Skip navigation
This isn't alien technology, literally or metaphorically.
0:00 / 0:40

This isn't alien technology, literally or metaphorically.

Devon Eriksen

@Devon_Eriksen_139K followers

Visit on X
1,714 likesDownloadOriginal post

126,941 views

This isn't alien technology, literally or metaphorically. The breakthrough here isn't in engineering at all — it's in corporate organization. See, most machines are broken because the organizations that make them are broken. As an engineer, I would estimate that at least 30% of the parts of any given device exist only to correct for the design flaws in the other 70%. Works like this. Suppose I, an engineer, design a set of pipes and injectors to feed fuel to a combustion chamber at a consistent and controllable rate. Except in testing, there's some corner cases it doesn't handle well. Now, if you're an engineer, you understand this is normal. Nobody ever gets a design right the first time, unless it's so trivial that it's probably been done before. Your first design doesn't survive the wind tunnel, your first code doesn't compile right away, and if does, it segfaults. Your rockets blow up. And there's a flow problem with your fuel feed lines. It happens. Doesn't mean you're stupid. Means you didn't have enough information. But you have an even bigger problem. Middle management. Middle management is a special variety of hazmat suit, which is worn by the finance, sales, or market guys who run companies, so they don't have to touch the icky engineers. C-suite guys hate hate hate engineers, because it's a terrifying sensation to be dependent on someone you cannot understand, who doesn't appear to respect you much. (It does not occur to them that it is also extremely frustrating to have your work paid for, and thus controlled, by someone who cannot understand what you do, who doesn't appear to respect you much.) The primary role of middle management is talk to engineers so C-suite guys won't have to, and the primary qualification is to be a member of the right social class, and to hate engineers. This qualification is dressed up in secret handshake buzzwords like "management experience", as used in the sentence "I know you have been an engineer for twenty years and the other engineers all come to you for advice and leadership, but you don't have management experience, so I am going to hire my golfing buddy's 23 year old kid, who just graduated business school." So the fact that the fuel lines lines are not working quite right isn't your problem. That's part of the normal engineering process. Your problem is that your manager hates you and everything you stand for, and doesn't trust you or take your word for anything. He think his job is to keep you in line rather than help or empower you. So when you say "the fuel feed lines need to be redesigned", he says "we already spent six months and seventeen million dollars design the fuel feed system, we can't let you do it again." He thinks you are telling him seventeen million dollars worth of work needs to be thrown away and redone, and he cannot be told otherwise, because he cannot be told anything. So he says "you have two weeks and fifty thousand dollars to fix it". So you add another turbopump. What else can you do in two weeks? Now there's a new part. And if something goes wrong with that, another new part will be added to fix it. Because the company's actual priority isn't what you are responsible for — the design. It's what middle management is responsible for — the schedule and the budget. You have to play the villain so they can play the hero. What middle management doesn't understand, is paid to not understand, is what engineering actually is. Engineering is the process of making mistakes until you run out of mistakes to make. So when you spend six months and seventeen million dollars, and ended up with a design that doesn't quite work right in all cases, you weren't throwing away money, you were burning through mistakes. You've gotten a lot of them out of the way and won't make them again. But you have to get rid of some more before the design is actually right and doesn't require extra parts. The work of engineering isn't making the thing. It's teaching yourself to make the thing. Once you've done that, you can make the thing with minimal effort and time, because you know how to make the thing. C-suite financiers would hate this idea if they understood it. Why? Because it's unpredictable. Financiers like safe investments that make money. Not knowing how long something will take or how much it will cost is terrifying to them. So they train engineers to lie to them, by hiring a middle managers who try to force them to lie, and just interpret every guess as a promise if engineers still refuse to do so. But lying to yourself doesn't make the truth go away. The truth is that engineering projects take as long as they take, and cost as much as they cost. And that no one knows how long or how much they will be, because no one knows how many mistakes are waiting to be made, until they actually make them. The only thing you thing you can do about this is hire really good engineers, who catch more of their mistakes on the white board, leaving fewer to be caught in the wind tunnel, and this makes engineering faster... but it doesn't make engineering more predictable. Gantt charts are nothing but a collection of lies that corporations have taught themselves to tell themselves, lies that middle managers try to make true by enforcing them as promises. The ultimate reason that machines are 30% unnecessary parts is that the corporations that build them are 30% unnecessary people. This is what's different about Elon Musk. It's not that he's a better engineer. It's that he's a better manager. He understands engineers and engineering, and he doesn't hate them because he is one. He understood all along that green-field design is full of unknown-unknowns, and that there is absolutely no number-crunching, pie-chart, MBA magic that can eliminate this risk... it can only be hidden from view. The critical understanding is in the video clip below, where he says that SpaceX only had a 10% chance of success. You cannot say something like that unless you get it. And when you get it, you are free. Free from artificial anxiety about schedules and budgets. Anxiety about schedules and budgets is based on the delusion of control. Managers can't make a project finish "on time". They never could. The only power they have is the power to screw it up. The fate of a project is already written in the unknown unknowns before it ever starts. And the best, the absolute best, an engineering team can ever do, under any circumstances, is to confront those unknowns with clear-eyed honesty, and a willingness to adapt. Well, when you say to yourself, "This project has a 10% chance of succeeding", you've already confronted the pain of those admissions, which means you have already conquered the fear. The fear that makes you try to treat a prediction as a promise. The fear that makes you insert an extra turbopump, when what you really need to do is get busy redesigning the pipes, even though you have no idea how long it will take or much it will cost. Even when your rockets blow up. The SpaceX Raptor engine isn't just, or even primarily, a triumph of engineering smarts. It's a triumph of character. It's about an entire team with virtues MBAs lack: self-awareness, persistence, courage, humility, and, ultimately, hope. Because that's what it takes to pursue the best design, the RIGHT design, not even knowing if it exists to be found, much less whether you'll find it before you run out of money. Great things are not accomplished by middle managers with spreadsheets and Gantt charts. They are accomplished by teams of experts with passion and vision. Who are willing to risk failure so they can succeed. So what's the point in me saying all this? Am I just writing a puff piece on SpaceX and the Raptor engine? No. The point is that this is not special, one-off, magic alien technology. It's the systematic result of a correct understanding of engineering. Which means that EVERY COMPANY CAN BE LIKE THIS. If the people who control the purse strings are willing to learn from this example, and stop managing with spreadsheets and fear.

36 Comments

Comments from the original post on X (Twitter)

  • Christine Eriksen's profile picture

    @AnEriksenWife

    When people complain that your scifi dares mention SpaceX as the primary force behind space colonization, this is what they don't get. THIS is how humanity is conquering the stars!! Hope everyone gets the audiobook: Or the Kindle:

    38 likes

  • Devon Eriksen's profile picture

    @Devon_Eriksen_

    Yet no complains that I mentioned NASA, another real-world organization. Or Maersk Sealand, or DCM Hyundai. They just have MDS because rich guy didn't vote for their candidate.

    83 likes

  • NachoCat's profile picture

    @NachoCat16

    “Engineering is the process of making mistakes until you run out of mistakes to make.” Yes! Now that you’ve clearly articulated what were vague misgivings, I see the risk of SpaceX changing this exact engineering approach is what has me most worried about the company going public.

    22 likes

  • A Pragmatist's profile picture

    @pragmatist_a

    Re: "hire really good engineers, who catch more of their mistakes on the white board" This is a good one, and correlates with Musk's line about competence and caring. Too many execs really believe engineers are interchangeable. On the contrary, there is a huge range of real world engineering ability- especially as it relates to designing new products. A good laboratory for developing and selecting the best engineers, is to look at small companies and small projects, and find engineers whose designs have an extremely high probability of working the first time, because they have imagined every failure mode they can think of, and have incorporated tolerances and design features to eliminate that failure mode. You will never anticipate *all* the failure modes (well, maybe sometimes), but if you don't anticipate *enough* of them, then you have zero chance of being given enough time and resources to fix all the design deficiencies, before the project will be cancelled. Also, as Musk has pointed out, many projects fail at the requirements stage. In small companies, a single engineer can put themselves in the minds of the customer, and connect all the dots to come up with the best solution - effectively writing their own requirements - and the solution. I feel that the US has really put itself at a disadvantage by outsourcing so much of our manufacturing to China. It is not totally unfixable, but the thousands of small companies which no longer exist to develop engineers and allow them to prove themselves, really is a loss. This is also true of technicians and every other worker. In the old days, non-degreed technical people had paths to become managers, technicians, and engineers. In my experience, some of the most creative mechanical designers were non-degreed - including one person I worked with having over 35 patents. By the time I started my career, there was absolutely zero chance of getting an engineering position without a BS in engineering, and even *with* the degree, it took time to actually get an engineering position. At the time, only 50% of engineering grads were landing jobs at graduation. Also, it may seem paradoxical, but a tight labor market for engineers - meaning a shortage of engineers - can be a boon for developing hidden engineering talent. My father - a non-degreed engineer - told stories about the 1950's in the Midwest US, where jobs were so plentiful that a draftsman or engineer could quit on the spot, knowing he could get another job immediately- often for more money. This forced companies to hire based upon talent and ability: Hiring on credentials was a luxury they could not afford. Companies would also train the employees they needed - much more efficient than low-demand economies, where workers try to guess which credentials they can monetize, and wasting lots of time training for skills never to be used. Add in DEI screening of candidates for engineering, and you have real problems - but those problems can give opportunities for companies choosing not to sabotage themselves with DEI nonsense and scoop up talent. I understand that Musk thinks that robots spell the end of worker shortages. In the long term, I think he is right, but I think his timeline is far too short. With the rapid reindustrialization needed in the US, we will need more workers - which is a good thing for driving up wages and increaded productivity. Increases in productivity can keep inflation from being a problem. Looking back at the US industrial explosion during WWII, and the amazing designs developed quickly and successfully using tools much less efficient than those of today, I can only conclude that our failure to build an amazing and prosperous society is due to massive mismanagement. Increased regulation, corruption, and government bureaucracy is at the heart of this national mismanagement. And increased government and socialism is the problem - not the solution.

    20 likes

  • James Seymour 🇺🇸's profile picture

    @realJSeymour

    Two employers ago the boss approached me and said "Customer want us to design for them. Can work work up how much that'd cost and how long to deliver?" So I sat down and wrote a functional design spec based on what little info I had. Then I developed a quick-and-dirty product design spec (for myself) from that functional spec. (By now any engineers reading this are thinking "Oh Lord...") Then I broke that design spec down into modules and estimated the amount of time it'd take to design, code, test, and debug each of them. Then I used the magic software engineering project projection formula: "Take however long you think it'll take, double it, and add 20%." (IYKYK) (You would be amazed at how close I tended to be on actual completed projects using that formula.) I looked at the numbers and thought "He's never gonna believe it." So, fully aware I was shooting myself in the foot, I shaved off that last 20%. I took it to him. He looked at it. "There's no way the Customer is going to buy this," he said. "It is what it is, boss. I can knock whatever you want off to give you the numbers you need to hit, but I have to warn you I already knocked 20% off my real projection." Thankfully, he decided to no-quote the project.

    18 likes

  • Lord Business (Gen X Alpha Male)'s profile picture

    @hydromerchant

    Real salesmen are hated by Middle management for different reasons. But it burns just as brightly.

    17 likes

  • Philo of New York's profile picture

    @PhiloOfNYC

    The biggest, but never really acknowledged, reason why American manufacturing lost first to Japanese and Germans, and then Koreans, and then Chinese, is that managers wrestled control over large American corporations from engineers, with the results described in your post.

    14 likes

  • GOPSantaCruz's profile picture

    @GOPSantaCruz

    Even many engineers don't understand Conway's Law (TL;DR: "You ship your org chart"). The Dilbertine human elements don't have to be so confusing. They have patterns, anti-patterns, and "laws" too. That's why @ScottAdamsSays was so funny.

    13 likes

  • John Fisher's profile picture

    @JohnFisher9372

    Any technological company is balanced on a knife edge. If the engineers have too much fun, the company fails to make money and goes bankrupt quickly. If the engineers have too little fun, the best ones leave, the rest go Dilbert or Wally, and the company goes bankrupt slowly.

    12 likes

  • Petrucio's profile picture

    @PetrucioBR

    But I thought Neil D Tyson told us Space X isn't advancing on anything. 🤦

    10 likes

  • Georg Felis's profile picture

    @georgfelis

    It's like NASA loves 4-barrel carbs. Elon went for fuel injection. NASA says "Look at the complicated problems we fixed with complicated parts." Elon says "We just avoided those."

    8 likes

  • PepeLp72's profile picture

    @PepeLp72

    @AnEriksenWife really needs to be compiling these tweets (essays). Succinct and insightful.

    8 likes

  • Kevin Monk 🇬🇧 ⏩'s profile picture

    @KevinMonk

    With the death of @ScottAdamsSays, I have a vacancy for a sage in my life. You’ll do for the moment. Excellent post 👏 Again

    7 likes

  • Bernardo Armero's profile picture

    @b_armero

    That's a very deep insight about big engineering companies. When they are or grow enough to make financial contrls jump in, Engineering bcomes another boring operationl activity among many. And there are some psychological phenomena to be unearthed here that I'm willing to study.

    6 likes

  • JSBAnuther1's profile picture

    @JSB152sbcg5268

    Just give me 1970s appliances that last 30 years! I don't need them to be smart; I need them to last. If they break, I want a generic relay replacement, not a proprietary board that's no longer available. Heck, the only thing wrong with the board is probably a $0.05 voltage regulator, but who's going to troubleshoot and fix that for $100?

    6 likes

  • Bear Goes Long's profile picture

    @GoesBear

    And then the MBAs mandate we use SAFE and pretend that all engineering problems can be broken into predictable two week "sprints" that show them to magically forecast the precise timeline and cost of every project Oh sure, I can hit your mythical bullshit schedule and budget, just didn't expect it without bugs Fast, cheap, functional - pick two and be honest

    6 likes

  • ThanksForTheMemeries's profile picture

    @The_Memeries

    There's a couple of long form interviews with Musk on the everyday astronaut channel. Dunno if you've watched them. He delves into that management and team culture. One thing that stuck out was design decisions were tied to a person, a name, not a department.

    6 likes

  • Nigel Woods's profile picture

    @woods96718

    You can see this quite clearly, when you look into any (especially European) engine bay. Each new government regulation has led to an addition of parts to meet it. No time has been spent to reduce part count (cost!)

    5 likes

  • T Patrick's profile picture

    @tvpatrick

    I wish I could remember who I got this from so I could credit him, but - another way to visualize the relationship between engineers and the c-suite in most companies is that they are separated by a vast fog bank generated as follows: c-suite makes a promise to investors that the company will accomplish x in a given time period. They then turn to their direct reports and say “make it happen”. Their directs repeat this process down the line to the folks doing the actual work. All incentives are aligned around reporting progress towards goal - so decisions are taken to enable positive progress reports. Issues are increasingly minimized as communication moves up the chain - generating the fog bank. C-suite has no idea that sub-optimal decisions are being made or that the actual goal may be at risk. All the players at every level of management is optimizing for their own career advancement by reporting positive progress. C-suite gets what they asked for - until they don’t… By the time critical failures occur many individual managers have moved to other projects or organizations…

    4 likes

  • Philip Turmel's profile picture

    @PJTurmel

    "stop managing with spreadsheets and fear" I'm not holding my breath that companies will do this, broadly. Or even narrowly. I'm *so* happy to be a solo entrepreneur. (Shades of "The Mythical Man-Month". Similar constraints.)

    4 likes