4 ms·

The FAANG paradox : these companies all have this performance-intensive culture with the gruelling interview, "perf cycles", promotion processes, OKRs, etc... t

by pyb 4y ago

The FAANG paradox : these companies all have this performance-intensive culture with the gruelling interview, "perf cycles", promotion processes, OKRs, etc... to ensure all engineers function at the top of their capacity. At the same time, their new public facing products have been mostly underwhelming over the last decade. It seems most of the innovation occurred back when FAANGs were 10-100x smaller. (The AI work is one exception)

yaseer 4y ago
It's interesting to see how Big Tech firms can struggle with innovation as they scale, but perhaps for different reasons. Facebook has lacked innovation for quite a while, first relying on acquisitions, then brazenly copying other social media company innovations (e.g. Snapchat, TikTok) Google innovates, then seems to give up on products way too early. It's almost as if their management culture has forgotten that new products need iteration. Amazon, to their credit, have managed to innovate better than these two (e.g. AWS, Kindle, Alexa devices). They started off as a book store, yet somehow became a Tech conglomerate.
mjburgess 4y ago
This I think is a bit like the "IQ selection" phenomenon, where if we suppose IQ actually measures (say,) test-taking ability rather than some complex notion of intelligence... then we find places that try to filter on IQ (law schools, med schools, FAANG) end up selecting for uncreative kinds of intelligence. As a bold, perhaps wrong hypothesis, this might be in play here. We have the very best leet-coders, but the very worst programmers: in that sense of programming Peter-Naur called "theory building". An activity which is perhaps the most comprehensively demanding on the intellect. Not rehearsing the techniques and solutions of others, but innovating oneself in the face of ambiguity and ill-posed problems. Perhaps the worst sin of IQ-test-like processes is that they use questions with answers to test intelligence -- which if it is anything at all, is what one uses in the face of the lack of questions with answers.
- afiori 4y ago
  Probably if you filter by raw IQ (either yours or others definition) you would get a lot of imaginative people as most intellectual faculties are positively correlated with each others. On the other hand if it turns out that you are actually filtering by "willingness to take years of testing" then you might get different results
  - mjburgess 4y ago
    I think these correlations are often spurious. Low IQ correlates with poor performance, making it clinically useful. I dont think there's much to be said about >115, other than by filtering on it, you at best reduce the variance in other abilities.
  - gedy 4y ago
    Exactly, I'd have no problem taking an IQ test for a role, and usually test high on those. Leet code grinding and interview "preparation" I just have no time and willingness for.
  - badpun 4y ago
    > On the other hand if it turns out that you are actually filtering by "willingness to take years of testing" then you might get different results That's a very useful signal still. Willingness to takes years of testing signals willingness to subjugate oneself to lots of unpleasant crap for money and prestige, which is something FAANGs are happy to see in their workers.
- pyb 4y ago
  There's adverse selection at play in FAANG hiring, but I think it goes like this : any ambitious engineer would wants to "make a dent in the universe" is never going to spend months studying LC, system design, or whatever to get into these places. Not only do they prefer building actual projects on evening and weekends ; but they know their time at FAANG would be wasted anyway. In 2022, few people go to big tech because they really want to build things.
  - 4y ago
nivenkos 4y ago
They hire the best people and then have them do effectively "menial" work - metrics reporting, etc. whilst strategy is decided by executives alone. This makes them disillusioned. I'm not sure what the solution is in big companies really, but it's a real problem when standard engineers can't get their ideas or features, etc. considered, or easily prototype stuff, etc.
Nextgrid 4y ago
I don't think internal company culture is the problem. The problem is that a lot of these companies' consumer products are based on the "attention economy", and the market has finally adjusted to see these things for what they are (essentially a system to put more ads in front of eyeballs) and thus considers them "underwhelming". The "A"s in FAANG which aren't based on the attention economy (Amazon, and Apple until very recently) don't (yet?) have this problem.
mattbrewsbytes 4y ago
The only reason I could personally see working at any of them is greed. I've worked at large corporations before and from many accounts on here and elsewhere I get the sense that there are many of the same problems which essentially boil down to being a cog. You might be a cog in fancy wheels with catered lunches and bean bag chairs but in the end its probably difficult to see where contributions are helping actual customers. I've also heard similar accounts about working for various government levels and suspect its the same conundrum. I think its unavoidable when humans organize a workplace into thousands of people.
theGnuMe 4y ago
It's really something to take heart and think about. "We only hire the best!" And yet the best are apparently not good enough. I don't think they have enough evidence to reject the null hypothesis that their hiring criteria is a joke.
ThrowawayR2 4y ago
Engineering ability has nothing to do with product management; critical feedback from engineers is even usually met with "Well, that's because you're not the customer." or "You just don't get the vision.". And, to be fair to them, their rate of failure is probably no worse than the odds of a tech startup failing (90%?), it's just more visible because anything a FAANG releases is trumpeted with fanfare. All that the high level of engineering talent at FAANGs allows is for bad product management to drive themselves off a cliff faster. (But at least the engineers get paid handsomely for it.)

by pyb 4y ago

yaseer 4y ago

It's interesting to see how Big Tech firms can struggle with innovation as they scale, but perhaps for different reasons. Facebook has lacked innovation for quite a while, first relying on acquisitions, then brazenly copying other social media company innovations (e.g. Snapchat, TikTok) Google innovates, then seems to give up on products way too early. It's almost as if their management culture has forgotten that new products need iteration. Amazon, to their credit, have managed to innovate better than these two (e.g. AWS, Kindle, Alexa devices). They started off as a book store, yet somehow became a Tech conglomerate.

mjburgess 4y ago

This I think is a bit like the "IQ selection" phenomenon, where if we suppose IQ actually measures (say,) test-taking ability rather than some complex notion of intelligence... then we find places that try to filter on IQ (law schools, med schools, FAANG) end up selecting for uncreative kinds of intelligence. As a bold, perhaps wrong hypothesis, this might be in play here. We have the very best leet-coders, but the very worst programmers: in that sense of programming Peter-Naur called "theory building". An activity which is perhaps the most comprehensively demanding on the intellect. Not rehearsing the techniques and solutions of others, but innovating oneself in the face of ambiguity and ill-posed problems. Perhaps the worst sin of IQ-test-like processes is that they use questions with answers to test intelligence -- which if it is anything at all, is what one uses in the face of the lack of questions with answers.

afiori 4y ago
Probably if you filter by raw IQ (either yours or others definition) you would get a lot of imaginative people as most intellectual faculties are positively correlated with each others. On the other hand if it turns out that you are actually filtering by "willingness to take years of testing" then you might get different results
- mjburgess 4y ago
  I think these correlations are often spurious. Low IQ correlates with poor performance, making it clinically useful. I dont think there's much to be said about >115, other than by filtering on it, you at best reduce the variance in other abilities.
- gedy 4y ago
  Exactly, I'd have no problem taking an IQ test for a role, and usually test high on those. Leet code grinding and interview "preparation" I just have no time and willingness for.
- badpun 4y ago
  > On the other hand if it turns out that you are actually filtering by "willingness to take years of testing" then you might get different results That's a very useful signal still. Willingness to takes years of testing signals willingness to subjugate oneself to lots of unpleasant crap for money and prestige, which is something FAANGs are happy to see in their workers.
pyb 4y ago
There's adverse selection at play in FAANG hiring, but I think it goes like this : any ambitious engineer would wants to "make a dent in the universe" is never going to spend months studying LC, system design, or whatever to get into these places. Not only do they prefer building actual projects on evening and weekends ; but they know their time at FAANG would be wasted anyway. In 2022, few people go to big tech because they really want to build things.
- 4y ago

nivenkos 4y ago

They hire the best people and then have them do effectively "menial" work - metrics reporting, etc. whilst strategy is decided by executives alone. This makes them disillusioned. I'm not sure what the solution is in big companies really, but it's a real problem when standard engineers can't get their ideas or features, etc. considered, or easily prototype stuff, etc.

Nextgrid 4y ago

I don't think internal company culture is the problem. The problem is that a lot of these companies' consumer products are based on the "attention economy", and the market has finally adjusted to see these things for what they are (essentially a system to put more ads in front of eyeballs) and thus considers them "underwhelming". The "A"s in FAANG which aren't based on the attention economy (Amazon, and Apple until very recently) don't (yet?) have this problem.

mattbrewsbytes 4y ago

The only reason I could personally see working at any of them is greed. I've worked at large corporations before and from many accounts on here and elsewhere I get the sense that there are many of the same problems which essentially boil down to being a cog. You might be a cog in fancy wheels with catered lunches and bean bag chairs but in the end its probably difficult to see where contributions are helping actual customers. I've also heard similar accounts about working for various government levels and suspect its the same conundrum. I think its unavoidable when humans organize a workplace into thousands of people.

theGnuMe 4y ago

It's really something to take heart and think about. "We only hire the best!" And yet the best are apparently not good enough. I don't think they have enough evidence to reject the null hypothesis that their hiring criteria is a joke.

ThrowawayR2 4y ago

Engineering ability has nothing to do with product management; critical feedback from engineers is even usually met with "Well, that's because you're not the customer." or "You just don't get the vision.". And, to be fair to them, their rate of failure is probably no worse than the odds of a tech startup failing (90%?), it's just more visible because anything a FAANG releases is trumpeted with fanfare. All that the high level of engineering talent at FAANGs allows is for bad product management to drive themselves off a cliff faster. (But at least the engineers get paid handsomely for it.)