Silicon Valley salary

=cities =sociology =economics

 

 

Scott Alexander: Every new housing unit prevented saves one person from having to live in San Francisco.

economists: People want to move to San Francisco, and it's good to give people what they want.

Scott Alexander: San Francisco is a bad place to live; those people are moving there for high-paying jobs.

economists: Then that must be a worthwhile tradeoff. If they're getting paid more, then they must be more productive, so moving more people to Silicon Valley would increase productivity.

me: Are you sure?

Scott Alexander: Maybe that's just a coordination problem, and a different place would be better.

economists: There are already several other software-focused cities.

me: Hold on, why are we just assuming that higher pay in Silicon Valley means higher productivity?

economists (collectively sighing): Competition means companies act efficiently, or else they go bankrupt and get acquired. So if they're paying some salary, then that must be worthwhile.

me: What about humans? Do people have to act efficiently or else run out of food and fall into slavery?

economists (smug): Ah, things used to be like that, but not anymore, thanks to capitalism!

me: That's not what I meant, but nevermind. Anyway, companies don't act that efficiently in reality, so your model is flawed.

economists: No, actually they do, but mortals like you might not understand the reasons.

me: Why did Yahoo refuse to buy Google for $1 million, buy Tumblr for $1.1 billion, and reach a market cap of $40 billion before being sold to Verizon for $4.6 billion?

economists: Bah, statistical fluctuations. That kind of anecdotal evidence is meaningless.

me: OK. Why do managers try to spend their remaining budget at the end of a budget period? Is that rational for the corporation as a whole?

economists: We don't need to look at such minor details to make models, just like physicists didn't have to look at water molecules to understand that water is incompressible.

me: But water isn't incompressible; small changes in density are what causes water pressure. Are you telling me that you don't think it's worth studying what most of the population deals with for 40+ hours a week?

shadowy figure in back: I DO!

me: Whoa, I didn't see you there. Who are you?

shadowy figure: I'm an economist too! But the media spotlight is on those guys so I guess it's hard to see me.

stage crew: Security? We have a disruption in the public forum.

shadowy figure: Wait! I have some things to say! Mmph!

economists: Sorry about that. As for what you were saying about water, well, like with water molecules, the fine-grained variation averages out, and small things can be ignored like water density changes are.

me: Are you sure? Corporations all consist of humans, and most have very similar structure. That means systematic irrationality is very possible.

economists: That doesn't explain overpaying employees in Silicon Valley.

me: I think it does! Hey Jeff, why do managers try to spend all their budget?

Jeff Bezos: Because they want to justify receiving that much money, of course. Controlling a bigger budget means more personal power, so managers usually want to justify as big a budget as possible.

me: Any other ways middle managers justify a big budget?

Jeff Bezos: I assume you're thinking of headcount, and yes, managers typically like to maximize the number of people under them. But there are exceptions, like hedge funds, where you don't need more subordinates to justify more money going through you.

me: What if middle managers could choose to manage half the people at twice the salary?

Jeff Bezos: I think most would, because it's less work for equivalent institutional power. But higher salaries need to be justified to top management.

me: What if you pay programmers 5x as much and just say they're 10x as productive?

Hacker News user: Hey! The idea of 10x programmers is a recognition of individuality, that we're not interchangeable cogs!

me: I know you'd like to think that, but when companies say they want "10x programmers" what they mean is, "all our programmers are interchangeable units of a special type that's 10x as productive as the most common type". You should have known that arbitrarily dividing programmers into standardized and immutable "productivity" categories wasn't what you were hoping for.

Jeff Bezos: Companies would have teams in different locations competing with each other. If a team in Austin does better than the Silicon Valley team, they'd move more operations to Austin.

me: No, that's what you'd do. You have different teams for every AWS product, all competing with each other, but that's not what companies normally do.

Jeff Bezos: It's not that uncommon an approach.

me: It didn't work very well for Eddie Lampert at Sears, but that's true.

Jeff Bezos: Hmmm?

me: OK, maybe that worked better for Eddie than for Sears. Maybe he was optimizing for spinoffs to maximally screw over minority shareholders and junior debt owners. But I was getting to my main point: there are diminishing returns to more programmers on a project.

Jeff Bezos: Maybe that affects performance numbers, but diminishing returns doesn't change what's better.

me: What if the marginal value of adding another programmer to a project goes negative?

Jeff Bezos: Illogical. More workers is more better.

me: Why don't you hire 10 CEOs for Amazon, then?

Jeff Bezos: That's completely different. Unlike programming, being CEO involves making high-level decisions about what to do.

me: Are you sure you know anything about technology?

Jeff Bezos: My company is the leading provider of cloud services.

me: Your website is bloated and has bad recommendations, and AWS (which provides all your gross profit) is only profitable because it charges more than competitors. People only use AWS because it's considered a "safe" and "standard" choice that won't get middle managers in trouble. You're the new IBM, or maybe Oracle.

Jeff Bezos: That's very mean and hurtful. Let me wipe my tears with this $100 bill while I work on rockets and curing cancer.

me: You know, I could've helped with your rockets, but now I don't feel like it.

Jeff Bezos: Hah, if you spent 10 years in college for a PhD in engineering and worked 80 hours a week at SpaceX, maybe someone would give you the honor of using something you designed. But you, you only wish I cared about your half-baked ideas, you hackneyed dilettante.

me: Well, you're right that I don't care that much, and you're wrong because I don't care that much. But I was talking about adding programmers to a project having negative marginal value. Let me give you an example: the Facebook iOS app had >18000 classes years ago, and now it's grown to 500mb. It's bloated and has real performance problems, because it had a lot of programmers and bad management coordination. Facebook spent at least $100 million making its Facebook and Messenger apps, and they're both garbage. Messenger cost more to make than some video games with online chat as a minor side feature. Is that thanks to the "higher productivity" of programmers in Silicon Valley?

Jeff Bezos: That doesn't mean that results would have been any better somewhere else.

me: True. Maybe hiring in Silicon Valley can give better results if some middle manager maximizes their budget and headcount, and that gets fewer programmers in Silicon Valley than Denver, and hiring in Denver would mean too many programmers and managers.

economists: Competition prevents that kind of inefficiency.

me: Competition from what app? WhatsApp? Oh wait, that was acquired. That famously only had 55 employees when acquired, and some of those were new employees that hadn't helped yet — having more engineers tends to increase purchase prices for companies, you know. How about Instagram? Snapchat? Hmm, I'm noticing a pattern here. Any potential competitor gets bought, and a monopoly has more profit to buy with than a competitor has value from potential profit, because competition reduces profit. By the way, Craigslist only has 50 employees, and it seems to work well enough.

economists: No, according to our model, antitrust action would be bad for consumers.

me: Yeah? According to my model, you're wrong.

Jeff Bezos: Soon...

me: What?

Jeff Bezos: Mmm, nothing. We were talking about house prices, and how it was smart of me to hire in Seattle instead of San Francisco; let's get back on topic. Even if companies irrationally overpay employees there, if they have too many programmers anyway, then it doesn't matter if they replace 5 midwestern programmers with 1 programmer in San Francisco. Maybe they should theoretically do stock buybacks instead, but they won't, so that's irrelevant.

economists: Objection!

Jeff Bezos: Overruled. I'm running a successful company, therefore by your logic my economic reasoning is correct.

me: At least then you have 5 decent jobs in decent locations, instead of all that money going to some landowners that are already rich. Tucker Carlson is a smart guy, and he'd ban automated trucks to save trucker jobs; I guess my main problem with that is just that being a trucker sucks. It's boring yet stressful, bad for your health, and doesn't give good opportunities for pursuing hobbies. Einstein worked at a patent office — maybe patent review should be farmed out to NEETs; they probably wouldn't do any worse, from what I've seen. Or maybe Japan has the right idea with its tariffs and subsidies of small rural farms.

Jeff Bezos: I have no idea what you're trying to say here.

me: Yeah, I guess you wouldn't. OK then, suppose you start replacing suburbs with apartment towers in San Francisco. Have you seen the traffic on the freeways there? Adding more residents isn't going to help with that. There are negative externalities.

Jeff Bezos: And yet, people usually want to live in cities. What about the positive externalities?

me: Cities require some size to support a level of specialization, but there are diminishing returns to specialization. People aren't like intermodal shipping containers, but there are still advantages to standardizing roles, for education/communication/management/handoffs/etc, so there's some optimum level of specialization and cities only need to be big enough to support that. I think the required size for an optimum level of specialization is 1m to 10m people, depending on a city's exports.

Matthew Yglesias: Why not have a higher population density in apartment buildings instead of suburbs? That way people wouldn't have to drive so far.

me: People like trees and space...

Matthew Yglesias: Not me! I only like trains! I wanna live in a hive!

me: OK, maybe some people don't. But criminals are poor, and expensive suburbs are hard to get to from cheap areas, so crime is low. And parents would rather send their kids to schools with kids of rich parents, because there are fewer disruptive kids, and everyone would rather network with rich kids than poor kids.

Steve Sailer: And of course, every form of exclusion except house prices is illegal. Surely this singular focus on money will make society more equal!

me: Tell you what Steve, I'll take some CK2 and KSP players and SHAFT fans, and you can have the local MENSA members. Fair?

Steve Sailer: Whoa! Awesome! I'm headed there now.

Matthew Yglesias: OK, he's gone. You made some good points, but what if people pay $1 million for a 500 square foot apartment in NYC instead of a 2000 square foot suburban house in Australia? Sure, that might seem worse in every way, but then we could have trains and subways!

me: You might want to make sure that your cities can actually build trains and subways first. But let me ask you something Matt: you like density, and getting housing built is hard, right?

Matthew Yglesias: ...yes?

me: So then you'd say that empty houses are bad, right?

Matthew Yglesias: I guess so, but that's implausible.

economists: It's illogical!

me: Have you been to Vancouver or Seattle? There are a lot of empty houses owned by wealthy Chinese. There are real estate agents that speak mandarin and ignore locals. In London, half the sites with planning permission to build homes are owned by a company that doesn't get any homes built.

Matthew Yglesias: Maybe empty buildings still count for density, because they can be filled in the future...? They still have the aesthetics of density...?

me: People want to live in cities because of services they buy and jobs that need them plus people with complementary specialties. So, the positive externalities of "density" come from the kind of people companies like to hire — and that's mostly smart young adult non-students. That being the case, empty houses are a negative externality, retired boomers preventing young people from moving in are a negative externality, and high density is only valuable when there are a lot of smart young adult non-students around.

economists: But statistics say density is good for GDP.

me: Yeah, well, maybe you don't actually know how to control for confounding variables.

Matthew Yglesias: What if I value density endogenously as justification for having become the bugman I am today?

me: I'm glad you're outvoted by the plebeians you look down on, then. Your only slightly-above-average intelligence is vastly outweighed by your biases from ingroup signalling. I said "ingroup" so I'm ✓ingroup.