
Source: I worked at AWS ProServe for a little over 3 years and saw how hard it was to detangle BigEnterprise from Microsoft
Google already seems like an also-ran, and did you even know that Oracle has a cloud offering too?
>time and motivation to grind
>pass an interview
Let's be honest, what more are we looking for?
I'm not huge fan of Google, but will readily admit that, on average, the ex-Googlers I've worked with have tended to be great engineers and very sharp people.
So, at least from that anecdote, there does seem to be a difference in each of these hiring filters.
It's all about gaming the system here.
I have literally seen people justifying lying on the resume, by feeding the job description to ChatGPT and let it come up with a resume.
Yet, there's not much influential software coming out of India, where leetcode hards for graduates are the norm, but these graduates don't understand what a virtual machine is.
On the other hand I have seen prolific open source contributors in the same age bracket (they contributed not to some web design project, but very much used cloud / system software) struggling to get a job, because they didn't do enough "DSA".
I am not saying people of my country are dumb. Quite the contrary.
I am saying when a measure becomes a target, it ceases to be good measure.
That's the case with leetcode problems.
It selects for people who can grind through it. Not the people who have other interesting things to do.
> prolific open source contributors...struggling to get a job
The market selects the best, if it's best at lying, who are we to tell the invisible hand that it's wrong? /s
I've seen environments where a certain amount of manipulating the truth is expected. Individuals eventually figure out what's expected of them.
You know it's not the truth, they know it's not the truth, and you know they know.
Google, for instance, has a slow tedious hiring process with a lot of false negatives, but in my experience those who made it through were absolutely qualified and great to work with. There certainly wasn't 30% to cull. In my opinion there wasn't even enough people to fill the percentage quota that need to be given ratings lower than satisfactory - before the layoffs. After them, the fact that they're still enforcing these quotas on teams is absurd cruelty that kills morale.
The assumption is that though interview processes are acknowledged as imperfect, they're largely considered to measure something like productivity or worker value.
Under that assumption, then it should be exceedingly difficult to "cheat" the interviews such that a person could do great at an interview, and be of low "worker value". Not impossible, but difficult. And I'm just pulling stuff out of my ass here, but if we were to do a survey here on HN, the most likely consensus on how difficult that is would be something like "it's actually more difficult to fake it in the interview than to just be good enough at what you say you're good at".
Imagine if you were one of the fastest sprinters in the world, but to be "hired on" to the Olympic team they decided to interview you for it. Sure, they'd probably ask in the interview what your latest 100M times were, stuff like that, but after the interview they'd all talk privately about whether or not you were a good "culture fit", and whether your smile was sincere or fake, whether you were too nervous or smooth and a "real go-getter". In such an environment, the Olympic team might well have 30% of its members that could be safely culled. The only jobs interviewing is even a good process for are those jobs where the actual work itself resembles interviewing... sales positions, PR, being the "public face" of an organization (news anchors, I dunno). The reason we use interviewing for technology jobs is because that's what the big bosses know works for their jobs, a CEO is the public face of the organization.