100,000 of the best out of a pool of approximately 10 million professional software engineers worldwide is a sizeable portion. Additionally, not all 10 million are even close to being up to BigCorps peculiar standards (perhaps the standard is "someone competent enough that they could potentially build a competing product line").
Goog, Meta, Amazon, Apple, Oracle.. all have huge rosters and it adds up to a significant portion of the market. There is also a huge amount of medium sized whales such as SAP and PayPal who in total end up also consuming a lot of the talent.
Well, Google's and Facebooks definition of 'the best' engineers.
I don't think grinding leetcode for an interview is the best indicator of a good engineer, and graduating from a prestigious university is not always an indicator either. imho it seems like the best engineers now are the ones doing their own thing outside of the large companies, or are at smaller startups.
No, but if all of FAANG are hiring by that criteria, it still works; startups can hire good talent because they break the pareto equilibrium, but that's ok for FAANG because they obtain that tech through acquisition after the idea and execution are derisked. The system works!
That would surprise me. I have attended targeted career fairs with both FAANGs and national labs recruiting, and the national labs give off way more 'work-life balance' vibes. Plus, as the largest bureaucracy in the history of the world, the federal government isn't a good place to get a high return on brain damage when you want to actually get something done.
Having said that, the national labs do seem like good places to go geek out in your own advanced intellectual cul-de-sac.
Well from experience of being an undergrad and going to career fairs, this assessment is spot on. You don't realize this whole thing is bullshit until a few years into your career.
Because my frame of reference is being early or maybe early-mid career, where you can't possibly have the necessary experience to be 'best' without working significantly more than 40 hours weekly, and from my perspective most of getting there in the future follows that path too. I'm not discounting that some top engineers could exist outside of working a lot, but for most people the path to that distinction is a lot of work, and in most places that lot of work gets done outside of the hours when people are distracting you with meetings and small talk, which means not stopping at 40 hours weekly.
Having said all that, I don't discount the possibility of work life balance in the 60-80 hour range, but that's a whole separate skillset.
FAANG's currently have a problem with ideological mono-culture. I dont know if recruitment has exactly suffered because of that, $$$$$ can allow for a lot of suppression of personal beliefs, but I do know a few people that have outright refused to work in those companies because of that, who are pretty excellent programmers
Using open source to judge quality seems wild. Maybe people just have no interest in maintaining an open source project. Looking from the outside at some of the stuff people put up with, it doesn't look worth it at all. I'll just work privately
As the sole maintainer of a popular open source NASA project (and contributor to several others), I can say that my open source work reflects very poorly on my work overall. We have a real problem in that there is a drive to open source things, but there is no money at all to support open sourced work. As soon as the open sourced work is no longer something I use day to day, I have to either maintain it on my personal time or it gets abandoned.
NIST and other government institutes are not known for open source work mainly because most of their work is a combination of science and technology communication. They deal in publications, conferences, and reference datasets. In my industry, NIST and the NIH produce the most important R&D reference datasets in the world, and everyone else looks to them for guidance. With that said, the NIH also occasionally produces world class software too (NCBI BLAST, etc.) although they do have some issues with parts of their software engineering culture being a bit out of date.
That question is pretty meaningless unless you can somehow measure the quality of an engineer. Is it the engineer who can build systems nobody else can, the one who can build the cheapest system that performs to spec, the one that can work well in a team, the one that is always available, the one that can teach others, etc etc etc etc. I'm sure anyone can think of many more aspects to being a good engineer.
I bet NASA and NIST have a great bunch of quality all-round engineers, but I'd be surprised if they were better at leetcode than the average FAANG dev. After all, FAANG devs have literally been filtered through an "are they good at leetcode" process. FAANG may be full of money chasers, but if the way to get more money there is by "being a good engineer" that does not mean much.
Indeed, but "works to the highest (quality) standards" is only one of the many aspects of being a good engineer. For example: government engineers are often not as good at completing projects within budget.
As someone who was a government worker, a lot of the issues why projects go over budget is because management believes that a single developer can do the workload of 4. So the product never gets delivered and that developer leaves to work somewhere else.
I think that dilutes the meaning of "quality" to nothing. Like if someone says "that's quality work" or a "quality engineer" I think of something specific.
For example I'd call a BMW a quality car. I wouldn't call a Lada a quality car, though it's much cheaper and has a much higher bang-to-buck ratio than a BMW.
In that sense sometimes government work has to be the highest quality, especially when it concerns security or safety. Sure it could end up being magnitudes more expensive but I'd say that's a question of efficiency not quality
Feds have some of the most useless engineers/bureaucrats in the world. They do have a very, very tiny amount of mission motivated folks who are the best of the best, but that number is a rounding error. Ask anyone who has left.
Not firing folks, low pay, focus on the best work life balance in history, heavy affirmative action, politics, and having to work hard to carry the coasters isn’t an environment that naturally attracts skill and competence. Work 500% harder than the next guy and get the same promotion. No thanks.
The gov and contractors, like it or not, are jobs programs first and foremost. A remarkably effective jobs program if you just measure folks employed and not output.
> Well, Google's and Facebooks definition of 'the best' engineers. don't think grinding leetcode for an interview is the best indicator of a good engineer
Their employees are also the subset of those who can get to a location where they have offices and have the relevant work permits. Those who do not object to and specifically want to work at those companies. Those who find their technical challenges of interest. Those who do not already have a satisfactory job elsewhere and are actually in the market for a job.
Some engineers see themselves as merely tools, so they "sharpen" themselves to be used effectively. Why would MAANAM (FAANG is a bit outdated) want more creative ones? They will get bored and leave.
So for each company they represent at most around (edit: 0.3%) of all professional devs, and presumably the "overhired for anti-competitive reasons" portion is a small fraction of that.
I think it's plausible that the superfluous hirings are caused by hirings of key individuals.
It's quite common for these big tech companies to poach each others department heads and other key personnel. This can cause significant damage to a company so can be an attractive tactic. The downside is that in order to retain these people you don't just have to pay them a lot of money you also have to give them big projects and resources to implement them, i.e. lot's of people get hired. This can a problem when these projects aren't supported by the wider company but are just someones pet project.
(edit: parent originally claimed 30 million professional software engineers worldwide, then edited in a revised estimate of 13 million.. which is in the ballpark of my original figure? :)
Just because someone "uses JavaScript" doesn't mean they are a full-time professional. In fact, most are dabblers. The number depends entirely on the definition - are all IT professionals considered software engineers? If so, that's about 24 million.
I am talking about full-time SWEs.
In any case Drew, it sounds like we're mostly in agreement. What a relief! :D
We can't really know what is in the minds of Zuckerfk and Pikaichu, in the end it's all speculation.
Yeah sorry I did edit the number down based on digging deeper into my link based on your comment here. I still think 0.3% (my original number was 0.1%) is not a meaningful amount of engineers for a "starve the rest of the world of talent" strategy to work.
I'm glad you put it that way. It's not necessarily smarts or talent, but it does take a particular willingness for the institutional peculiarities to integrate with a big organization. I'm not one of those people, I tried it, and I will never do it again. I did note, you either had people who had just joined, or people who had been there for nearly a decade or more. I think of the word "institutionalized", as in, they had bought into the institution lifestyle, and were so full of it's arbitrary knowledge that moving on would be like starting over.
Everyone claims they hire the best. For a long time Google and others had atypical hiring practices which they have since abandoned. I suspect this is because they discovered the techniques were less effective than originally thought. So 'best' by what measure?
I've worked with some really good engineers who came out of google, but I've worked with far more that were extremely arrogant but could not actually get anything done. One of the startups I worked at got an "advisor" from Google (as part of a startup program) that probably set us further back than it ever helped. Anytime this guy didn't understand something he just got extremely belligerent instead of actually trying to get the problem. In general his advice was ignored because it didn't make sense, and he never delivered on any of the promises he made. Not to harp on this guy, as he's just one example amongst many, but it's reached the point where if I see google engineers on the founding team for a company I typically won't consider working for them.
It turns out that being able to solve cute little puzzles while interviewing doesn't really help with systems level thinking.
> Anytime this guy didn't understand something he just got extremely belligerent
I think he might have backed himself into a corner by coming in as an 'advisor'. How can I be an advisor if I look like i don't even understand what is going. That must have been his mindset. So the only escape is being arrogant and belligerent.
Arguably that's Google and Meta's strategy (maybe even Apple) but that's certainly not Amazon's. They just mass hire anyone without a care in the world. Not sure if Oracle even belongs in this group.
I believe it. Every single day I get emails from various Amazon recruiters. Often it's for positions I'm barely qualified for. As much as I think AWS is a great service, I'd be terrified to learn what lies beneath given how low their recruitment standards are.
Their recruiting reach is high, but it doesn't have that much to do with desperation, and their actual interviewing standards aren't low.
The recruiting reach is high because every single sub-group of teams within amazon has their own recruiters, and none of them communicate with recruiters from outside of that. Sometimes i get multiple emails from different AWS sub-group recruiters per day, but it isn't because AWS is desperate for me. It is simply because for them, the existence of the other ones reaching out at the same time is completely immaterial, just like if they were recruiters from other companies.
And while yes, Amazon's interviewing bar might not be as high as Meta/Google/Dropbox/etc, it isn't far behind at all, and it is pretty much on par with Microsoft.
Disclosure: never worked at Amazon, but interviewed with them and the rest of the companies mentioned, and worked at (or got offers from) some of them.
Based on what I've seen from the outside about their corporate culture, I'm not in any way interested in working for Amazon/AWS.
That said, the interactions I've had with the people working on AWS have been uniformly positive. They're easy to work with and obviously very skilled engineers.
It definitely wasn't Meta's strategy when I was there in 2018. They hired a lot of junior software engineers but all other positions had relatively limited headcount (which I mostly think is a good thing).
Where are you getting the 10M number? Just curious not a criticism. I was thinking it would be around half of that(5M), with a tenth decent enough to work at most tech companies 500k, I think the bay area has 1M tech workers so half of them are engineers and thats one of the largest cluster of engineering on the planet.
A few years ago I was super curious how we stack up numbers-wise compared to doctors and attorneys (quantities artificially limited in the USA because of licenses). I did the research to calculate based on the number of CS graduates being produced by universities, combining it with average number of years worked before retiring. Unfortunately I don't have the references handy at this point.
Interesting, thanks for that info, licenses in other industries artificially reduce participants in that field. I wonder if tech not having them has resulted in our field to dominate most of the economy in the past 5-10 years(or helped that domination).
Hording talent could be a leftover Chesterton's fence from when they had an illegal agreement between Google, Apple, etc. to not recruit each other's employees, but Facebook was never found to be part of that.
100,000 of the best out of a pool of approximately 10 million professional software engineers worldwide is a sizeable portion. Additionally, not all 10 million are even close to being up to BigCorps peculiar standards (perhaps the standard is "someone competent enough that they could potentially build a competing product line").
Goog, Meta, Amazon, Apple, Oracle.. all have huge rosters and it adds up to a significant portion of the market. There is also a huge amount of medium sized whales such as SAP and PayPal who in total end up also consuming a lot of the talent.