What Silicon Valley “gets” about software engineers
blog.pragmaticengineer.comIt’s kind of a miracle engineers are treated as well as we are compared to most traditional office jobs. IMO there’s probably a good reason for it, namely some institutional memory of when engineers would just defect en-masse and literally destroy your company if you pissed then off enough. But it also makes me a little sad that other white collar jobs don’t have the same benefits.
There are probably a ton of jobs where people work just as hard as engineers but are treated as little interchangeable boxes by the money people.
> It’s kind of a miracle engineers are treated as well as we are compared to most traditional office jobs.
Other office jobs are treated better. They don’t get pinged at random times outside of business hours. They don’t have to attend late night meetings because of foreign time zone
I know a bunch of AWS people and have heard worse things from them
Frankly, I love programming but would rather have a regular office job. This industry is predatory
I think you're glossing over how much people in other office jobs get gaslighted (gaslit?) into showing up at odd times too. To finish a report. To coordinate with an offshore accounting team in a foreign time zone (not unique to Engineering). To be always reading their email. Not for nothing are some states and countries writing rules against being "always on" into their laws, recently. So, sure, a small contingent of Engineering folks might be on-call, but you should be able to make good working agreements about that, put boundaries around it, and keep your sanity. If not, you're maybe being gaslighted like any other old regular office worker.
I replied to your other similar post too, but why not look into working for smaller companies or contract/freelance work? Not all tech jobs are big tech type work. You’ll make less money perhaps but it sounds like you might be happier based on what’s important to you.
Smaller companies are less stressful but they often give you no autonomy or purpose.
It can be a good economic / stress tradeoff but it will suck your soul.
Left job with a unicorn because of stress, I make more than double that by contracting with smaller businesses and working less - but I'm way more depressed about the job.
I think a happy medium might be to run your small product company. I do that on the side as well, but it brings 1/10th of what I can make contracting.
I know tons of office jobs that have to attend late night meetings because of foreign time zones.
Just like I know lots of developers that don't.
What a weird generalisation to make.
Not a miracle. Things fuck up extremely quickly when you the systems are not maintained and it takes competence. A lot of businesses value (most perhaps?) goes to zero in a flash sans - IT. There is a developer behind all of that.
That combination of needing talent which is in short supply and being essential.
Even comparing traditional sysadmin work — where operations are even more crucial — to software engineering, it’s easy to see that engineers tend to get paid more, get treated better and have more freedom.
Obviously there are reasons for this, but it seems to boil down to broad generalizations in how the C-Suite sees your role, and very little in the particularities of the org or employee.
The reason is growth. A sysadmin will make sure the lights stay on and the place doesn’t burn down, but they aren’t creating a new revenue stream. And when all these companies need to grow forever, they need to be creating new things, which developers do and sysadmins don’t.
And that is why platform/devops is safer than the sysadmin. Not only do they help developers create new things, but make it complicated enough that you can't disband their team ... {tongue in cheek, sort of...}
It's not a miracle.
You're treated well because you have the potential to deliver multiples of what you're making in revenue.
Other office jobs rarely can do so.
That means that sensible places will happily pay you a lot and make money on you, and that will prop the entire market up, which means even crap places will have to pay more just to get some developers.
Supply and demand and the fact that their multi-billion dollar business depends on this small handful of people?
Name another white collar job that’s simultaneously as complex and as creative as being a programmer.
Architects build houses.
Software engineers build whole cities, and they exist only in their minds! (i.e they can’t even see them, they’re just abstract concepts of enormous complexity)
I'm very thankful to my high school self for choosing computer science. I feel lucky I happened to be interested in computers. If I was born 5 years earlier, I may have made a different choice.
> institutional memory of when engineers would just defect en-masse and literally destroy your company if you pissed then off enough.
Got some good examples?
Probably a reference to the "Traitorous Eight". You can Google more about them.
One of the realities of capitalism is that people are not paid proportionally to how hard they work.
Or in proportion to how much value they create.
They're paid in proportion to how much unique value they gatekeep.
Hence why Silicon Valley's masters were so keen on everybody learning to code in primary school and getting more women into the profession. They dream of making us more of a commodity, so that as the chief gatekeepers of the tech economic engine, they get to keep a larger slice of the pie we create.
>Hence why Silicon Valley's masters were so keen on everybody learning to code in primary school and getting more women into the profession.
Why is it bad that more people learn to code? I see this as a net benefit to humanity. Why do you want to gate-keep it?
Put another way, wouldn't it have made more sense, social-benefit-wise, to encourage everyone to be nurses, doctors, teachers, or members of another profession that provides direct benefits to humanity? Or flautist or waitstaff or any thing else that benefits people too?
It really was to put the prices down and benefit themselves, not to benefit humanity overall. They would never encourage people to learn mental health providing, physical therapy, exercises for their own health, they don't actually care about those.
A glut of coders only is good for the coders if there is enough overall business to keep them actually employed somehow. Otherwise you have a lot of people who don't spend time coding.
I was just explaining why the tech oligarchy is so obsessed with it. It's a long term pragmatic and selfish effort to drive down our wages that is couched in progressive terms about inclusion, equality, education, etc. They also want to appear to be the hero while they try to push down dev wages.
It was a positive statement. You're free to make whatever normative judgements about that that you please. I'm not here to stop you praising it.
'Tech oligarchy', whatever that means, can't forcefully turn everyone expert SW devs even if they wanted to.
We studies fairly advanced CS at school, without the 'tech oligarchy' making us (in Europe where this is non-existent anyway) and yet very few of us turned it into a career later in life because not everyone likes to do programing for a living and there's other lucrative jobs out there. Same at university, less than half of those who entered, managed to graduate in CS because it was tougher than people initially thought.
Cook, Zuckerberg, Musk and Bezos can't turn you into an productive CS if you don't have what it takes, they can't download CS theory into your brain like in 'The Matrix'. In fact, many of them are responsible for dumbing down the population when it comes to computers, so I guess we can actually "thank" them.
Only the last ~6-10 years saw an insanely large number of people get into the industry without CS background because of the super high demand driven by the mobile revolution, negative interest rates, crypto hype cycles and stuck-at-home pandemic, and not by the 'tech oligarchy'.
But that was potentially an once in a lifetime event which might never come back. The mobile market is now saturated, crypto is done, people are not stuck at home anymore and the negative rates are also over, plus a trade war and an actual war, meaning an industry slow down, but again, not dictated by the 'tech oligarchs' and more people choosing CS careers, but by market, geopolitics and financial circumstances.
It is so disingenuous to snip out the reason they gave and then pose a sanctimonious question as if it wasn't clear. They said it's to increase supply and lower labour costs.
You can argue with that or not but don't pretend you do not understand the capitalist ideal of an excess supply of labour so that salaries and benefits can be driven to legal minimums with desperate people still waiting to be hired no matter the indignity. Creating Labour surpluses to reduce Labour power and enable exploitation is not a "net benefit to humanity", it only benefits capital.
In our shitty world, you are only valued if you are in a position of more demand than the supply. You might get that by being lucky but most only get it by gatekeeping whether it's doctors or delivery drivers or companies buying up their competitors. If you stand idly by, the world will punish you.
>It is so disingenuous to snip out the reason they gave and then pose a sanctimonious question as if it wasn't clear. They said it's to increase supply and lower labour costs.
Oh please, don't act so 'righteous-than-thou'. With this logic, aren't we all here who got into SW development responsible for increasing the labor supply and decreasing wages?
What's disingenuous, is people in traffic complaining there's traffic. It's sounds like you got "yours" and then you wanna kick down the ladder behind you.
Where does the buck stop? Who gets to decide who's deserving of getting into SW development without also being responsible for diluting wages, and who not?
Those are valid questions, you are describing the problem. It is a tension. It applies to immigration and any number of other social issues. What was wrong was to frame the flooding the market with new software developers as an entirely noble intent. It's like saying the abuse of h1b visas has also been noble and not about labour costs / control over staff.
The problem is that it is a fundamentally unfair game which requires some people to lose. If you want a solution then it's to remove desperation from the labour market. Remove the implicit threat of destitution and ruin if you don't take the work that is offered to you at the minimum cost an employer can get away with. This is what takes people to ideas of basic income and basic services.
>Oh please, don't act so 'righteous-than-thou'.
Theres an interesting pattern here of us making positive statements and you inferring normative judgements which then make you upset.
That's good, in fact. People should be compensated for the value they generate. Why should a fast food worker be compensated equivalently to a doctor or engineer? I do not want to live in a world where that could exist.
CEO compensation vs returns:
> We find evidence that industry and size adjusted CEO pay is negatively related to future shareholder wealth changes for periods up to five years after sorting on pay. For example, firms that pay their CEOs in the top ten percent of pay earn negative abnormal returns over the next five years of approximately -13%. The effect is stronger for CEOs who receive higher incentive pay relative to their peers. Our results are consistent with high-pay induced CEO overconfidence and investor overreaction towards firms with high paid CEOs.
https://www.wsj.com/public/resources/documents/CEOperformanc...
Employee compensation vs Output:
https://www.bls.gov/opub/btn/volume-6/pdf/understanding-the-...
They're not paid for the value they generate either. They're paid proportional to the leverage they have to extract money from the business.
In other words, can you run the business into the ground? If yes => $$$$$$$$$$$$$$, if no => $.
The CEO can tank a company (in fact they frequently do)
An engineer certainly has more power than average to sink a company (isn't half the star wars story pretty much this?)
Cleaning personnel does not
A manager at an In-N-Out burger restaurant makes more than most developers outside of Silicon Valley ($160000) - and tbh I don’t complain about that because I’ve seen the speed those guys put out burgers in Mountain View on a busy day!
You are in a world where that can exist if you make international comparisons. There are some fast food workers earning more than some developers.
Bartending at the right bar is a six figure, semi-untaxed salary which is competitive with a jr swe salary.
Not competitive because you still have to account for quality of life, working on site with no remote option, working evenings, nights, and weekends, general riskiness of being around drunk/high people, etc.
Of course you may not have had to study as hard, so your quality of life as a young person might have been better, but bartending is not secure income for life either.
Either way, the easiest way to look at it is, do parents tell their kids to aspire to become bartenders?
> Either way, the easiest way to look at it is, do parents tell their kids to aspire to become bartenders?
That sounds incredibly elitist and cringey. If I had a child earning six figures as a bartender I’d be very impressed, and even if they weren’t earning six figures but were enjoying life I’d be happy… And taking it from another perspective, I know people who’s parents would be disappointed their children are ‘lowly software engineers’ too. With that sort of attitude you can rarely win.
Using a broad range of parents’ recommendations for what their kids could aim for does not having anything to do with parents being proud or disappointed.
It is a rough rubric to gauge an average quality of life for a person who does a certain thing for a living, not that parents or even entire generations of parents are always right about the continued resilience of a given occupations’ quality of life.
Okay tiger mom. Bartenders don't get fired because they got a cheaper office in Estonia or India to write code. They don't have 9am meetings. Git is an insult you call somebody; What's a JIRA‽ You don't want to work in the nightlife/entertainment industry, which is fine, but it's a $36 billion dollar industry and some people want to do that instead of sitting in front of a screen all day long.
At the end of the day, take home pay is one metric we can actually compare, because different people want different things.
You can always rely on Gergely to cut to the heart of the matter with his articles. What he's written largely chimes with the (many) diverse engineering environments I've collected a paycheck from. Some, usually the worst ones, have that "just stick to shipping PRs code monkey" mentality based off zero wiggle room in terms of defining the problem scope. At these kinds of companies and teams, the seeping lethargy and disillusion has otherwise talented folks building out code they know has limited impact on the business.
The best places, in fact my current role in an SV-adjacent tech company, encourages a kind of consulting approach to help define and solve business problems, not just JIRA tickets. This results in much better engagement with engineering overall, and getting better things built quickly.
> The best places, in fact my current role in an SV-adjacent tech company
Would love to find a company like this again, past two roles were JIRA micromanagement shops and actively discouraged talking about anything.
This is the reason why I am trying to move away from being a software engineer. I don't like to complain but it is honestly degrading how SE are treated in some places. I've been on teams with 4 project managers/analysts and 3 developers. They will expect you to do all the real hard work while they sit in meetings all day making UI mockups and architecture diagrams. There is just no way in winning this.
> They will expect you to do all the real hard work while they sit in meetings all day making UI mockups and architecture diagrams
You could also look at this the other way.
They are the ones doing the real hard work i.e. making tough decisions, dealing with stakeholders, incomplete and poorly conceived requirements whilst you just get to build what's written down on a page.
Everyone plays their role in companies and assume they are the indispensable ones.
I think the point of the article was that to be valuable as an engineer you have to do things like "making tough decisions, dealing with stakeholders, incomplete and poorly conceived requirements". My problem is that a lot of people have your attitude and pretend that they are doing me a favour by having this being written down all the while limiting the actual design space for a problem, and in the same breath patronise me on what teamwork is really like.
If you want to do that work, have you looked for ways you could get involved?
Or if it’s the case that you’re actually doing that work behind the scenes because the output from the PMs isn’t very useful, have you looked for ways to make your contribution more visible?
Oh, how I laughed.
Mostly they just pass those things on to the developers to deal with. Then we have to figure out how to bypass them and communicate with the stakeholders ourselves to figure out what the real requirements are and what we can build which solves their problem and yet is still broadly recognisable as the specified deliverable.
After that, coding it is easy.
Yes PM roles seem to be exploding at certain companies and it's a bad thing for developers. There's not enough engineers to meet demand so just hire more PMs seems to be the strategy.
What is degrading about that? Would you rather sit in meetings than develop?
I would like to work in a place where communication is a two way thing.
It's a culture issue. Some PMs get the impression that they're higher up in the hierarchy; they start delegating instead of negotiating with the devs. Really they're just around because upper management prefers talking to them instead of the pseudo autistic engineers.
Of the three places I’ve been a full time employed dev I only enjoyed working at the first.
This was a UK retailer, the other two were “tech” companies.
The retailer operated in the way the article describes, direct contact with the business, ability to unearth a problem and run with it, understanding that dev work isn’t just fingers on keyboard and sharing of full business context.
The other two had strict comms lines, not allowed to stray from the Jira board, rarely any contact with the business or users and no sharing of business context because “we don’t want to distract you” - completely missing the point that without the contact you rarely understand what the point of the work is.
I've been in the first type of work my whole academic career. I'm definitely more of an informatician/sysadmin than a scientist, and despite being entrenched in the science for almost a decade I can tell you that I never developed a 'feel' for it - it was too specialised, and I lacked the ground knowledge to bridge the gap. A smarter man would have found the time, but in my defense even in a single field has too many subfields to delve into, and you simply cannot master them all.
The best projects, and when I'm most efficient at work, is when someone explains the input and describes the desired output in as most abstract way possible.
Well, science is more specialized than business, for example there is only one "MBA" but at least a hundred masters in science.
I think, though I'm not sure, that some people prefer the latter environment.
Having worked with teams which have preferred to work in as pure a technology environment as could be constructed while still providing value, I think this might be the "product focussed" (or not) distinction that people make about themselves on their CVs.
To me “product focussed” is the situation the article describes, everyone thinks about the product (and therefore the business by extension)
Those who don’t like that work environment (so would prefer the latter two of my three dev jobs) would be tech focussed in my book.
It’s all semantics though
Yeah, that was the conclusion I reached as well. I might not have made it clear
Like everything from pragmatic engineer so far, that is an insightful article. As someone with a more traditional, read half mechanical, engineering background who is working for quite sometime now at the interface between mechanical, electrical (hardware basically) engineering, commercial and software (because no modern piece of machinery works without that), I'd like to add some things.
At the core of the difference between "traditional" and SV type companies is a fundamental difference on how they view software developers (and I am firmly in the non-SV camp myself): are they engineers or not? In SV, yes. For almost everyone else definitely not. And there lies the problem traditional, engineering heavy companies have integrating software.
that does cut both ways so, most of the SV-software engineering practices are simply as incompatible with hardware engineering as hardware engineering practices are with software development. And for most non-SV-pure-software companies, any piece of software is just another part number in the final product configuration, in that environment, software developers are nothing special. And it takes a special kind of software developer and company culture to properly manage that.
That being said, I agree with the articles conclusion. And I'd add that hardware engineering could benefit a lot from treating their engineers the way SV does their software devs. I do have a question so, where does all that significant horsepower of SV companies go nowadays? From the outside, it seems to go into ad tech, social media (with the aim to sell more ads), crypto (nothing to show for besides bitcoin and ethernum as new stuff to speculate with when gold isn't fun enough anymore) and "AI" (the latest hype which still has to show some real world benefits).
The true innovation to come needs a combination of software and hardware, each bit for itself got as far they could. There isn't much world left software can eat without some allies, at least until the next cycle repeats.
At (traditional) this is causing me to develop feature and then throw it again and then implement it again slightly differently because someone said something in the chain without thinking and sometimes to fight with management teams for doing something insane. And in one case why throwed out 3,5 months of development work, because some exec was mistaken.
But at (SV startup company) this caused at least for my team complete chaos, what should we do, who should be doing the integration, talking to customers, why are we doing this.
Usually at SV startup company I was like 2-4 times more productive, but often output of my work wasn't really significant, as in traditional company so small dumb feature was raking in money and really improved large customer base.
This article was somewhat revelatory to me in that I can now see that the source of my frustration with the direction my career has gone in recent years is that I spent much of it cutting my teeth at a SV-style company during the 2010s before it was bought out by traditionalists, and that subsequent companies ended up being more solidly traditional, even if they initially appeared SV-like on the surface. I think I had naively assumed that SV-style was the norm, and the experience of discovering this to not be the case has been a little alienating.
There is only one central difference of "Silicon Valley" type companies with "the Rest of the World". It is the recognition (or not) that we are in an accelerated information processing age where the wiring of information flows in society is upgraded dramatically. While this realization is not new (from Negroponte's Being Digital [1] to the perennial VC cry "software will eat the world") the disruption story is for at least two decades broadcast to deafening volumes. But it is by no means universally adopted.
For most organizations "IT" is just a utility. While information flow and processing is vital for most businesses, the control of these particular "means of production" is not deemed central to the business model. Like the provision of electricity, water etc, it is a stylized something that happens in the basement or outsourced. The C-Suite is clueless of digital tech and will at best include some tech consultant types rotating in and out.
In fact the very strange period of tech we are going through is a testament to how the alien forces of this new landscape have turned everything upside down. Google, Amazon, Meta are not "tech" companies. They are advertisers, retailers, publishers that have made information flow and processing central to their business model (and for now at least, have cornered the market).
What is mildly interesting though is that there are many major sectors where information processing is 100% what they do (finance and insurance the easiest example). A universe where the information market is cornered by, e.g. banks is actually less absurd than one cornered by advertisers.
With the above framing, the role of developers in decision making etc. is in reality just an epiphenomenon and circumstantial. Form follows function and it is the functional difference (what the company does) that is the driving force in organizational matters, remuneration etc. This also gives you some hints as to when things might change: When senior industry personalities are equally comfortable both in digital technology and the details of the particular sector they are in.
I don’t think your analogy would not be true for Tesla.
My underlying theory is that software is how we map what is happening in the real world and represent a model of it in computer(s).
Great Software engineers have this amazing ability to represent this and write computations that can reason on top of it for effective decisions.
Whether it’s dashboards, social network, extracting patterns from data, advanced electric cars with autonomy, mobile phones etc.
If you can use the power of computers to do more of what is happening in the human brains to create economic value, then that is a highly leveraged engineer.
The last startup I worked for as a UI dev the marketing manager wanted me to make a purely cosmetic change to the UI that at the time was quite difficult to do and half jokingly I told her if she insisted, I would quit. She immediately relayed this to the CEO, her husband, sitting across the table "Did you hear what he said?" I was not fired on the spot, but quit shortly afterwards anyway.
I think autonomy is ultimately a red herring -- what people want, crave, to feel dignified is agency. This comes in various forms but sometimes it can be as a part of a collaborative team that values their input, not just as a free-floating agent that can sweat hard and steamroll things into production.
Full title: What Silicon Valley "Gets" about Software Engineers that Traditional Companies Do Not
It sounds to me like successful software companies are Agile in nature. Most significantly, that autonomy and people in the business are the priority, and that the processes support them - not the other way round.
Imo all of this talk about being “well treated” in SE is an utter load of shit
Regular office jobs don’t have oncall. They don’t get randomly pinged outside of business hours. They aren’t as rat racey. There’s less ageism. They don’t force you to live in areas where a shitbox sfh costs more than 2mil
Seriously, I wish I could go back to undergrad and tell myself to go into a regular engineering field. And my experience isn’t even the worst I’ve heard, just ask anyone from AWS and they’ll tell you
A lot of tech jobs don’t have on call/out of hours expectations (in the UK at least). They’re probably not the top paying jobs at the prestigious companies, but that’s the trade off. I’m sure there are similar trade offs with the “best” jobs in other fields.
Don’t work at big tech. If possible, don’t work with people buying into Protestant work culture.