fix: deciding to work in tech
let's now consider that working in tech is the rational default choice
I and many others have attempted to persuade others to join tech in myriad ways. In this article, I’m flipping the mental model and showing that tech is the default choice. By setting tech as the default choice or null hypothesis, the non-participant must present a strong reason for opting out. I’ll give one example of what counts as a good reason at the end.
1. Why Tech is a Unique Industry
Consider the Cobb-Douglas economic production function:
This function states that the level of technology, A, directly impacts output in a linear fashion, in contrast to other forms of growth which exhibit diminishing returns to investment. In short, this is why technology is unique.
The technology industry is unique for essentially the same reason, plus one more. For the reason already mentioned:
Tech is Evergreen: Any productive economy will have tech
Tech is a particularly good investment because the returns do not diminish. Relatedly, tech should be an outsized industry in terms of headcount, firm count, and so on, as a result of added capital allocation commensurate with the added value proposition of investment therein.
There’s one extra reason technology as an industry is unique: It is cross-cutting on all other industries. In a real and interesting sense, it is a pseudo-industry.
For these reasons, we should choose tech by default. To make it clearer:
Since tech is expected to be outsized, a random choice under a veil of ignorance would most likely pick this industry.
Since tech is particularly lucrative, and because we work primarily to make a living, it’s even more of an obvious choice.
Since tech is cross-cutting, you may only avoid it by deliberately avoiding technology in your field of choice, which essentially automatically forces you into the low-performing group within that specialty, which is a bad thing that none of us want.
2. Anticipating Some Objections
On its face, my prediction has failed. I have claimed on the basis of Cobb-Douglass that the tech industry should be overallocated, but many empirical reports would disagree. Take, for instance, Indeed’s Report of the 13 largest US industries by GDP contribution, based on 2022 BLS data. The information industry is fourth, not first! The answer is threefold:
There is a measurement problem with BLS industry classification. What they call the information industry is not identical to the technology industry under the Cobb-Douglass lens. Real estate, the top BLS industry, has elements of technology (Redfin, Zillow, Compass…), as do “scientific and technical services” and many of their other categories.
Cobb-Douglass takes a long-run perspective. We shouldn’t be surprised to see much short-term noise at any particular point in time or any particular small window of time.
Cobb-Douglass is a pure economic model that does not incorporate political incentives. The third largest industry in Indeed’s report is government, and this industry is entirely left out of Cobb-Douglass. I argue that Public Choice, or political, incentives explain the remainder of the difference between my earlier prediction and our actual observation.
Notice that Utilities are the second largest industry, and this is essentially a quasi-governmental industry.
The largest industry in the report is real estate. I think we can explain this away on the basis of public choice incentives as well, but I don’t feel compelled to do so. I predicted that technology would be the largest industry. If it empirically ends up being the second largest industry, I already feel substantial validation in my theory.
To reiterate: measurement error and short run noise are expected to substantially blur my prediction anyway, so I see the empirical reality as totally statistically consistent and confirmatory on my model, argument, and prediction.
A second objection would be that people should not persue a career only to increase their own wellbeing, but they should instead be motivated by welfare and helping others. To this, I respond:
Cobb-Dougless is a social welfare model, not a personal wellbeing model. We are already optimizing across all people.
If you want to disproportionately benefit society, you should definitely still be in tech.
Don’t work for a nonprofit, drive a nonprofit. How? By being a wealthy donor.
Don’t get a job in politics. That’s not an effective route to changing policy. I reiterate that you should take the emprirical work on Public Choice incentives seriously. Instead, create or contibute to a succesful organization and leverage organizational influence and wealth to change policy outcomes.
Take advantage of regulator lag and permissionless innovation. Instead of trying to change the mind of beaurocrats, get ahead of them. This is a more succesful strategy and it seems to be getting better over time: In the age of the internet, then social media, then crypto, and now AI, the competency gap between industry leaders and regulators seems to be growing, allowing for effective growth in technology over public policy to drive social welfare and social change.
A third objection would be that working in tech is infeasible. It’s too hard for most people. Tech workers are smarter than average, and working in tech requires advanced math skills, or so the popular myths go. The answer to this objection is fourfold:
We should ignore selection effects and survivorship bias. [TODO: Illustrate w link to tiktok talking to Dan the Recruiter or w/e about social networking]
Tech demands diversity over the status quo. Given that the average tech worker has some collection of traits, we should not conclude that such collection of traits are a requirement or even advisable for entry. In fact, having other traits makes one a more ideal applicant to many employers.
About half of tech workers are dumber than the average tech worker anyway.
Empirical work shows that an IQ of 100, the average, is within the normal range of tech workers today. Again, even an IQ under this number doesn’t disqualify anyone from entry. A substantial number of individuals with subnormal IQ will still have other skills or traits making them suited for at least some roles in tech. This data is consistent with the thesis that enrollment into tech is a proper default.
A fourth and final objection points out that seemingly anything can be called technology for at least two reasons:
Everything we are and everything we do began at some point. So, everything is an innovation and a technology in the grand picture of history.
Cobb-Douglass technology, A, can take a positive, negative, or zero value, so anything that has a positive, negative, or zero linear impact on output can be called technology, and that appears to be almost anything.
As the objections are two, so are the responses:
Yes, if we take a grand view then everything can be considered technology. All the more reason for you to default into the field of technology.
In practice, economic talk is oriented around goods, and while bads and economic neutrinos exist, they are considered special cases. Similarly, when we talk about technology, we are typically talking about a thing that causes an increase to the level of technology and a net social gain. People, firms, and societies tend to throw out practices with net social cost over time.
There is almost always some time window relevant to any particular discussion, although it may not be explicitly stated. Paying attention to this is actually a great analytical key in general.
It’s actually not at all clear that most things have linear effects. The linear effects we know about may be overrepresented because checking for linear effects is easy, and analysts will generally look for linear effects before they look for other kinds of effects. In business, we often use linear approximations while knowing that they are not accurate representations, and this is often low-hanging fruit for improving business analysis. Deeper analysis actually tends to show that decreasing marginal benefit is a more general feature of real things in the economy.
3. Good Reasons to Avoid Tech
I can basically think of three, and I’m tempted to argue against all but the first:
You gave tech an honest try and you have found that you have an outright disutility from engagement.
I wouldn’t advocate for anyone to continue on in a state of torture throughout their life! The concept of tech as a default means you should try it out. It doesn’t mean that you have an obligation to persist if you try it out and it doesn’t work for you.
I encourage everyone to follow a three-tries heuristic here to establish an honest try. One bad experience with technology doesn’t amount to a real try. The tech industry is diverse, and diversity means it will include many poor fits as well as many good fits for many people. Even within programming in particular, I encourage folks to try out a few different programming languages on a few different projects before concluding that they don’t enjoy programming.
You have a really strange situation that makes your incentives very different from normal.
As an example, your parents run an artisan craft business and you are expected to take that over when you grow up. In this case, you may actually optimize your own life by learning obsolete skills instead of learning or contributing to the development of new technology.
I would still encourage you to take seriously the possibility that getting into tech might be better for you and your family than maintaining such an old craft store, or whatever analguous incentive you are facing.
You have serious medical impairments that directly prevent you from participating effectively in tech.
Tech is actually extremely friendly to workers with many sorts of disabilities, relative to other industries. You don’t need to be strong or even be able to walk. Tech is friendly to many kinds of neurodivergent workers.
Still, there are some forms of impairment that make this work infeasible. I don’t have a particularly good list of what these would be, but I acknowledge the category for a minority of the population.
Earlier I mentioned that “I’m tempted to argue against all but the first” and that would include this category. I stand by that because I think many individuals with one or even multiple disabilities underestimate how friendly tech can be to them. I’ve seen succesful tech workers that are blind, deaf, unable to walk, have certain neurological issues, and more. I encourage you to give it a try!