← Content

Intellectual Darwinism

In our world of technological frontiers and relative abundance, Social Darwinism – with its fatal outcomes for the disadvantaged and despairing – is hyperbole of what would today be downward socioeconomic mobility for individuals who do not adequately contribute to the global economy. With the emergence of institutional safety nets and welfare states, the stakes of intellectual underperformance are non-existential; however, it appears the product thereof can be similarly consequential, looking anecdotally to cities like Detroit and Pittsburgh.

Socioeconomic rearrangement as a product of economic contribution is not new at both the state and individual levels. However, the advent and adoption of frontier technologies (for the sake of this essay, I will speak to AI, but quantum computing, biotech, spacetech, etc.) will likely present new frameworks for productivity, and thus incite individual mobility along the socioeconomic ladder. In this context, mobility is liberating, and its pursuit endemic to the American dream. Understanding the nature and potential impact of frontier technologies on socioeconomic mobility for the individual is the difference between riding the next wave or being quietly left behind it.

Let us begin with an analogy to the information era of the 1990s, because AI is not the first time human-machine interaction has undergone a serious recalibration. A 2025 article titled Equalizer or amplifier? How AI may reshape human cognitive differences suggests that the period "reduced inequality by ability, and increased it by education". Put concretely: low performers gained access to tools that let them produce at a level closer to their middle- and high-tier peers, while the highest-tier performers used those same tools to pull themselves dramatically further ahead.

Consider the typists and bookkeepers, who saw their professions commoditized seemingly overnight; with the advent of software tools like Word and Excel, the cognitive barrier to entry for basic administrative and financial work dropped dramatically, and with it came new opportunities for unskilled labor to move toward the socioeconomic mean. Contrarily, those same computers gave rise to the Bloomberg terminal, algorithmic trading, and the spreadsheet-powered analyst. As a result, the finance sector's share of U.S. GDP doubled between 1970 and the 2000s, and generational wealth was minted across Wall Street. (x)

The personal computer rewarded procedural skill and information access, hence the extreme wealth generation across hedge funds and top investment banks. I believe what is unique about the advent of AI (and the commoditization of intelligence more broadly) is its reward structure; while many will use the technology (as did the PC), those who possess a deep intellectual vitality stand the most to gain. Further, I believe that those who use AI – specifically in educational or professional settings – for a quick answer, or to maximize productivity without taking ample time to push back on the LLM, ask questions, and read the answers, are putting themselves at a significant disadvantage in the long term; this type of person is using AI to boost productivity at the expense of intellect. In my experience, to truly learn, you must struggle with the material; read it thrice and even then find comfort in not always understanding the words in front of you.

This notion of "brain, then LLM" workflows being advantageous to both cognitive development and effective AI use was explored in a 2025 MIT Media Lab paper titled Your Brain on ChatGPT, which split 54 participants into three groups to write essays using either an (1) LLM, (2) a search engine, or (3) no tools at all, ("brain"). The study measured both brain activity via EEG and essay quality (judged by MIT English profs). Intuitively, the "brain" group showed the strongest neural connectivity and the LLM group showed the weakest. What is fascinating is that, after completion, the subjects were asked to rewrite one of their essays, but the LLM group had to do so without the AI tool, and the brain-only group could now use an LLM. The first group remembered little to none of their own essays and showed weak brain waves, whereas the "brain first, then LLM" group showed a significant increase in brain activity across EEG bands and strong memory recall. (x)

What the MIT study suggests at the level of individual essays, I believe, will play out at the level of careers and livelihoods. There is a fundamental asymmetry in how individuals choose to engage with the machine, and that engagement will – in a truly Darwinistic sense – shape their long-term socioeconomic opportunity.

The rows of the matrix below describe performance before AI adoption. High performers are the professionals and academics already excelling in their fields; low performers are unskilled workers, or those in jobs widely considered "at risk of disruption by AI." The columns describe how each individual chooses to engage with the tool. Low curiosity captures those who turn to AI for a quick answer or purely short-term productivity gains. High curiosity captures those who engage actively — practicing the "brain, then LLM" workflow from the MIT study, pushing back on the model, and continuing to ask questions.

Each cell explores the individual's likely change in socioeconomic opportunity. A low-performing but curious person (top right) – the modern analogue of the early Excel adopter – now has tools to pull themselves toward the socioeconomic mean. Contrarily, the current high performer who is not genuinely curious has built a career on an intellect that has now been commoditized; this individual is in a significantly worse position than they were five years ago, and many will not realize it until it is too late.

Low CuriosityHigh Curiosity
Low Performer ↔️ Neutral
Copy-and-paste users. Short-term productivity gains at the cost of compounding intellectual loss.
⬆️ Lifted
The modern Excel adopter. Previously lacked tools to compete; now has the leverage to mimic the top and pull toward the mean.
High Performer ⬇️ Eroded
Smart, lazy. Historically rewarded because intellect was scarce – that scarcity has ended. Drifts toward the commoditized middle.
⬆️⬆️ Amplified
The compounders. Already capable, now equipped with a tool that amplifies the exact trait they already possess.

A necessary caveat (and topic for another day) is that curiosity, or rather the freedom to be curious, is not evenly distributed globally. Children raised by engaged parents, in good schools with ample time to be bored, and exposure to complex problems early on, I would presume, are far more likely to develop the intellectual desire that undermines this essay.

The quadrants are hypotheses, and the impacts are delayed. It may take years or even decades to see the cultural impact of prolonged AI in professional and educational settings; nevertheless, the topics spoken here will reshuffle the deck, lift many from poverty, and fundamentally alter socioeconomic normality across the world.

This is what I mean by intellectual Darwinism: a selection pressure, imposed no longer by a necessity to survive (as with traditional Darwinism) or by some social or ethnic predisposition (as proposed by Social Darwinism), but rather by the tool that has commoditized humans' greatest differentiator. The sorting trait is intrinsic and controllable, and I fear the future is grim for those who hand their mind to the machine.