At that point, they are not hiring you, me or anyone who directly applies themselves via the job form.
They are hiring their friends from elite circles.
In the case of OpenAI you also have interesting tech and a brand that will massively accelerate your career if you want to stay in that field. So while yeah, you have to hire the best people; and OpenAI like everyone else will be paying $LOTS to a few useless engineers in the mix, I think "$900K and everyone knows it" is a pretty good substitute for talent-spotting, which anyway can't be bought.
If one (“talent") is a stellar runner, but when the team goes on a run the "talent" has to wait for their teammates to get to the finish line, most of whom go running once every two months, and the "talent" gets dress shoes instead of spikes etc., the "real talent" is worth, to the team or company, as much as the average employee on that team.
An easy and naïve recommendation would be to remove the constraints that limit the work of talent, but it is much easier for talent to move to organizations that, for a variety of reasons, are more conducive to the expression of their abilities than to fight the inertia of organizations. Those very organizations which, for one reason or another, have existed for a long time and pay salaries to hundreds or thousands or hundreds of thousands of people.
offering a lot of money attracts everyone including in-demand talent.
offering less turns away in-demand talent.
How do you quantify best? Number of degrees? Publications? Association with prestigious institutions? Past work experience at top companies? Speed of problem solving? All of this is gameable once it starts being _measured_ and enough incentives exist for people to devote their life to winning the "game".
However, if you happen to hire a math olympiad winning PhD with numerous publications from a tier-1 research institution with a known track record in industry, it would be hard to argue they aren't the best. But success breeds success, and top people will keep being poached to other top places. Kinda how money makes more money.
Comments of this nature would be much more informative if were they to begin with "what I've seen is in my professional life", "in the three or four teams I've worked on", or "according to some friends, who consider themselves to be top engineers, working in teams similar in size and scope."
Otherwise, we are left with the sometimes realistic and sometimes much more fairy tale-like story of incompetent leadership holding back talent for no other reason than incompetence or nepotism or envy of those who are smarter and more accomplished.
I have been in similar situations and considered myself a top professional, which may well be true, but in large, mature organizations, regression to average performance is largely an inevitable consequence of size and the need for coordination, a problem whose solution is not to be found in a redistribution of salaries.
How else, with a few exceptions, do employees at Google, Facebook, Uber, etc. think that back in the day things were so much better, that talent was treated better, that there was a real engineering culture, whereas now it's all about sitting in chairs, people in finance having the most power, interviews were so much harder, and we have so much technical debt?
OpenAI is, at this point, a research organization, in spirit and in purpose. If and when it becomes a "normal company," the logics of scale and scope will lead the early employees to complain about how things have evolved. But I suspect that the "let's get rid of the makers of technical debt and use the budget to give top talent more money" will not produce the expected and desired result of a renewed engineering culture, because top talent will not be as useful as it once was.
Whenever a top executive leaves a mature company (or dies, see Apple), the risk of catastrophe is aired, but catastrophe rarely occurs. There is a lesson there.
Both may be benefiting openAI here. There’s lots of places to work on LLMs but “GPT” is a product/brand that people have heard of, and if OP is to be believed then they certainly seem to be paying “enough”.
But keep in mind that Google regularly launches new products and features that are more profitable than many startups, and then shutters them, and regularly launches individual features or pieces of infra that are industry leading (or defining, as in the case of k8s). They just aren't new consumer products.
However what this viewpoint doesn't account for are team dynamics. A strong TL can turn NNPPs into incremental positive contributors. A great programmer without leadership capabilities will not be able to outpace the technical debt. There are also more subtle dynamics depending on the structure and personality traits of the individuals. Ultimately programmer productivity is not an absolute value, it depends on the whole ecosystem (including other functions, leadership stance, etc). After doing this for 25 years (IC, TL, EM, CTO), I strongly believe a healthy team is about harnessing and orchestrating different individuals unique strengths rather than trying to set too high a bar—the latter will lead to counter-productive competition and ultimately burn out your best folks.
Personally, from my life in tech, I do not feel that OpenAI has done a great job (and rather frankly, work that has been supported by both "press" and popular sentiment, because who doesn't like the heroic effort of a group of smart people facing poor odds against the Goliaths?) because management in Google or Meta cannot recognize who is writing great code.
Think about the problem of "hallucinations" with GPT. After all, it was considered a minor hiccup on the road to the AGI, a path opened by a team of mavericks. But if Google, had it been first to market, had delivered such a product, the press would have gone from "oh, that's funny, it will get better with time" to a more worrying "Google is destroying humanity with those hallucinations."
It is much easier to be innovative when you are small, hungry, with little to lose and much to gain, rather than when you are worried about your current salary or equity or reputation. It's not just a matter of paying top talent more and getting rid of more average people; I'm sure there are enough brilliant, highly paid people who have enough capital to build small, high-IQ teams in any of the major technology companies to get to GPT-like models before OpenAI. But incentives, reputation, the nature of public companies play a role in being slower, less innovative, less risk-taking.
As a concrete example, I've gotten more accolades for silly personal projects that sound impressive, like training a convolutional network to pilot a simulated car on the GPU, than for impactful work at my actual job, which was a lot less challenging.
I guess hiring is just incredibly noisy, and I think companies could really get far hiring less than the best people, and just squeezing good quality work out of them (I believe Amazon is known for this).
Obviously OpenAI should not hire subpar people lol, they should keep doing what works for them, just grumbling loudly here.
Seems silly enough to get you somewhere cool. Though the best pass is recommendation. Impress your colleagues so bad, they will recommend you somewhere one day.
than for impactful work at my actual job
Oh you have to fight hard to get one of those. You cant just do what you're told. Ive got a cool story to tell from my time at amazon but I fought heroically to get it (I was younger though).
I guess hiring is just incredibly noisy
Careful what you believe because your belief becomes your reality.
If you want to be great, you cant just act and think like anybody else. Work on yourself and be great. Get the mindset first.
"life has strange ways"
We’ll, by the number of Leetcode Hards they’ve completed, of course! /s
* have a set of interview questions that help clarify they have the "right" kind of person
* have a performance review process that ensures that hired individuals continue to engage in ways that are understood to be most useful to the company; i.e. that validate (or invalidate, as the case may be) hiring practices