(Origin: "Mentofacturing") | Doc
Author: Vincent Lextrait, Founder at METASPEX
October 6, 2025
The purpose of this document is to fill a huge gap. I will attempt to explain how software companies operate here, where I live, in Silicon Valley. This way of operating is largely unknown outside. It is so unknown that this document will likely leave many readers in disbelief. I will provide evidence to support my claims. It is urgent that software managers, entrepreneurs, investors, policymakers, popularizers, influencers, educators, and practitioners realize the gap that began to widen decades ago and which largely explains American dominance in the software industry. It is not too late.
Of course, what I'm about to describe is also popular outside of Silicon Valley, but it remains essentially confined to the United States, and it started there. And even within Silicon Valley, there are some differences, for example, between Salesforce, Yelp, or even Square on the one hand, and Google and Meta on the other. Essentially, in Silicon Valley, the more a company's orientation is toward so-called "business" applications, the more it deviates from local standard practices. There's no reason for this difference; it's simply that the world of "enterprise" applications has always been more conventional than the rest. It evolves much more slowly, as Grace Hopper explained it many times. Attracting talent is also much more difficult and requires infinitely more work, which discourages most and makes them take the easy way out: degrading outcomes.
I've long hoped that someone would spend the time explaining the details of the chasm of differences I saw between, say, Google and Meta on the one hand and how the rest of the world operates. That explanation never came. I don't see this as a malicious attempt by Californian tech companies to hide anything. Rather, I see a complete lack of curiosity on the part of the rest of the world, and a well-established certainty that nothing can shake the outdated organization of software production work as we have it in place elsewhere.
A disturbing symptom of this ignorance is that no name has been given to Silicon Valley's practices, even here. Alexander Grothendieck, probably the greatest mathematician of the twentieth century, based much of his career on the idea that we must name things in order to begin to glimpse them. This principle is even rooted in Judeo-Christian culture; the Bible itself, in Genesis, describes how God begins by naming the elements of creation. "On the first day, he names the light and the darkness. On another day, he names the sky, the earth, and the waters." He gives names to the stars. Later, Adam is tasked with naming the animals. This act of naming is the first step toward being able to envision something.
In the absence of a name, and with a certain arrogance, I therefore named Silicon Valley's practices "mentofacturing." It's a neologism, a play on words to distinguish this culture from the one we inherited since Adam Smith: "manufacturing." The word "manufacturing" means "made with the hands." Software, on the other hand, is made with the intellect, even if a keyboard and sometimes a mouse are involved. Hence the prefix "mento-."
For the curious reader, I have already verbally demonstrated the general difference between the two. A video in French is available. A document in English going into the most minute and dry details is also available. But the subject here is not the general difference between the two approaches to work organization. This isn't even to say that the manufacturing organization is the worst possible organization for delivering software. It is, but that's beside the point. The purpose of this document is simply to provide compelling facts about these differences in the software world, which is the equivalent of the pin factory that Adam Smith cited as an example in his "Wealth of Nations" in 1776, except that software is eating the world.
How did I identify these differences myself? I first noticed the divide between Silicon Valley practices and the rest in the 1990s. Thanks to the French company that employed me: ILOG. Initially, ILOG was inspired by this culture, with founders who had been exposed to the US approach to software, but I also had the opportunity to spend time on site, here in Silicon Valley.
I have since emigrated to the United States, settling in the epicenter of this silent revolution, the "Bay Area." Here, many software practitioners, who started their careers at Google or Meta, would themselves be stunned to hear how software is produced elsewhere. Many have never known anything other than mentofacturing. A manufacturing organization is inconceivable.
The primary characteristic of the difference in the mentofacturing organization is the lack of separation between thinking and doing. This separation is, on the other hand, characteristic of manufacturing. This difference was identified as early as 1975 by Alfred (sic) Brooks in The Mythical Man-Month. He wrote: "Thinkers are rare; doers are rarer; and thinker-doers are rarest." This statement contrasts completely with manufacturing, where the thinker is separated from the doer, who simply executes. Steve Jobs, in 1990, reiterated this principle: "Doers are the major thinkers." Dave Cutler likes to point this out as well. American culture, with its pragmatism, was conducive to adopting this approach. Freedom, learning by doing and learning from failure are founding values here. Steve Jobs went as far as to say that believing that ideas can be separated from doing is a disease.
One of the principles applied by Google in all its initiatives is "The doer decides." This is exactly what Brooks and Jobs explained. This means that thinking and doing are not separated. The doer is the one who has the elements to decide, because it is doing that is difficult and provides the knowledge to decide. Being separated from doing cuts this off, even if one has extensive past experience of doing. Judgment deteriorates very quickly, and one resonates with trends instead of reasoning.
It should therefore come as no surprise that structurally, that is, in the definition of mentofacturing roles, very limited technical decision-making power is placed in the hands of non-practitioners. And when we talk about "practitioners," we are not talking about people who play with technology, who practice it as a hobby, for pleasure. We are talking about people who write code that contributes to software that is distributed to real users. There is no CTO role in mentofacturing. Google has "technical leads." These people write code almost every day. This code goes into production. These people are responsible for the ultimate decisions. They aren't the only decision-makers; other "doers" make decisions too, of course. It's the collective intelligence of the "doers" that prevails.
Do you want to be part of the tiny group of elected "super users" who approve urgent code changes? You have to be a practitioner to do that. What's more, this prerogative can't be granted by management. They don't even have a voice. It's only obtained through a vote of confidence from other practitioners.
PowerPoint and the likes have been banned. No slides to present ideas in large meetings, wasting people's time. Detailed, structured, text-heavy documents are circulated in advance. Meeting participants are supposed to have read the documents before attending. Meetings concentrate only on questions. No handholding, no simplification, collective intelligence is the goal.
Agile methods are not present in Mentofacturing. Developers aren't treated like rowers in a galley. No one produces at the rate of a drummer. There's not even a database of change requests, cost estimates, or meter readings by diligent administrators ready to whack or prod. There are no foremen or laborers. Frederick Taylor must be turning in his grave. "Smart people self-organize," Steve Jobs said. This is true even when there are hundreds of them working on the same product.
Mentofacturing does not use offshore, while it is in full swing in software manufacturing. The irrational belief of the latter in the separation between thinking and doing creates the illusion that thinking can happen onshore, while doing can be geographically separated in low cost offshore countries. Mentofacturing companies have sites everywhere, not for unit cost arbitrage, but to find talent wherever it is. I met the head of Microsoft in India, he was saying, like Steve Jobs, that hiring talent is one of the hardest things and he was struggling in India.
Mentofacturing offers managers very limited control over recruitment. Developers are recruited by technical committees, staffed by practitioners. A manager with a budget can only choose candidates from a list of people pre-recruited by the company. This ensures that no manager will be tempted to lower the recruitment bar to meet their budget, and incidentally, avoid losing it. New recruits have complete freedom to refuse an offer from a manager. Managers are in competition. If they can't convince new recruits to join them, they lose their budget. If they lose it too often, they are demoted and revert to individual contributors.
Mentofacturing never lowers the hiring bar (this was probably violated for the first time during Covid). The fact that software production is an intellectual act is well understood, as is the inevitable consequence that there are orders of magnitude variations in productivity between practitioners. There is a substantial, but far from complete, correlation between the level of STEM education of new recruits and whether they have successfully passed the hiring criteria. No knowledge workers, no skills development, all this is worthless. In contrast, manufacturing considers that subject matter experts are the scarcest and most valuable. What manufacturing calls “technology” is the product itself. Software engineering is just for doers. On the other hand, for mentofacturing, candidates are not expected to be "filled bottles," they must be "lit fires," as Albert Einstein said. Companies would rather not meet needs than lower the bar. The result is that Silicon Valley companies chronically lack development talent. Google currently has 6,000 unfilled positions. Despite the abundance of candidates, Google didn't adjust its recruitment criteria to the market. Mentofacturing assumes that talent is the source of prosperity. This is why Steve Jobs said his number one responsibility was recruitment.
Mentofacturing organizations put their money where their mouth is. If talent is the source of prosperity, it comes as natural that loyalty is expected to be commensurate to trust. If you work for a tier 1 technology company and you are delegated to entertain discussions with another tier 1 company, you receive no instructions whatsoever on what secrets you can or cannot disclose. Nothing. No. Thing. Better, after the meetings you are not asked to explain what you disclosed. Trust is foundational. The employee is not some "agent" who is guilty by default and must be controlled by all means necessary.
In Europe, I was faced with the need to explain why I was applying mentofacturing in the organization I worked for. The general culture, as almost always in Europe, was, in a caricatured way, manufacturing. I was told that we could never "offer careers" to the people I recruited after personally analyzing 5,000 CVs per year. We were to recruit talent, but in small numbers, for management positions, and recruit “small hands” to execute. This point of view speaks volumes. It presupposes that we necessarily operate in a manufacturing-driven world where we can't offer salaries commensurate with the contribution of practitioners, leaving them to be the thinker-doers. Despite immense pressure, I never budged. My answer has always been that without talent, we lose battles, whereas with talent, we create prosperity, and that prosperity creates opportunity. The future has proven me right, but to tell the truth, I wasn't entirely sincere in the sense that I knew the only way to be more rewarded in a manufacturing culture is to give up being a thinker and doer. But what I said wasn't wrong, even if it led to a very poor strategy, inevitably cutting off capable people from the roles where they would have the greatest impact.
How does Silicon Valley reward the talent of thinker-doers? I frequently see the question of developer compensation on professional social networks. Figures are cited, and the comments are incredulous. This question is easy to answer. Just visit glassdoor.com. It has all the necessary information. Here are two reference points at Google, which is known to reward its engineers significantly less than Meta: The title of "Senior Staff Software Engineer" has a median salary of $736,000 (salary plus stock grant). At the lower end of the scale, the median salary for a simple "Software Engineer" is $280,000. If you're curious, you can check out the same categories at Meta. It's common knowledge that the highest-paid developers total over a million dollars per year. They number in the hundreds. Under these conditions, there's no reason, as in manufacturing, to hold back on recruiting talented people in order to lazily seek out a majority of executors. On the contrary, we recruit everyone we can find without limit. Prosperity is used to reward thinker-doers. The vague interpretation of this outside of Silicon Valley is at best cynical, at worst hypocritical. We have developers incapable of managing people? To retain them, we must create a technical career path, where those "broken", who cannot become managers will find a way to gradually be better rewarded. This is a complete distortion of reality, as we have just seen. The mentofacturing salary pyramid does not necessarily reflect the managerial pyramid, and when we have a thinker-doer, mentofacturing does everything it can to ensure that they express themselves as best as possible and continue to be better rewarded according to their contribution.
Mentofacturing organizations have eliminated all forms of political gamesmanship. The manufacturing model implies fit for a role and therefore the ability to integrate into a process. The process is primary; the employee is replaceable. Manufacturing is process-centric. Manufacturing-inspired organizations inevitably encourage the use of surreptitious means to survive, especially in positions of responsibility. Managers are literally driven into moral corruption. Mentofacturing organizations are talent-centric, talent comes first, the process comes second. The slightest sign of dishonesty, manipulation, or simply maneuvering is punished by immediate dismissal. We recall Meta, who fired high-caliber employees for a few dollars of false expense reports.
If you think that putting process first and talent second vs. talent first and process second is almost the same and not even worth mentioning, you would maybe be convinced by Graham Greene's point about the differences between Catholicism (love first, virtue second) and Protestantism (the opposite). On paper this should not have divided two branches of Christianity. In reality, this apparently minuscule difference has created quite a few blood baths. It changes almost everything. So does process or talent centricity. The first, overwhelmingly favored in Europe has caused it to lag increasingly behind the United States in the software industry. Europe risks now total annihilation.
In a setting like this, you might conclude that managerial responsibility is frustrating? It implies a responsibility with little oversight beyond working on the motivation of contributors, identifying the most talented, and developing all the thinker-doers. This observation is generally true. It explains why many mentofacturing managers decide to become individual contributors again.
But, you might ask, why is it that if the talent of mentofacturing thinker-doers creates prosperity, seeds don't appear outside of Silicon Valley to convince others of the striking merits of this organization? Having been a leading player in this, I can say that the question ignores a fundamental problem. Manufacturing and mentofacturing organizations are stable. They return to their equilibrium after any disruption. In other words, if we assume the separation of thinking and doing in manufacturing, there will initially be considerable hurdles in recruiting if we want to mimic mentofacturing. Even if these hurdles are overcome, the organization will still be hell-bent on turning thinker-doers into managers. This extinguishes prosperity and turns the organization into a milking farm. And even if the organization doesn't turn thinker-doers into managers, the fruits of prosperity will be attributed to managers, because the general culture of the company is to believe that prosperity was caused by a separation of thinking and doing, and thinking is the lot of managers, whether they contributed to it or not. Prosperity will not be used to create the rewards we see on Glassdoor. Despite the presence of a mentofacturing seed that demonstrates its virtues, the organization will evacuate this to return to its stable manufacturing equilibrium. Even if blatant successes demonstrate the vast superiority of mentofacturing, managers will claim the merits of the thinking.
But even if the thinker-doers are retrained as managers, should they reproduce the mentofacturing principles and overcome the manufacturing culture? This is not the case, because talent is not correlated with a rebellious spirit. The human desire for conformity is the most widespread, especially if that conformity is with a company's culture. New managers will adopt the dominant manufacturing culture and will hire in turn pure doers. Steve Jobs called this the "bozo explosion." He added that people don't become pirates, they join the Navy. So, here too, there will be no conversion.
Similarly, the mentofacturing organization is very stable. If we introduce a separation between thinking and doing in a sub-organization, the thinker-doer talent will flee it, managers will be demoted, will revert to thinker-doers themselves, and will remain so if they are competent, or will be pushed out by the mentofacturing organization if they are not.
Is the mentofacturing organization as implemented by Silicon Valley without weaknesses? Certainly not. It can be improved in many ways. In particular, Silicon Valley has established a "nice guy" culture. Disagreement can only be mild. The talent density is such that a non-aggression pact has been implicitly established to avoid overheating in interactions. However, one can only innovate if one is iconoclastic: literally, a breaker of icons. Breaking is a violent action. So we can only create with an approach like Steve Jobs's, without the hassle of a nice culture. This is certainly where Silicon Valley can be beaten. But this requires first adopting the principles of mentofacturing.
The introduction to Réaumur's The Art of the Pinmaker, written by Henri-Louis Duhamel du Monceau, dates from 1761. It is the source that inspired Adam Smith's division of labor. It is not too late to start distancing ourselves from it. It is hard, though. Whether you are religious or an atheist like Arthur Schopenhauer, if you do not want to be the victim or the executioner of software manufacturing, it is worth pondering Schopenhauer's words:
The majority of men are not capable of thinking, but only of believing, and are not accessible to reason, but only to authority.
— Arthur Schopenhauer, Parerga and Paralipomena, 1851
At the end what matters is whether you prefer authority, even if sterile, or liberty.
Last modified 11 September 2026