8 ms·

The metal economy is huge. Almost every product requires metal. But metallurgists don't take home a big fraction of it. What's special about the software econo

• whateveracct 6y ago
  What would cause it to not last outside of precedent? In all those cases, it sounds like middlemen exist that corner the end-market that you gotta go through to make big money. How would that even work with software in the age of the internet, open source, and commodity cloud infra?
  • save_ferris 6y ago
    AI-driven software development down the road could theoretically commoditize developer skills. I sometimes wonder how many decisions I make per day that an AI could replicate in some repeatable way. Who knows, but nothing lasts forever.
    • runawaybottle 6y ago
      We should technically be the last to be automated. We’ll destroy other labor systematically over decades, and then finally ourselves. We seem to be subconsciously precocious about our fate too, as we deprecate things at a faster and faster rate.
      • sgillen 6y ago
        I think you might be surprised, it’s possible some types of programming could be automated away. Honestly I would consider writing programs more straightforward to automate than a lot of systems that need to interact with the physics map world, or with people.
        • ehsanu1 6y ago
          But programming is essentially about interacting with people more so than interacting with computers, even now. It's less so that the job will go away than that its nature will change, such that the proportion of time interacting with people goes up, and the time currently spent "coding" will instead be spent interacting with AI which writes the actual code. But this AI won't be smart enough for a while to talk to a normal human who doesn't know how to talk to it and translate requirements into the right form. Hence the need for the programmer. Programmers will make use of the automation to just become much more productive.
      • save_ferris 6y ago
        > We should technically be the last to be automated. I disagree, developer work will be automated as soon as it is technologically feasible, especially given the high cost of software development labor. I foresee AI software development platforms starting to take over new application development at some point, building applications that humans have minimal access to the source of. Existing apps by humans will likely need human maintenance until AI is able to reason about larger legacy applications. We’re already starting to see no-code apps taking off, and AI will drive the proliferation of that technology as it becomes more sophisticated.
    • d-man 6y ago
      I have doubts about AI driven end results, but I could totally see AI created domain specific languages
  • tlb 6y ago
    I don't know. Maybe this time is different. I have this book from 1880 about steam locomotives, where the preface talks about how all the most ambitious people are going into steam engineering. Which they were, for a while. The answer might be, until something better comes along.
• lainga 6y ago
  Maybe that analogy would be accurate if you could make new and pleasing music out of hundreds of small samples of other music, and also 90% of the musicians were used to either getting their samples via torrent or picking one album and playing it in an orchestra of 5,000 for the next 40 years
  • golergka 6y ago
    That's literally how most of music production works. You torrent sample libraries, get used to a couple of genres that you do well, and then use the same tools all over again for every client.
• hnick 6y ago
  Good software is a force multiplier than enables everything else to be more effective. A traditional mindset applied to software results in the businesses we often see people complaining about. "We make the product but Sales gets all the money!" and that's very similar to your metallurgists. Code for hire as a product of labour. It takes an army of people to sell, to advertise, to make me want what it is you have (just like with music). But a business who looks at software as complementary will see the value. A dollar spent brings many dollars back, just indirectly. It enhances the machinery that's already in place, and allows it to do entirely new things.
  • Craighead 6y ago
    tlbs comment is still completely relevant
• sidlls 6y ago
  " people who write the software take home a large fraction of it. Maybe more than half, " Where do you get that figure? The vast majority of people writing software are not paid enough salary + benefits to come anywhere close to 50% of the revenue their work is responsible for earning.
  • mikekchar 6y ago
    It's a good point. It varies widely depending on the industry, but even in the software industry a good rule of thumb is that IT costs about 10% of revenue (some places are as high as 30%, some as low as 2%). But that includes all the management (not executive management) and support staff. The spend for sales and marketing is much higher than development (though sometimes most of that goes out of the company, paying for sales channels, etc).
    • TheSpiceIsLife 6y ago
      > IT costs about 10% of revenue (some places are as high as 30%, some as low as 2%). That’s the total expense, right? Not just employee remuneration.
      • mikekchar 6y ago
        Yes. However, it may or may not include equity depending on how things are set up.
        • TheSpiceIsLife 6y ago
          Do you mean employee equity? Isn't that usually considered part of total remuneration?
          • sidlls 6y ago
            It’s not a cost as a function of revenue the same way cash, perks, and other benefit are. It’s also not clear that typical grants are enough to close the gap to “more than half” as the op remarked.
• MattGaiser 6y ago
  Metal economy is huge, but there are inputs. Software engineers have no inputs into our projects besides our work and some hardware and electricity. A developer is 99% of the cost of a piece of software.
  • ComodoHacker 6y ago
    You've missed coffee!
  • underdeserver 6y ago
    In other words, your value-added as a software engineer is significantly higher, as a fraction of the money in a transaction, than the metalworker's. This is because of leverage. Implement software that speeds up a process by 20% once, and every time that process is used from then on the speedup adds value. We're lucky to be working in a field that enables us to add a lot of value, for relatively little capital costs, and pocket a good share of the difference.
  • dwaltrip 6y ago
    They listed music as a counter-example to this. No inputs = zero marginal cost.
• tree3 6y ago
  > What's special about the software economy is that people who write the software take home a large fraction of it I don't think this is true. Take all the non-tech companies that rely on software and hire tons of engineers, such as insurance and banking.
• 6gvONxR4sf7o 6y ago
  My hypothesis for why tech has high salaries is stock compensation and fast growth. For example, Google stock is nearly 2x what it was 3-4 years ago. If you have a four year grant and had about half your comp in stock, then 2xing the stock means you're making 50% more now, assuming you didn't get any promotions or anything. If you're trying to hire folks away from these situations, your pre-stock-growth starting offer has to compete with their post-stock-growth comp, so starting offers have pressure to grow over time.
• pif 6y ago
  Musicians write music. Developers write and maintain software. Imagine there's no Version, it's easy if you try... Really, no concept of version: you buy a piece of software as it is, take it or leave it! Either it works enough for you or it doesn't. No bugfix in the next release because there is NO next release. The musician, ehm... developer, has switched to the new shiny thing. You modified yoour environment slightly and you need a tiny change on the software to adapt? Too bad it does not exist! Really, the point I'm trying to make is that comparing software developers to musicians doesn't hold a lot of water. Both professions include creativity, but a song shipped is as immutable as a piece of hardware. Adaptabillity is what makes software as worth as it is!
  • jayjader 6y ago
    One could argue the comparison still holds for developers of video games - as long as the game's code is not touched upon release. Used to be the norm, nowadays admittedly seems to be disappearing.
    • risyachka 6y ago
      Also the amount of skilled professionals needed to write a song and to write a game is slightly put - different. A song surely requires a skilled professional. But a game required dozens of those, with years of experience and lots fo time. Creative industries get way lower pay because the skill level needed is low compared to many others, like law, medicine etc. If you are talented you can paint great pictures or be a great guitar player in 1-2 years. To be a good doctor you need what? like 10? 20 years?
      • Kye 6y ago
        Like games and software, the staff needed to make a song varies a lot. If you want your [game, software, song, ...] to reach a certain level of polish and fame, you need a similar number of people working on it. Mixing, mastering, recording, composing. You need someone who's very good with sheet music and knowledgeable about all parts of an orchestra if you have any orchestral parts you want played on real instruments. And so on. Can I make a song in a week that'll impress a few tens of thousands people? Sure. People write in to me all the time saying "this is just what I needed!" They don't notice I banged it out in Ableton in an hour because it's what they needed. It's what I needed too, so it was easy to make it good enough. Just like you can make a game or a piece of software in a week that'll scratch a lot of people's itches even if it's not perfect. Recently, I got back into photography, and it's the same. I can take a great picture on my own and impress a lot of people, but most famous photographers have a staff that rivals the biggest non-unicorn SV startups. Editors, lighters, second (or third) shooters, general A/V crew, marketing, office staff, etc.
• lol636363 6y ago
  And just because in other industries people are used to taking smaller portions of revenue they generated, doesn't mean that we should do same. Also unlike in some creative industries, software engineers tend to have higher education and are better able to understand their economic contributions.
• risyachka 6y ago
  First of all, software is not a creative industry. Creatives always took home a small fraction for many obvious reasons. Software solves hard problems it is constantly evolving and updating so you need lots of personnel to maintain it, with a lot of knowledge so it will never be cheap, and great engineers to create it in the first place. Take away a bunch of artists nothing will change. Take away a bunch of critical software - the world will stop. So no, it will last and it will last long, as the world becomes more and more technological. It is unprecedented because software and computer world in general is a huge breakthrough and is only 30 years old. Music on the other hand is thousands of years.
  • TheSpiceIsLife 6y ago
    > Take away a bunch of artists nothing will change. . Take away a bunch of critical software - the world will stop. Is that right? Is there any evidence that a society can exist without art? There’s an overwhelming body of evidence that suggests it can do so without software.
    • colinmorelli 6y ago
      Not agreeing with GP, but I think your comparison here is a bit flawed. Art is a broad concept - it doesn't suggest any specific form or delivery. Software is fairly concrete - it has to be run on a computer. A more appropriate comparison would be art versus scientific research, or art versus engineering. I think that's a more interesting question, to which I would say: perhaps we could survive without either, but I can't imagine we'd thrive without both.
• phamilton 6y ago
  > Musicians don't take home half of music revenue. In most creative industries, the creatives take home less than 1/10th. Perhaps the difference is that musicians don't change a completed work. There is no maintenance cost. A song doesn't become useless and obsolete if you don't pay the musician to continually improve it. For static software, I'd expect it to look more similar. How much of Tetris went to the people who wrote the software?
• momokoko 6y ago
  Most of the people writing software for the companies listed in this article are freelance workers in lower wage companies. They make close to, and sometimes less than minimum wage for high wage nations. Startups pay well because speed is so important and faang pays well because their margins are enormous and the risk of those shrinking is worth paying top dollar for the absolute best talent in the world.
• davee5 6y ago
  Yes. The commodification of professional work is worth pointing out. I'm surprised by the percentage of people in this thread who think the SW industry is exceptional. It tells me a lot about the fictions people want to tell themselves in an industry that is still relatively novel. By contrast I'm a HW builder (mechanical engineer / industrial designer / design manager). I make good comp but maybe 70% of what my similarly leveled colleagues in SW do. The vast majority of the arguments in here still apply to my work, with a little less focus on maintenance. And yet my job role's comp is in a slow decline. Why? Because it's fully commoditized. Anecdote: I used to work for a HW focused VC leading their internal incubator/consultancy's physical design team. One of our ML focused investments got acquired for a ton of money and every other startup in the building wanted to know how well compensated the team was going to be. It's odd to see some people at the water cooler suddenly be worth millions when your startup is still grinding, it makes for stark contrast and a very real "what if" to the onlookers. One of the Sr MEs at a struggling startup was clearly jealous and lamented to me, "man, screw machine learning, how come nobody is paying me a million bucks for my work?" To which I said "let me ask you something, when you were studying mech-e in school, did you ever use steam tables?" yeah. "Did you learn about the Carnot cycle?" yeah. "Have you considered the fact that our industry is 200 years old? It stands to reason that our jobs are not hot shit anymore. Be glad you like your work and are still paid well; ignore the new kids if you can."
• dredmorbius 6y ago
  Curious what fields return what fractions of benefits / value / profits to what participants, and how you'd measure and classify those. There's SIC / NAICS / ISIC codes, or tier classifications, ranging from three (Qesnay) to five (Clark (1940), Hatt and Foote (1953), and Bell (1973)): [A]n economy's major sectors, as delineated by Clark (1940), Hatt and Foote (1953), and Bell (1973), correspond to major stages in the essential life process. The primary sector -- agriculture, fishing, lumber, mining, oil and gas -- represents the extraction of matter from the environment to produce energy, including the calories to sustain individual organisms. The secondary sector -- processing primary goods, as in construction and manufacturing -- represents the synthesis of matter and energy into more organized forms (negentropy). The tertiary sector, including transportation and utilities, represents the infrastructure for distributing matter and energy about the system, while the quaternary sector -- trade, finance, insurance, and real estate -- constitutes a parallel infrastructure for the collection, processing, and distribution of information that is necessary in all living system for the control of material flows. Finally, the "highest" of all sectors in its remove from the physical environment is the quinary sector, including government, law, and education, representing the societal programming -- socialization, education, law making -- and collective or representative decision making to effect control. -- James Beniger, *The Control Revolution,8 p. 179. For participants, perhaps: unskilled labour, skilled labour, professionals, managers, finance, consumers?