1,813 karma · joined June 15, 2022
Underrated point. It didn't really sink in for me until I saw the numbers with my own eyes.
From a quick DDG search of publicly available info, here's [1] the numbers for FAANG headcount as of the end of 2022:
- Meta: 86,000
- Apple: 164,000
- Amazon: 1,541,000
- Netflix: 12,800
- Google: 190,000
The numbers get a bit smaller if you focus only on creative roles (engineering, design, etc) — but it's still an enormous amount of people. And all of them are constantly moving at the speed of realtime chat to jump on every project and figure out how they can use it to advance their careers. The politics and bureaucracy are practically inevitable at this scale.[1]: https://www.macrotrends.net/stocks/charts/META/meta-platform... (all companies' numbers slightly rounded for easy reading)
Remember those photos of thousands of multicolored bicycles abandoned in fields? Or the scooters being yeeted into the ocean, global riots from taxi drivers, the collapse of the taxi medallion market, regulatory debates in every city/state/country, billions of VC money raised in weeks (or days), the Darwinian M&A scene as companies were devoured in the jungle as quickly as they were founded...
...and, of course, the fact that you could finally push a button and make a bag of groceries appear. (Though the 3,600,000 millisecond latency still isn't great on that one.)
"Driving school" wasn't a standardized thing back when I got my license — in many places, you can just stroll in, pass the driving test, and be on your way. So I guess I missed all the interesting tips they might share in a formal education environment...
Any other learnings from driving school that you find yourself using often?
That way, it's unavoidably in front of everyone's face, and you get the perk of ironclad timestamps in the document-editing history.
I used to know this grumpy old guy who had a simple explanation for that statistic:
It doesn't mean that, if you start a business, you have a 99% chance of failure. It means that, out of everyone who starts a business, 99% of founders are guaranteed to fail, and 1% of founders are guaranteed to succeed.
The trick, I think, is to become the kind of person who can succeed at starting a business — and then start a business.
How would you define it?
At the same time, the "candidate asking questions" portion is 10 minutes if you're lucky, and it's between two strangers who've known each other for maybe 45 minutes up to that point. It's understandable why the interviewer might give canned answers.
I think it's still worth asking the questions. The Joel Test [1] may be older than many candidates these days, but a large number of companies still fail most questions.
For the trickier questions, I think you have to read between the lines and look at what's not said. If you ask about work-life balance and the interviewer gives a little "hah" before answering — well, there's your answer. If you ask about the build system or deploy process and they keep hand-waving — well, they're still telling you something very important about the engineering culture.
[1]: https://www.joelonsoftware.com/2000/08/09/the-joel-test-12-s...
I've been thinking about this too.
My plan is to just sit down and really put the hours in — work hard, and really grind through all those meditation apps and podcasts and books.
And I think, if I just grit my teeth and keep at it, nonstop, then eventually I'll master the secret to relaxing and taking things slowly.
I think it works fine if we add on the concept that a good manager operates on a different timescale. The on-boarding timeline for a first-line EM can be months at some companies.
In your example, the manager moving over should be given a compensation package based on the expected value he's going to get out of the flailing team once he works his magic. (And, indeed, this is exactly the kind of package often given to externally-hired CEOs brought in for turnarounds.)
> Otherwise you're mooching off other people's work. [...] Show me what you did.
But that's just it. The manager's skill is improved utilization of resources. The leader's skill is improved direction of resource utilization. By definition, those skills are expressed on the canvas of the team they lead.
Or, more simply: A leader without followers isn't actually a leader at all!
And, indeed, we see that in the real world too: A successful executive in one company goes elsewhere, doesn't fit in with the new company's culture, and performs terribly. It's a tale as old as time.
You wouldn't run an A/B test to quantify your impact, but you could pretty easily do before/after metrics (or just collect them on the way during a graduated rollout) for any feature you're rolling out. It should be straightforward to quantify productivity benefit for a large initiative.
> How much of that revenue is attributable to the manager's superior skill
The output of a manager is the output of their team. We can't A/B test managers (or at least, we probably shouldn't), but we can look at a cohort of managers and see how well their teams are generating business impact. Combined with other data (e.g., upward feedback surveys, level-skip 1on1s, retention metrics, etc.), we can measure the impact of a manager.
I'd take that deal in a heartbeat. I think I (and most other decent engineers) can probably point to tens (or hundreds) of millions of dollars in savings generated over the course of a single career.
As far as I can tell, it gets frittered away by empire-builders who take that $ and spin up new teams for nonsense. Kicking back a % to the engineers who harvested all the savings seems like a much better model.
The point is not about whether a task ends up being less than an hour. That happens all the time, as you say.
The point is that, when beginning the task, it is entirely unknowable whether it's definitely going to fit within an hour or not. Any of the 3 examples you gave — updating a test, fixing some automation, or fixing a bug — are things that can easily blossom into all-day (or all-week) tasks.
And so, with 60 minutes and 3 tasks that might be doable within the time, it becomes that much more difficult to justify investing the time to get into flow exactly because of that unknown ROI. Plus, you need N minutes to get into flow and M minutes on the other side to prep for the meeting, so you only have 60 - N - M minutes to invest anyway.
There's a guarantee of an interruption to flow within the hour, but there's no guarantee the probably-small-but-possibly-enormous task will be done by the time that interruption arrives.
I know this is my entitled Big Tech worker bias, but... surely there should be a lunch break?
I frittered away large swaths of my 20s, and now, as I try to make up the lost time in my 30s, I often think of an entirely different quote though:
Software development, like professional sports, has a way of making thirty-year-old men feel decrepit.
- Neal StephensonThat's still not bad, though. Graduate at 22, having banked a decent bit of savings already from 2 or 3 Big Tech internships. Make L5 by 26, coast for 5-10 years, and then call it a wrap. Invest in real estate as soon as humanly possible during that time, and scale out the real estate portfolio at the same time as the brokerage account and the 401k.
If I were to do it again, that'd surely be the fastest way to go.
Then again, if I were to actually do it all again, I'd love to do it all entirely differently. Maybe become a carpenter or something.
Maybe even a fifth: storage price.
Funny enough, I actually have noticed this with my books. More books require more bookshelves. More bookshelves require more real estate. It's a straight-line arrow of causation from picking up a copy of Moby Dick to house-shopping for a place large enough where I can have my own private library.
- 'j' and 'k' to move up and down, with '{' or '}' or even something like '20j' to move quickly
- 'w' and '$' and '0' to move horizontally, with things like '5w' to move quickly
I also rely heavily on 'v' and ctrl+'v' to select arbitrary shapes of text for yanking/deleting. I'm not sure what the exact takeaway is from comparing our styles, but it's interesting to realize about how much of coding comes down to just navigating up/down/left/right.
I agree with your last couple paragraphs almost verbatim though! That StackOverflow answer is exactly the reason I decided to try vim and stuck with it.
The verb/noun nature of vim makes it so easy to enter a flow state. The text editor fades into muscle memory, and you can surf waves of code with no bottleneck between mind and machine. (To write this comment, I actually had to open a random file and navigate around it while watching my hands — it's been years since I've thought about it!)
At some point, the business world called, and I stopped spending time on IRC. And I haven't kept up, but I think something happened with Freenode, and the communities faded away, right?
Those were simpler times, I suppose. We were working with all the new frameworks of the day, and it was the grizzled old engineers alongside the young hotshots — and none of it mattered, because all anyone saw was your nick and your code. You'd put some interesting snippet up on one of those anonymous paste sites, and the critique would roll in.
I don't know how kids these days learn about dependency injection, the downsides of singleton architectures, the art of API design, or the nuanced tradeoffs of using an ORM to abstract over a relational model. But I rather enjoyed learning it all "on the streets" of IRC — frankly a far more welcoming and engineering-focused atmosphere than many tech companies.
IMHO sharding traffic by subreddit doesn't pass the smell-test, though. Different subreddits have very different traffic patterns, so the system would likely end up with hotspots continuously popping up, and it'd probably be toilsome to constantly babysit hot shards and rebalance. (Consider some of the more academic subreddits vs. some of the more meme-driven subreddits — and then consider what happens when e.g. a particular subreddit takes off, or goes cold.)
Sharding on a dimension that has a more random/uniform distribution is usually the way to go. Off the top of my head (still speculating wildly and basically doing a 5-minute version of the system-design question just for fun), I'd be curious to shard by a hash of the post ID, or something like that. The trick is always to have a hashing algorithm that's stable when it's time to grow the number of shards (otherwise you're kicking off a whole re-balancing every time), and of course I'm too lazy to sort that out in this comment. I vaguely remember the Instagram team had a really cool sharding approach that they blogged about in this vein. (This would've been pre-acquisition, so ancient history by Silicon Valley standards.)
As for subreddit metadata (public, private, whatever), I'd really expect all of that to be in a global cache at this point. It's read-often/write-rarely data, and close-to-realtime cache-invalidation when it does change is a straightforward and solved problem.
Do you mind sharing the job title for something like this? Sounds like a dream job.
<vent>
And, leaping into action, carrying the banner of innovation, they're declaring their bold vision of the future: getting sneezed on 20 times a day while attending the exact same meetings as before.
Plus, many people get an extra bonus in this future: the addition of 1-3 hours of commute time per day. (Thanks to terrible traffic in every single part of the Bay Area, this is true for many more people than you might imagine.) Time entirely wasted, that neither their family nor their business will benefit from.
The saddest part of this stupidity is that all teams still have to collaborate globally anyway. So everyone's still going to spend time dialing into meetings, fumbling with microphones, and running into A/V issues — except now it'll be compounded by the fact that there'll never be a meeting room available, and all the equipment will be covered in a fine layer of snot from everyone sneezing on everything.
The fundamental issue with the "RTO or not" question is not the debates about the science of productivity. It's simply the fact that any business choosing to spend time on this is already displaying a failure of judgment and lack of focus.
Entire markets are getting overturned by the arrival of AI, and these people are navel-gazing about desk occupancy rates.
</vent>
Ah, the classic "California Corporate" style of speaking:
Q: Yes or no question about a specific case?
A: Four sentences of condescending and convoluted explanation about how the
rules work for the entire system, while simultaneously refusing to give a
clear "yes" or "no" to the specific case being asked about.
It seems passive-aggressive to demand that every single reader piece together the logic of their "labs" and "campaigns" and "badges" to figure out if a given course is free or not. I'm all for shipping early and often, but it does feel like adding 2 if statements would have made the whole pricing answer about 100x more user-friendly.</vent>
According to Jony Ive, Steve Jobs was exceptional at nurturing this kind of creative environment. He describes it perfectly in this WSJ article [1] from a couple years ago:
> As thoughts grew into ideas, however tentative, however fragile, he recognized that this was hallowed ground. He had such a deep understanding and reverence for the creative process. He understood creating should be afforded rare respect—not only when the ideas were good or the circumstances convenient.
> Ideas are fragile. If they were resolved, they would not be ideas, they would be products. It takes determined effort not to be consumed by the problems of a new idea. Problems are easy to articulate and understand, and they take the oxygen. Steve focused on the actual ideas, however partial and unlikely.
The other part of it is that ideas, as fragile as they are, can blossom and transform if you can get multiple people and multiple different perspectives involved in the early stages. That's why it's so powerful to have this kind of culture on a team; it lifts everyone up.
The problem with polishing the idea into a final draft before sharing it is that — whether or not it's good — it's almost guaranteed to not be as good as it could have been. And, by the time the "final draft" is ready, it's too late to get involved — all the interesting rabbit holes and tangents and possibilities have already been pruned away.
[1]: https://www.wsj.com/articles/jony-ive-steve-jobs-memories-10...
It's not, though. Many people have no clue how to interview well, and way too many tech interviewers are obsessed with whether or not the candidate got the "right" answer.
Anyone who dropped multiple mediocre solutions before a good one in an interview with me would likely get a strong hire — I love to see this kind of iterative thinking, and finding people who can role model a healthy exchange of rough-draft ideas is always a great boon for the team's psychological safety (and by extension, their creativity and productivity).
But I think, industry-wide, it's not great. As a SWAG, probably ~50% of interviewers are more interested in the correctness of the answer rather than the caliber of the thought process and communication.
I was just thinking earlier today about the absolute meaninglessness of my Big Tech job — and whether it's even feasible or worthwhile to try and find something better.
I have meetings, and many of those meetings are about other meetings. In fleeting moments between calendar chimes, video calls, chat pings, and the flood of email, I code little morsels of SQL or Blub to at least do something interesting. It's tolerable, like all studies in nihilism. But you've got me wondering what it'd be like to work on something I'm genuinely enthusiastic about.
I really dislike the notion of taking a reduction in income to get that, though. I can't pay the bills with "meaning."
Having components with lifecycle methods called at known times by the framework is a perfectly straightforward mental model. It's worked just fine since the beginning of time. iOS still uses it.
I remember investing hours (and days and weeks) migrating old code to the "new way" or trying to get some simple thing "working with hooks." Other than being able to share in the authors' feeling of smugness over the unnecessarily-convoluted mental model when we finally got it all working, I genuinely can't recall any value we got out of the whole investment.
If you don't mind me asking, how'd you end up in the space? It feels like the vertical has so much esoteric/specific knowledge. We ran into a few companies that were founded by truckers (or by people who had truckers in the family), and it seemed like already having all this knowledge gave them a big leg up.