Why are so many successful founding CEOs software engineers?
iism.org
iism.org
I mean, come at it from the other direction: Why are there so many software engineers who are not successful CEOs? My God, they have the secrets of for loops, S-expressions, they know about vim---and yet they aren't running Pfizer? Why not?
If you have already a CS degree and you spend a couple months refreshing your memory, you would be amazed what codebases you can touch.
Although NL might be a place where there isn't a lot of programming market, I'd be willing to bet the resume needs a bit of work.
Maybe they turn into the engineer who can extract water, if only for reasons of path-dependency. And in a world where water-extraction is a valuable industry, that path-dependent contingency will result in them maybe starting a valuable water-extraction engineering firm. But it's not that nailing up drywall as a kid made them uniquely qualified. It's that they were industrious and smart and had sufficient fortune and whatever else. Which is good, all that stuff is good.
The problem with nails is that they walk out over time. Secondarily, they can become "squeaky" when used in floors.
Because programming is too much fun.
Because there is no c-suite.stackexchange.com for them to cut-and-paste their corporate directives from.
There are 10 CEO level people mentioned.
Woz could certainly get a job at Apple. He's clearly brilliant and capable.
Larry & Sergey could definitely get jobs at Google. They were PhD CS students at Stanford (how on Earth is that "unqualified" to you?).
Zuckerberg could get a job at Facebook. He obviously wrote a lot of Facebook's early code.
Gates & Allen allegedly didn't do much actual coding themselves but they certainly understood it. I imagine Jobs, Bezos and Jack Ma also understood coding even if they didn't do any themselves. I don't know anything about Ma Huateng.
I doubt any of them could drop in to a high level coding role and be productive, but they could all do intern or junior level stuff. Code isn't that hard that you need to be a special snowflake to do it.
I mean, plenty of very capable people are filtered out of Google and go on to do great things. It is bound to happen when you receive hundreds of applications and hire one person.
I don't know if they wouldn't get hired, I mean it's not like those companies never hire straight out of college, they obviously do, so expertise is not an absolute requirement. But they probably couldn't be hired as drop-in experienced SE's. Half their tech stacks are on technology that didn't even exist when they last touched code.
All of these CEO's are famous for their excellent early talent acquisition of software engineers. Microsoft and Apple for leveraging those engineers to initiate a new digital world, and Facebook and Google for attracting (and motivating) the worlds best to achieve global domination.
That's exactly what I said in my first post - "I doubt any of them could drop in to a high level coding role and be productive, but they could all do intern or junior level stuff."
I mean if we reanimated Big Al Einstein today and he had a blind interview for a tenure track physics position at Princeton Physics dept he would probably fail. The field has changed too much although with months of prep he'd be able to catch up, the same as L & S could for SWE interviews.
At the same time, I think the coding interviews generally are harder than what's needed to do the job well, just because there're so many job applicants so the bar gets unnecessarily high? And people who fail the interviews can still be really bright. (Also, some people get nervous and fail mostly related to that. I wonder how to avoid that happening)
I mean there are Yale JDs that fail the bar exam but no one would claim that the average Yale JD couldn't get hired as a Big Law associate or judicial clerk.
They would definitely get a phone screen based on their resumes, but there is a good chance they wouldn't even make it to the virtual onsite round. It's not like the interviewers would use their products or look at their code or dissertation or whatever.
The "average Yale JD" would not get hired as a judicial clerk. Competition for clerkships is fierce, and generally only the top students get them. Any judge would probably take a "top" Yale JD, but unless the judge is a Yale graduate they'll all pass on the "average Yale JD," in favor of a better-performing graduate from another law school (usually their own).
> Homebrew, which was used by 90% of Google's engineers at the time of his interview
This is definitely not true.
At any rate, the point is that a guy who used software used by a large portion of Google's engineering team because it was superior to the in-house solution could not get a job at Google because he failed to answer a programming question in the specific manner the interviewer wanted for a type of problem he would never actually encounter on the job. And in that vein, Sergie and Larry would similarly have difficulty getting hired at Google because other than their original algorithm, which Google has not used in over a decade, they have no modern or recent software experience and would likely fail all of the programming interviews.
Do you seriously believe that they couldn't come up with something better than mediocre package manager?
The real question is: knowing Google's track record of almost complete failure for the past few years, why would you believe in Google?
Google is run by the engineers. You don't get to blame the MBAs for its failures.
What are your sources?
It's hard to put everything in a build system like blaze but if you do, I think the outcome is superior to Homebrew.
Google is a linux shop, and you were generally not allowed to even keep the source code in your notebook.
> But ultimately, should Google have hired me? Yes, absolutely yes. I am often a dick, I am often difficult
If any of the interviewers got a whiff of that, it could have easily torpedoed the interview loop.
His post also makes him sound like a dick (which he admits). Google has a “no jerks” policy that a lot of people take seriously.
Many stanford phds fail the interview process. I'm not speculating. Others have suggested this says more about the interview process than the person in question. They might be right. But here we are.
Hell, the type of attitudes and personality that made these folks good at what they did would probably be liabilities from some the interview process.
Zuckerberg, on the other hand, partly coded a CRUD PHP app before moving on to executive stuff. He would probably not pass a Facebook interview without several years of intense preparation.
Brin and Page fall in between those two. They at least have several years of academic coding and their early Google work.
Yes, they're all very smart, but that's not enough to get into most FANG software engineering jobs.
That doesn't mean he could do leetcode problems.
What? Gates & Allen similarly wrote "a lot of early code". Gates also acted as a Chief Software Architect in his final days at Microsoft.
[1] https://en.wikipedia.org/wiki/TRS-80_Model_100#ROM_firmware
Almost certainly true, but would they pass the coding interview?
I'm not clear why/if this is controversial. I thought it was well established that interviews are tough from both ends and that they suffer heavily from Goodhart's law.
Have you seen Facebook's early code? It's very much "Early 2000s PHP spaghetti written by undergrad on a caffeine bender at 3 am". Facebook succeeded because of catching lightning in a bottle at the right place/time, and sure, later engineers built it to scale, but let's not kid ourselves that it was due to Zuck's coding genius.
It depends on what you mean by "serious". Software development is very much a free-for-all as far as backgrounds go. It attracts people from all disciplines. There are NO CREDENTIALS needed to be called a "software engineer", and it will remain this way for the foreseeable future. Every organization that develops software has its own rules and standards for how they staff their teams, usually these rules are flexible to almost non-existent. They hire based on whether they think someone can do the work based on some loose evaluation criteria.
But, yeah, a mid-to-late career-changer with a successful business background as highly paid executive is going to raise eyebrows if they go after some software development job-- not because of technical chops (assuming they still got it), but more "wtf are you thinking?" considerations. These people would normally just self-fund their own software projects, or partner with someone in exchange for having some fund with code (it happens).
CEO demographics vary wildly. The article just cherry-picks the top ten market cap companies, which today, happen to dominated by software companies. Most of the CEO's have been SE's? So what? That means almost nothing. Are these 10 companies supposed to be profoundly representative of the world of work in some way?
In the USA, maybe that's true.
In Canada, for instance, there are legal requirements to being able to call yourself an Engineer.
That's why "Software Developer" is used.
https://careers.microsoft.com/us/en/job/958535/Software-Engi...
Likely the company wants to use the same language in US as well as in Canada. Noone I know in Canada is called Software Engineer, we all refer to ourselves as Software Developer.
FWIW the UK has a similar thing for traditional Engineer roles. Nobody is legally permitted to call themselves a "tradtional" Engineer (such as Civil, Aerospace, Aeronautical, etc.) without, typically, at minimum a bachelor's degree or equivalent vocational training in that discipline. These are deemed "protected" roles but they are very specific - i.e. there needs to be some kind of auhtoritative body that wards the title(s). "Software Engineer" has no such protection. Generic/non-descript "Engineer" isn't protected. There are plenty of "Software Engineer" roles and it's entirely feasible to be hired as a SWE before you even get any higher education at all. Most of my previous roles have officially been documented as "(Junior/Senior/Lead/etc) Software Engineer" on my contract(s) of employment.
I do know some people who have the title software engineer and they had to do like a whole ceremony where they swore an oath and got a ring and stuff.
They also have real-world legal rights and obligations that it grants them. For instance they can put their stamp on a structural drawing. Because they studied that stuff.
It does seem silly to have real-world building architecture requirements for a Software Engineer but that's where we're at.
Anyways, Software Developer (or even "Product Developer") is a much more common term in Canada as a result. No one wants to tangle with provincial regulations over a job title.
I'm a bit baffled on the fact that they can stamp things they aren't formally trained in. That's a surprise.
I had assumed they had some baseline level of training in different areas because a Software Engineer I know stamped an architecture drawing for a shed he was building. Granted a shed isn't a high risk building but still it's surprising he can do that if he's untrained.
Also does the ring really have nothing to do with it? My understanding is you earn the ring at the same time as the ability to call yourself a professional engineer. Maybe the ring itself isn't necessary but it's symbolically related for sure. Am I wrong?
Professional Engineer is a protected title in Canada. Is that what you’re thinking of?
I'm mixing between EGBC (Engineers and Geoscientists British Columbia) and PEO (Professional Engineers Ontario) in links below, but feel free to contact your local engineering licensing body to check if your titles hold up. There is a bunch of weird language through out. Who knows, you may qualify.
From [1]: "In British Columbia, anyone who practises software engineering, or who uses the title “software engineer” (or a similar title that implies that they are a software engineer, like “firmware engineer”, “mobile app engineer”, etc.), must be registered with Engineers and Geoscientists BC."
There is a lot more in [1] including, "Not all software development constitutes software engineering." and "... many individuals who develop software probably do not actually engage in software engineering..."
Someone linked to a Microsoft job using the word engineer in Canada, which seems strange since Microsoft already learned about the protected use of the word "engineer" in Canada in 2001 [2].
Professional Engineers Ontario says, "On July 25, 2002 Microsoft Canada announced that they will continue to use the term 'engineer' as part of the Microsoft Certified Systems Engineer (MCSE) designation."[3]
> Professional Engineer is a protected title in Canada. Is that what you’re thinking of?
I think "engineer" is protected too. I think job titles that contain "engineer" can only be used by those who are registered professional engineers.
From [4], "Can someone call themselves an Engineer/Professional Engineer or P.Eng if they don’t have a licence?
The term Engineer/Professional Engineer/P.Eng. can only be used by those that have been granted a licence by PEO, under the authority of the Professional Engineers Act. The title “Engineer” is restricted to Ontario licence holders under s. 40(2)(a.1) of the Act."
[1] https://www.egbc.ca/Registration/Individual-Registrants/How-...
[2] https://www.canadianconsultingengineer.com/engineering/micro...
[3] https://www.peo.on.ca/public-protection/complaints-and-illeg...
[4] https://www.peo.on.ca/public-protection/complaints-and-illeg...
Interestingly, you might be wrong. There's a continuing push to get people in the industry who can code for startups wanting to disrupt the space. You'd probably be the lawyer who can code and any lawyer-adjacent tech company would love you.
When we think "programming", we think FAANG but in the real world, there are thousands of places that need basic code slung.
Not just "basic code". There are thousands of companies doing just as challenging work than FAANGs do. There's innovation happening all over the place.
There's probably a place nearby that needs a basic developer for writing
No, most of them don't, because most of them aren't themselves working at the apex.
The people working at the apex fall for the Apex Fallacy, including the ones that think they don't but think that the whole field tends to.
Consider two broad but otherwise generally viewed as unrelated (however, not mutually exclusive) skillsets: A and B. Lots of people are great at A and lots of people are great at B, but the intersection of the two is generally small (perhaps "law" and "programming" could be reasonably argued as examples of A and B here). There are large numbers of people employed doing A, and similarly for B, often quite successfully in both cases.
Now, there's some much smaller set of individuals who have a reasonable grasp of both skillsets A and B. Recall that they're not mutually exclusive (are any skillsets, really?), just not often seen together. It seems that this set of people would be uniquely positioned to potentially provide immense value over the perhaps small (but even more marginally serviced) field of needs where A and B intersect. The value here isn't in the individual's capability for A or B in isolation, but in what they can provide where A and B cross.
I realise the above is a big generalization and perhaps hugely oversimplified, but it's interesting to think about the different, seemingly unrelated skillsets where there may be great potential from some cross-pollination of ideas to bear new fruit. I do think that "programming" in general, perhaps in some form not as it's conventionally understood today, is becoming somewhat of a new literacy, at least to the extent that there seem to be countless existing businesses or fields of inquiry that could benefit from programmers who are also experts in their traditional domains. However, one could also come up with lots of different examples for A and B above, which may not even include programming at all.
I think in software (and CS in general) this advice is especially valuable because software and computing is the glue that can connect and materialize ideas into something that seems tangible. It's also something that's often approachable by an individual and doesn't always require a mountain of capital to achieve.
https://techcrunch.com/2020/07/19/the-dual-phd-problem-of-to...
[0] https://www.lesswrong.com/posts/XvN2QQpKTuEzgkZHY/being-the-...
No we do not. Some of us do
As a lawyer who can code (years of experience as a professional software developer and years of experience as a practicing attorney) let me assure you, this is not the case. There just isn’t that much overlap, despite the common wisdom that such a thing should exist, it’s very hard to actually find.
Also, not everything outside of faang is ‘basic code’ - there’s a whole wide world out there!
No, we do not.
In my experience disrupting spaces it's easier to find software engineers that are open minded when working with other professionals. Not like, super easy, most of the SWEs I've worked with can't do it, even on embedded when working with EEs... but open minded SWEs are out there, albeit usually more expensive or with good pay and job security by working on a less-softwarey company than "Big Tech".
Being able to natively express and review logic related to some legal software would be useful indeed, but honestly your team will need someone good at gathering requirements from non-technical folks either way and my guess would be that a lawyer that can vaguely code would likely be retained as an expert in the field but offered compensation like an entry-level to mid- developer by a startup.
I think that SMEs are great to have on the team, and I'd happily jump at the chance to contract with one that's got some vague coding experience as well - but I'd contract that out to pick their brains specifically on their specialty rather than try and shoe-horn them into a dev role. If I found a lawyer that had returned to development full time that's a different jazz - but you also need to keep in mind if the now-lawyer that used to be a developer actually wants to go back to full time development.
You'd be surprised...
Mick Napier who founded and runs The Annoyance where many famous comedians studied over the years, comedians who Second City would claim as graduates of their program like Tina Fey and Chris Farley, said they he doesn't like clamming taking the credit for successful graduates unless he also takes the credit for all the unsuccessful.
That has always stuck with me.
1) At least a year or two of experience at another programming job, 2) MONTHS of time studying whiteboarding, and maybe 3) A few good referrals.
I know, because I'm 34 and I have 9 years of enterprise experience. I'm trying to get into Big Tech right now (I've interviewed with Google and Amazon and I'm about to do Facebook). I'm lucky though that I have an undergrad science degree from Stanford, so I have at least some credibility in that world and many of my former peers are in that space.
The only really hard part is finding the will to study leetcode problems for hours.
I have all sorts of thoughts on the subconscious motivations behind the whiteboard grind. It's sort of a hazing ritual. Or, like so many things in America, a class filter initially designed to not be a class filter: rich people can afford to study for months for free. Poor people can't.
But ultimately, it's helpful to think of it as a game. Either play it or don't. There's no point in trying to resist it if my goal is ultimately to succeed personally. I keep getting rejected from Big Tech, but local firms are THROWING job offers at me just for being a decent, experienced programmer. I don't have much to complain about, other than the fact that I'd sort of like to make $400k for a little while and then drop out of society for a year or two.
And I think the large majority of people realize leetcode != software engineering ability. They're just trading off a low false positive rate for a high false negative rate. My wife has also been spending months studying for the LSAT. It's really quite similar except I can work as programmer without formal credentials.
No offense but I laugh at this thought. Because most people don't realize what they are actually capable of-- they'd rather make excuses.
Excuses are easier to make than to sit down, intensely study, and build projects, in a public library, while living in a tent.
And, I think a lot of people in the "special snowflake" mentality are afraid to recognize that a lot of poor people choose to be poor-- and even glorify it: much like the "gloating over/glorifying victimhood" movement.
As someone who has spent almost a year living in a tent while studying technology, I beg to differ.
I studied for free for several months, a few times, while living in a tent & in a cabin without running water/electricity. It simply requires a strategy to find: internet access, free food, low cost showers, relatively safe tent space near a town/city, etc.
About 8 years ago I stopped working in entry level Marketing ($14/hr) to live in a tent (so as not to have to pay rent) and learn web design. It helped my get my first jobs.
It's a long story-- after a few months, went back to work in digital marketing. Later, lived in a tent again to study web app dev. After dropping out of grad school (lived off loans while living in shack without water & electricity), lived in relative's living room in SF Bay area for 1-2 months while applying to jobs, and landed first programming job.
There's nothing wrong with living in a tent, on a couch, etc-- whether rich or poor.
Living poor didn't stop me-- the willpower to put in the work to actively learn (actively i.e. keeping journals of notes & building projects) is more important than having a place to live.
Anyone can live in a tent and eat food from a food bank for a few months, and shower at a community center, while learning. (Assuming they don't have kids to take care of. If they have kids before getting their career skills together, well... that's just poor planning)
Strategies involved: Motivational thinking. Motivational music. Cultivating positive thought processes. A strong belief in the economics data (bureau of labor statistics). And a strong belief that studying web dev will result in a job, assuming a person is organized, studies, and creates a relevant portfolio with projects.
__________
Note: I have relatives who are currently quite poor, and have been as long as I've known them.
Their problem as I see it?
They refer to themselves as "Poor people" instead of "People who refuse to apply to jobs" or "People who refuse to research economically relevant skills and then build said skills in an organized manner".
The reality is that being poor or not poor IS in their control. And they choose to be poor. Why? Because they choose not to do the alternative: work towards a strategy resulting in not being poor.
I've counseled a cousin who is about my age. He prefers to wake up in the evening, smoke cannabis, stay up all night, and occasionally work manual labor jobs during the day. At age 30-ish. Yet is intelligent. Just chooses to prioritize instant gratification.
It's not either or: 1) people should take more accountability for their own lives, but at the same time, 2) it should be easier to raise yourself up if you're less fortunate.
If you're willing to overcome the obstacles you can accomplish nearly anything. I'm just in favor of fewer obstacles.
But this statement:
>" it should be easier "
applies to literally everything in existence :P
challenges are good-- they result in growth.
I don't think any of them became engineers. One of them ran a catering business, for instance, which is more in line with his background.
Is that growth?
What makes you think that they should become engineers? Just because they're brilliant they're supposed to work in engineering?
I think whiteboarding problems can be silly but....
If you are applying to these jobs you either just graduated, in which case you just spent 4 years finding time to jump through all the hoops the professors gave you, surely you can find another 3 months to become an expert on whiteboarding problems.
Or you're a professional software developer. And the number of professional software developers who are good enough to work at Google or Amazon, live in the U.S. and are "poor" is a vanishingly small number.
Studying algorithms is HUGELY IMPORTANT. They say they "just want to see how you think." That's a lie. Big Tech interviews are like a whiteboarding final exam, and you need to shoot for an A.
Study leetcode. Study resources like interviewcake.com or Cracking the Coding Interview, and take practice interviews at interviewing.io.
Good luck, and feel free to hit me up for advice.
Oh boy, here we go again.
Smart people can afford not to study for months. 'Stupid' people can't. Shall I, as a 'stupid' person complain about that?
I probably wouldn't make 400k in my first FAANG job. More like 250k to 350k on the higher end if I were really lucky.
for a new grad in a non-CS STEM subject, do you suggest I spend this summer doing a no-name internship, study algorithms, or both?
I don't see why it would be surprising that tech companies benefit from technical founders. When I see MBA/finance run companies my first thought is that they are almost certainly a finance, contracting, or marketing business even if on the outside they are a consumer product, mining, or manufacturing firm.
Many mature businesses benefit from a financial viewpoint as CEO as there are no major innovations or risks to be done that can't be reasoned about on a spreadsheet. Tech has been on the other end of that spectrum for the last few decades.
Because you can essentially hop in and fix something, even if your other engineers are but stuck. Plus knowing that you can personally increase your wealth by 6 million dollars with a couple of good lines of code doesn't hurt.
Yes, there are exceptions to the rule. Especially in fields whose core business isn't software.
But in general it helps for founders to be able to, you know, actually build the thing they are trying to sell. I think that's all there is to it.
A necessary precondition for showing that something new & innovative is worth doing is usually to do it. It's not a sufficient condition though: you also need the risk tolerance to work on things new & innovative that may not work; you need the deep understanding of your problem to understand what the solution looks like; you need the location & social connections to access capital and hire real software engineers to scale your solution; and you need the judgment to hire the right people to fill in your skillset gaps. That's why most software engineers aren't off founding billion dollar companies.
Note that this applies to other founder-led enterprises. Fred Koch was a hell of a chemist before Koch Industries became a huge conglomerate. Warren Buffett studied under one of the greatest value investors (Benjamin Graham) on Wall Street. Carnegie started as a telegrapher and railroad superintendent before becoming a steel magnate. You need that operational experience and domain knowledge before founding a world-changing company.
Because there are many more software engineering jobs in the world than CEO jobs.
> Almost none of them could get a job as a software engineer at their own company, if they had to, because they would be completely unqualified.
Which probably reflects more that the hiring process is broken, than that these people couldn't contribute to a business writing code.
The real problem is that, in addition to writing code, they would also start identifying business opportunities and new ways to improve processes and increase revenues and cut costs, threatening the perceived job security of the people in management above them.
I think the obvious lesson here is when society becomes reoriented around a new field (in this case software), people with experience in that field are going to do well.
Around 1990 Japanese finance majors made the best CEOs, because the Japanese financial markets were near their peak and Japanese bank stocks were highly valued.
> Prospecting for oil is one concrete form of the abstract pattern of discovering value
So is literally everything.
Police officer: discovering value by figuring out the right person to arrest
Doctor: discovering value by figuring out what disease.
Etc
And, of course, any modern organization is totally dependent on software in some way, so understanding this stuff is always going to be helpful for a CEO when making related hiring decisions and signing contracts.
7/10 of the biggest companies are software companies. 6/7 of the software companies were founded by engineers. Ma is an outlier.
Any of these other points might be true, but there's no empirical reason to expect that that they predict "success." The simpler explanation is that software company founders tend to be engineers. Financial companies tend to be run by financiers, etc.
Conclusions like "discovery of solutions via techniques like continuous A/B testing" may be true, but they're not required to explain why engineer founded companies dominate the top 10.
The article itself explains it: software is swallowing the world. 7 of the top 10 companies by market cap are software companies (or closely related).
Do bankers tend to lead banking companies? A cursory glance at the top-5 banks in the US indicates that to be the case. They were largely founded by, and are currently run by, bankers.
The reason I was given for this advice is that they don't understand how iterations work in software, and so they can grossly misunderstand your level of doneness. This is first a big headache and then a source of post-release problems because you end up rushing things out.
Getting more software-savvy people into management in more industries would definitely help with that.
- Google/Alphabet's current CEO is a materials engineer [0]
- Apple's current CEO is an MBA with a degree in industrial engineering [1]. Jobs, the founding CEO, was also not a software engineer.
- Amazon's CEO, Bezos, was a software engineer [2], but it's widely known that he didn't do programming work for Amazon. I don't think he would consider himself a "software engineer" today. There is a HN regular who was an early Amazon employee and has talked about this (his name now escapes me).
That makes 7/10 of the "top 10" companies in this chart headed by non-software-engineer CEOs!!!
[0]: https://en.wikipedia.org/wiki/Sundar_Pichai
Also, the article is referring to founding CEOs, and hence referring to Page/Brin and Jobs/Wozniak, not Pichai and Cook.
[0]: From Cook's Wikipedia article: After graduating from Auburn University, Cook spent 12 years in IBM's personal computer business, ultimately serving as the director of North American fulfillment.
Also, needless to say, Jobs was not a software engineer either.
> *Wozniak obtained a master’s in computer science in 1987 and does code. Though Jobs was brilliant and technically savvy, Jobs clearly didn’t professionally code; he did contribute to creative and technical designs
"Contributing to creative and technical designs" does not make you a software engineer. LOL. Might as well give my grandmother the moniker as well (she's contributed to some creative designs for my startup).
--
Also, it's a terribly written article. I mean, this is content-marketing level, https://shortlyai.com type of stuff.
> Conversely, traditional CEOs are regularly incentivized and seduced into pleasing financial markets by having higher returns in the near term. This typically overhyped mission of “improved earnings” occasionally results in companies pursuing unsustainable schemes that are focused on boosting quarterly profits (liquid value).
Software engineer CEOs do this exact same thing, probably more so! (Mostly as an artifact of their Silicon Valley VC investors pumping them into SPACs and IPOs to get a fat payday.)
They paint a picture of non-software-engineer CEOs as idiots, which is completely idiotic. Most successful companies throughout history have not been founded by software engineers. This is a recent phenomenon and the article's author(s) have no idea what they're talking about.
In terms of title? No. In terms of function? Well sure, he's a co-founder.
In a small company, every person has enough influence to rival the CEO and every role overlaps enough that the distinction given by titles is meaningless inside the company. Mainly the distinction given by titles determines how people market themselves outside the company where the dividing roles will appear clear for marketing purposes, which they are not in reality.
The article actually refers to Jobs & Wozniak.
My view having known quite a few engineers who got MBAs (outside of the software world) is that it very much marked a point where people had decided they wanted to stop being a practising engineer and wanted to be someone who managed practising engineers (directly or indirectly).
I was already managing a team before I got it. After I got it I've had a mix of roles as both engineer and manager. I didn't intend to switch track, it's more about being better at both tracks.
Pretty much same comment for Apple, it remains to be seen if Tim Cook is another Jobs/Gates or if he is instead a Ballmer.
Regarding Bezos, if you read the article you will find that you actually made the one of the major points in the article with your comment about Bezos.
To your note on Bezos, the whole article makes no points about whether Bezos actually programmed or not, just that he thinks analytically and runs experiments. I mean, great for him! But this article isn't driven by any empirical data. Aside from its disingenuous claims about CEOs, it's a very emotionally written piece (doesn't feel like a software engineer wrote it ;).
I mean, read this:
> After much consideration, it would seem there are unseen forces derived from the nature of value itself that are biasing software engineering CEOs towards greater success, and the raw power of these forces appear to sculpt many of the successful practices of these leaders.
What???!??!? What does that mean???? "Derived from the nature of value itself ... the raw power of these forces ... sculpt many of the successful practices"
This reads like a college freshman's Business 101 term paper.
>So, obviously, those that understand the super infrastructure that is software have a competitive advantage over those that don’t. But, as these leaders move up, they aren’t actually the ones writing the software. So, what differentiates their management practices such that those under them are better able to create and scale value? Other forces must be at play, it’s not as simple as the company’s products and services are software based. In the next sections, we will drill into the value stages for likely causes.
I mean, he clearly isn't writing the code, but he also clearly understands how to run a software org based on him telling wall street that his team is working on a quarter 3 years out.
Read the potential value portion of article or read further about my understanding of value here: https://iism.org/post/glossary-value-76
"Economic value is also theorized to be the universal organizing force of secular society, and the nature of value is thought to be responsible for the emergence of the three dominant value management systems, each having proceeded out of its own particular value state. Individual contributors are often attracted to a particular type of value, and they are therefore intrinsically motivated by the unique properties of that value type. Often, humans perceive their own value to society in terms of their ability to produce or replicate a thing of value such as a meal, or concert, or a manufactured good. Others might perceive their value in terms of their net worth, and creative individuals might perceive their value in terms of their ability to discover new solutions or their ability to create new works of value."
How many more trillion dollar do you need , to give Tim Cook his due credit for operational excellence
Operations is great to have and Cook is really good at it. Focus on ops can destroy the creative process though.
I figure another 8 years and we'll know for sure (Ballmer clocked in at 14). I look forward to being wrong, Cook is a great human being.
Fixed the typo.
Tim Cook is about 2/3rds through of Ballmer's tenure. If you were to take a bet today, do you see Tim Cook doing the direction of Ballmer?
During that time, their stock price fell at least 10%. The launch of Windows Vista was bungled (XP outlived Vista). They had a string of failures in the phone market, before finally abandoning it. They had a string of failed acquisitions, including Nokia. Remember the Microsoft Stores?
Azure was the only real lasting success acheived under Ballmer.
Whenever I was buying a new laptop, I've only considered MacBooks because they did one thing I cared about a lot (high resolution displays in portable machines), and they were usually cheaper[1] compared to laptops I went for (lately X1 Carbons: worse performance [lower TDP CPUs], better battery life, worse but acceptable screen, keyboards way ahead — in essense, better keyboards and Linux compatibility + nicer design [personal take :] for worse performance and pricing on par or worse than MacBooks). M1 Macs now win even on battery life too, so it's going to be even harder to stay away from them going forward.
Still, keyboards and Linux compatibility will likely keep me in the HP/Thinkpad business lines for longer, but pricing is not the factor when it comes to choosing Apple computers or not for some jobs.
[1] Configuring a Thinkpad X1 Carbon 8th gen brings the Linux + i7-10610U 1.8GHz/16GB RAM/1TB SSD/2560x1440 screen up to $2022 today (down from "retail" price of $3140, which I hope nobody is paying anywhere, ever), and for roughly that money you get a 13" Macbook Pro with i5 ("up to 3.8GHz"), 16GB RAM, 1TB SSD and a 2560x1600 13" screen.
With the new Macbook Pro everything is just instant (I love the instant wake-up, the instant log-in with the fingerprint sensor...not sure how the Linux desktop compares). I remember how much effort Microsoft put in these things in the old days.
No that doesn't remain to be seen. Tim Cook is clearly one of the most successful CEOs of all time. Compare Apple now to Apple when Jobs died.
Apple certainly won't fall because of their sheer size and momentum (just like Microsoft never did), but we look at things differently at that scale and for a while, Microsoft was "stagnant" and "backtracking".
I don't know how people define "most successful CEOs of all time", but I do know that predicting future is a fool's errand (you can go into statistics and risk evaluation, but here we are talking about one person thus not much room for "statistics": I mean, everyone who's hugely successful but then hugely fails has been mostly hugely successful right until they failed).
The author goes into extreme detail describing exactly why software itself scales. This seems to have little to do with the CEO, though they keep jumping to comparisons between a software engineer's 'predictable' behavior, and the 'non-software' CEO's 'predictable' behavior.
Other businesses are constrained in a way tech companies will never be (geographical, meat space constraints) and are in legacy industries where decades of competition has eaten away margins.
The explanation that the tech CEOs are better (although I'd agree with that statement in general) isn't the best one.
Even if we did want to accept the definition of "better CEO" from the article, we'd need to look at performance over time: increase in market cap over their tenure.
Next, a cut off at 10 in an attempt to generalize is silly: why not top 1 when we are at 0%. Or top 2? Or top 1000? Top 10 is arbitrary and not telling.
And I am sure real data actually hides in the smaller companies market cap movement.
- Attention to detail
- Control of flow (if this then that) mentality
- More binary thinking
- Precision when describing (programming does not allow ambiguity)
- A lot of frustration tolerance :)
That's at least my experience
Probabilistic programming and machine learning do!
I can get that you usually don't get it from regular high school mathematics, but even there one should learn to deal with proofs and deducing stuff in a precise, non-ambiguous way, and you have to pay attention to detail.
Coding allowed me to build things with my analytical mind liking precision and patterns, so it's the way for my "mathematical" brain to express creativity.
Really? Should we count how many bugs happen due to type conversion, precision errors, off by one etc?
A) I think you greatly underestimate what other engineering disciplines bestow not to mention require
B) Really? Have you seen modern software development practices? barely functioning MVPs, vertical slices, pivoting and hacks galore....
Like personally I would wager that most programmers do not understand exactly what they are building when they start, I know I don’t always and I never really met anyone who does so either at least not constantly. You know the general direction then you tend to hack things until they work and optimize and fix edge cases as needed.
We very much can write programs without understanding what we need to do or even how things work otherwise we wouldn’t have bugs that are not hardware errata and even then one can argue that if you truly know what you are doing you plan for hardware failure.
What makes you think I underestimate those other disciplines? I do not. The other disciplines tend to focus on tangible rather than on abstraction. Programming (if you want to actually be good at it) requires you to tackle abstractions all day long. My theory is that that ability translates (in some people!) to being better CEO.
> B) Really? Have you seen modern software development practices? barely functioning MVPs, vertical slices, pivoting and hacks galore....
You are mixing stuff here. I can guarantee you that if you know what you want to build and have the clarity to make the right choices, the hacky project will turn out better than if you don't. Specific development practices (e.g. hacks vs. no hacks) has no relation to personal ability to formulate clear problem statement.
Formulating a clear problem statement is applicable to essentially every field and I would argue is probably the hardest thing to master and what most people fail at.
At least personally the most improvement I have experienced in formulating a problem statement had nothing to do with actual technical work but rather when I worked in professional services and had to come up with new service offerings, writing up proposals and managing stakeholders.
Formulating a problem is definitely a soft skill that I would say most purely technical programs don’t teach that well, quite often because you are given the problem statement to solve. This is also true for technical work and why some people might find it hard to progress in their careers. I’ve met plenty of brilliant and highly experienced engineers that suffer from major tunnel vision and couldn’t see the forest for the trees.
1. Software has become a global form of super infrastructure that facilitate scaling, but there's more going on. 2. I believe that guys like Zuckerberg, Bezos, Musk and Gates are using software effectively to speed value fulfillment cycles – pretty much in everything they are doing, from rockets to eradicating disease, or just shipping packages.
Software engineering CEOs are realizing value at faster rates, IMO, by executing well on all 4 stages of the value fulfillment cycle -- see https://iism.org/post/glossary-value-stage-82. I mean, these guys are all billionaires, so clearly the market thinks they are valuable (Larry, Sergey and Ma Huateng as well).
As to why: I believe it’s the relationship between emerging infrastructure and the value fulfillment cycle. Software happens to be the emerging super infrastructure, so software engineers are naturally emerging as the next generation of industrialists. Not to say there aren’t other emerging trends of significance; but for now, software and software variants is the emerging super infrastructure. Discovery of value has always been the result of iterative discovery – software just happens to facilitate quick iterations of value attempts, and software engineer tend to be more Agile than traditional plan driven companies.
Apparently it's not, see parent comment. Doubling-down and ignoring the criticism is downright dishonest.
I'd love to believe that software engineers make better CEO's in general, but you'd have to show me a list of companies that became successful only after hiring an ex-SE CEO, and preferably in another industry than information processing.
IMO, the answer to why the biggest companies tend to be started by software engineers is far far simpler and something you can find in your ordinary econ 101 textbook. Your standard natural monopoly is one which has high fixed costs and low marginal costs such that the first one to dominate the market can usually sustain that because their ongoing costs are low while any new entrant has to pay down the high fixed costs to enter the market. While usually associated with utilities, software matches that pretty well and software companies tend to be started by software engineers.
Writing software (the fixed cost) is extremely expensive as you need to pay a lot to get good engineers and you need a ton of them to handle the complexity associated with the largest markets but once you have written the software, it is relatively cheap to run (the marginal cost). Once one company starts dominating a software market, they can amortize the fixed costs across the greatest number of users and continuously invest to make their software continuously better such that it becomes almost impossible for a competitor to catch up.
You can see this today with the cloud as AWS rakes in billions in profits because they have the most features, biggest head start, and most IAAS revenue while the other companies lose billions (5.6 billion for google last year for example) trying to catch up.
Yes.
The original HN title was something like "Software engineers make the best CEOS, if measured by market cap".
Current, sure, but he showed up after Google was already successful. Becoming the CEO of Google made him successful, but he didn't make Google successful.
> Apple's current CEO...
Same thing.
> Jobs, the founding CEO, was also not a software engineer.
True but he co-founded it with a software engineer (Wozniak) who was arguably influential enough that he had the influence of a CEO if not the title.
> Amazon's CEO, Bezos, was a software engineer
He may not have held the title or had the function, but apparently he had the knowledge.
And the previous CEO of Google was Eric Schmidt, also not a software developer, and successful before taking the role at google.
>True but he co-founded it with a software engineer (Wozniak) who was arguably influential enough that he had the influence of a CEO if not the title.
Just about every "CEO" has people with outsized influence. Saying that Woz had the influence of a CEO but not the title kind of misses the point.
>He may not have held the title or had the function, but apparently he had the knowledge.
I know how to make a woodshed (I even did last year!). I don't think anyone would confuse me with a carpenter.
Wikipedia calls him a "businessman and software engineer". He has a PhD in EECS with a dissertation on "the problems of managing distributed software development and tools for solving these problems."[1] He never coded at Google but he started his career in a technical role.
People seem to be missing the point of the article. It's all about ways of thinking. It isn't literally saying that the more algorithms you memorise the better at CEO-ing you become. That's way too literal. It's saying things like, "if you've written and debugged software a bunch of times, you learn about iteration and predicting/root causing failures, which helps you be a better CEO".
So yeah. These people are CEO's, not software engineers.
Most US Congresspersons studied law, most high-ranking Chinese Communist Party members studied engineering. Most of both spent limited time practicing their profession, they're professional politicians, not lawyers or engineers.
It still says something interesting about the respective bodies. If there's some kid out there who wants to grow up to be a CEO, should she study computer science or, like, history?
It’s not like history is some useless knowledge. Software engineering is still the new hotness. Give it time and the headline will change to why meteor engineers make the best CEOs while we all pine for the days we sat on top
Also: being a monopoly helps.
If you actually want to show that software engineers wind up being more successful CEO's then you can't look at the top 10 companies, there is virtually zero statistical significance here whatsoever.
Instead, take a look at the Fortune 500 at a minimum (or better yet, all publicly traded companies). And then you need to rigorously define what makes a CEO "better" -- which isn't market cap. Indeed, it could trivially be the case that software engineers fail 10x as often as other founders, but that they found 100x more companies so they have more absolute successes.
The only thing that the top-10 shows is that there are a lot of tech companies in the top 10, and that tech companies tend to be founded by software engineers, while non-tech companies aren't. Which is... about as surprising as saying that mega-financial companies tend to be founded by bankers, or that restaurants tend to be founded by chefs.
Internet Boom (Confounder)
/ \
/ \
/ \
Software Engineers ----> Successful CEOsThe article has to do with this idea: "Software engineers deeply understand that you iterate your way to value through hundreds of failed attempts.”
I by no means think software engineers are smarter or more virtuous. What I do think is that the high rates of fail from writing code teaches them to discover value iterative -- hence the Agile movement. Learning is pivoting moment by moment, and planning one value attempt or value experiment at a time IMO. It is the nature of value that shaped this behavior IMO, not some superior virtue -- see https://iism.org/post/glossary-value-76
I also agree with others who said limiting analysis to the top 10 seems arbitrary, especially when you're covering mostly just the runaway outliers despite having the general overall claim of "why are software engineers better CEOs?" Does your analysis still hold when you expand your field of vision to the top 100, or top 1,000 companies?
You can hack some crazy stuff together really quickly and it's possible for it to have an outsized impact.
Because with software the time-to-market is instantaneous.
If other engineering disciplines, it seems a bit pretentious.
Software Engineering is bounded by the laws of magic.
Engineering requires creativity to harness a material to perform useful work, within the material’s tolerances.
Software Engineering requires creativity to create worlds that have no existence in reality, other than in the imagination of the software engineer himself.
And in today’s world, software engineering is essential to make physical engineering a reality.
TRY ME! I bet no one can think of a single thing that I can't say is just ultimately due to physics.
> And in today’s world, software engineering is essential to make physical engineering a reality.
It is literally quite the exact opposite.
If instead you look at industries which don't have as much of this future hype, they are much more competitive. Fashion, for instance. LVMH has a market cap of $283 billion and Bernard Arnault briefly surpassed Bezos as the richest man in the world in 2019.
Of course now take a look at this macro trends chart comparing walmart to amazon
https://www.macrotrends.net/stocks/stock-comparison?s=revenu...
Now tell me your best guess on when when Amazon will pass Walmart to take Fortune 1.
Edit: one thing about revenue and profit is that there are businesses that operate off of graft that kind of throw a monkey wrench into those measures
I'm not saying that Amazon won't overtake Walmart, but that tech companies benefit from future projections factored into their valuations, while other non-hyped businesses don't. If the argument is that "software engineers make the best CEOs", it seems more likely that companies in the software industry simply get higher valuations for the aforementioned reasons.
Of course through the lens of history, considering climate change, the best CEO was the one who didn't unleash a huge pile of CO2!
In a world of literacy, you are a fool to hire an illiterate person as a CEO. So we start to see the shakeout.
Look, in early Medieval Europe, Charlemagne could only write his name, and was functionlly illiterate for much of his life. But every Holy Roman Emperor afterwards was a as literate as you or I.
We are at that hand over stage now - every CEO in the previous generation could get away with being Charlegmane. Not any more
These companies have grown to be insane empires, not because it has been software engineers running them, but software engineering is a new field that greatly impacts the world at this time and they are the people that solved the harder mainstream problems at the time. If this was in the industrial age, we would see more blacksmiths and manufacturers and railroad builders on the list, than farmers or accountants or whatever. It is a mixture of time, depth of problem and an obsession to solve it. Not that there is something intrinsically special about software engineers
The opening paragraph of this article:
"From the onset, we have chosen to use one of Wall Street's measures of a better CEO – namely, market cap. In other words, by this measure, CEOs that create the most value are the best CEOs. Sure, there are other measures perhaps more virtuous; but market capital is well-known, generally reliable, and historically trended. So, let’s just roll with it as our measure of “better CEOs” as we have lots to discuss.
An objective review of the data leads to the conclusion: top companies are increasingly founded and managed by software engineers."
The first 100 words in this paragraph and the "objective conclusion" drawn is all you need to read to know the remaining 7000+ words are completely logically flawed. It's like making the assertion "the best restaurants in the world serve the most meals, so an objective review of the data leads to the conclusion that fast food restaurants have the best chefs."
They got fooled quite often. Not on purpose, the engs they hired didn't know any better.
This was also true for technical people who would hire outside their domain. Hardware people hiring web devs were also fooled often.
There is a hurdle though, that being a CEO of a startup is a full time job. A consequence is that if you happen to be a strong engineer with relevant skills in a cash strapped environment, you will come under strong pressure to be an engineer and perform poorly as a CEO unless you can raise the funds to hire a top-tier engineer.
Now I get to insert yet another gripe against Europe and its horrible environment for fundraising. I believe that the lack of technical CEOs is likely due to the lack of resources (both at startups and large companies) to develop them.
I'd say if you don't have that perspective and are CEO of a software driven company, I'd make sure that you have a very close group of constant SE advisers to help you out.
I'm also not sure many of the others qualify either, were Sergey and Larry "good CEOs"?
As others have noted, the correlation between developer CEOs and success is that many of the most successful companies today are software companies. Is there even a single large non-tech company with a developer as CEO?
If anything I would assume developers are substantially worse than average as CEOs, from lacking interpersonal skills to exaggerated focus on details.
Wozniak is many great things, but sadly yes he probably would have been:
However many if not all of these companies have gained most of their market value under the management of fairly ordinary business-type CEOs and management, who at best have some expertise in a technical field.
So clearly software engineers are the smartest people in any organization. No wonder they become CEOs!
That is a recursive algorithm.
Unrealistic scope expectations are a huge problem for business people without the technical background to intuit pitfalls.
I suspect that within 25 years that mix will change as being a "software company" will become less of a meaningful distinction (and to some degree already is -- what is a "tech" company these days?). Every company will be a "software company" and won't lean as heavily on its leadership being having a specific background in technology.
ETA - ah, looking at comments I see others have already made the same point.
These are also people who have figured out how to make software development work for them, not kill them. That's huge.
An interesting study would be to see how people with software engineering background (by some agreed upon definition) fare as CEO's of companies that they didn't start or own. Granted, that might take more time and study.
That being said, i can't agree with the statement that sw engineers make the best CEOs. But i tend to agree that a basic knowledge of IT/Software/Hardware is a must.
What would be notable is if there were many software engineers leading non-software companies. That does not seem to be the case.
The best tech startups I've worked for had technical CEOs or co-founders with engineering backgrounds ("best" meaning good to work for and most successful.) Worst were guys who thought they were going to somehow be the next Bill Gates just because he didn't graduate college, either.
Software companies grew the way they did because, for a while, there was little regulatory capture, with tech doing what tech wanted, at least with the unicorns. Thiel, PG have commented on this.
There was a recent article on HN from O'Reilly - the new wave of breakthroughs anticipated, in biotech, might not have room for engineer CEOs, nor the unicorn growth of software-based start-ups.
Half of the Founding CEOs shown are not actual software engineers but other types of engineers.
In the vein of the article, "Why do article writers make basic off-by-one errors, engage in clickbait, and treat logical fallacies as the truth?"
Software is eating the world. IT costs and benefits are high. Getting IT right in many organizations is now more vital than ever before.
The author says nothing to address that. In fact, not far into the article they effectively state they're just going to assume that his hypothesis is correct and then proceeds to hunt for reasons why. This raises all kinds of red flags for me.
I'm sure you don't have to look far for a common cause explanation of the hypothesis.
Since most startups fail and the failure is usually due to lack of cashflow, startups that has the founder as a software engineer can have more bootstrap resources.
Which potentially increases the chance of these companies to cross the chasm.
Because domain knowledge is needed.
That being said, many companies have hierarchies of clueless managers between the CEO and the engineering team. Having someone at the helm who understands the good-fast-cheap triangle [1] in tech driven companies is from my point of view a good thing.
This doesn't imply that every software engineer can be a CEO. It requires a lot of other (soft) skills to do this job.
[1] https://www.explanimate.com.au/wp-content/uploads/2017/09/go...
Not sure why it's so insightful that people who understand the business also run the business.
I agree with the article, but there's also a much simpler answer. To be successful in a world-changing sense, you have to actually care about making a good product. People get MBAs to make personal money, not to make world-changing products. People in these leadership roles got into software because they were interested in making cool products. Nowadays we see some people getting into software primarily to make some personal money - they won't succeed either, not in this world-changing leadership sense.
I disagree. Any major in any field may or may not pursue making personal money, build wold-changing products or have nothing planned for their future. I think you are projecting your own anecdotal experience to the general case.
Exactly the same way a salesman would be better at starting a sales company, etc.
There is a lot of charlatanism. If you ask Mr. Musk what’s your speciality, he would say I am a physicist.
Bezos is not a programmer.
>In 1986, he graduated summa cum laude from Princeton University with a 4.2 GPA and a Bachelor of Science in Engineering degree (B.S.E.) in electrical engineering and computer science.
Now that is secondary to IT.
In domains that are either all IT or where IT is a key part? :)
Many of the 'top companies' at the moment are youngish software companies, many still run by their founders or early employees, which include lots of software engineers.
If the current big thing was biotech the 'best' (ie most) CEOs would be chemists.