High-Agency Individuals

The phrase “you can just do things” is unarguably appealing. It’s simple, informal, and empowering. The casual air makes action feel accessible rather than daunting, while the vagueness implies a universal relevance. What will you do? Anything you want. How can you do it? You just can.

If you’ve seen this sentiment shared in recent months, it was probably by someone at least adjacent to the world of tech, where the slogan has been growing in popularity for several years. It’s appended to posts about anything from Anthony Bourdain to the invention of Invisalign to the actions of Luigi Mangione—reframing European holidays, dental hygiene, and alleged assassination as examples of permissionless action and bold, outside-the-box thinking. You can just do things, you know?

The phrase is used both ironically and sincerely but embodies more than just the internet’s endless capacity to run jokes into the ground. It’s part of a wider cultural appreciation for self-directed action and intentionality (also glossed as “agency”) and, like so many memes, embodies a particular response to the political and social world we live in.

In this case, it’s a world in which many believe that meaningful action is obstructed by burdensome procedure, either in the form of sclerotic state institutions or oppressive cultural diktats. The reminder that “you can just do things,” then, has accrued a totemic power. The phrase is used by both left and right, by CEOs and shitposters alike, and is often accompanied by exhortations to become “more agentic” or “high agency.” The advice welds together self-help cliché, entrepreneurial spirit, and cod-Nietzschean will to power into a single maxim: you can do it if you really try. That so many feel the need to reiterate this claim so frequently makes you wonder what has troubled their sense of agency in the first place.

On a personal level the principle that you can just do things is, in my opinion, broadly true, admirable, and helpful. It’s a reminder that we’re frequently constrained as much by our own imagination than by external factors, and that action—any action—is often preferable to overthinking and paralysis. You could try to trace the history of this idea, but it’s so broad that it resists meaningful narrative. It’s simply a truism of the human condition that has been framed and reframed by successive generations. You can find echoes everywhere from the aphorisms of Epictetus (“Now is the time to get serious about living your ideals. . . . How long can you afford to put off who you really want to be?”) to the essays of Ralph Waldo Emerson (“The power which resides in him is new in nature, and none but he knows what that is which he can do, nor does he know until he has tried.”). But while in its simplest form the notion of just doing things has no set ideology (apart from a vague sense of libertarianism), its current vogue exemplifies specific political and social trends. More specifically, it speaks to a pessimism about society’s capacity to improve itself; throwing responsibility back on the individual, like the “entrepreneur of the self” described by Michel Foucault.

In the tech world, where the phrase is most popular, the exhortation to “just do things” seems in part a response to the specter of automation currently threatening this white-collar industry. A recent piece in the Wall Street Journal lauded “high-agency” individuals as those best positioned to capitalize on recent advances in AI. In the op-ed, agency is described as a personal characteristic that not only preserves your current economic value but allows you to succeed like never before. The writer explains that if AI tools are increasingly able to accomplish basic tasks like writing code and copy, then it will elevate select individuals with the vision and will to execute on their own ideas. Similarly, AI execs talk fondly of the concept of the “one-person unicorn”: a billion-dollar company with a single human employee and legions of AI subroutines. The implication is that vast riches await those high-agency individuals who can marshal the free labor of AI. Those among us who can just do things.

The downside is that the labor being automated now includes much work typically done by tech employees themselves. Companies in the sector have been laying off workers by the thousands and crediting AI with the cuts, while reports suggest new graduates in the computer sciences are struggling to find employment. As a recent New York Times headline put it: “Goodbye, $165,000 Tech Jobs. Student Coders Seek Work at Chipotle.” The anticipation among some analysts is that this situation will only get worse as companies build increasingly powerful AI agents—tools that directly take over your computer, wresting mouse and keyboard from your control. Are these programs easing a worker’s burden or relieving them of it permanently? Who’s in control here? Who has agency? In this context, repeating the mantra that “you can just do things” is an anxiety response; an expression by those threatened by automation to distinguish themselves from the unfortunate masses.

Sometimes the threat goes deeper, too, from a loss of mere employment to a loss of meaning in life more broadly. Some in the tech world talk not only of the importance of being high agency but also “high variance”—that is, statistically anomalous and daring in both action and outcome. The shift in meaning is slight but revealing. To be “high variance” is to retain a sense of self in a world which is increasingly reduced to statistics, a world in which being “average” means being predictable and hence replaceable. This is an analysis that is both paranoid and reasonable. The dominant heuristic in artificial intelligence today rests on systems that comb through vast stores of digital text to identity and reproduce the patterns that are most statistically abundant. (This is why large language models are disparagingly referred to as mere autocomplete machines.) In such a world, to be average is to become fodder for the machine, subject to replication and redundancy. Instead, you must become high variance, unpredictable in thought and deed, always zigging when expected to zag.

The fear of having your freedom proscribed by statistical analysis has an interesting history. Scholars like Ian Hacking have noted how, in the nineteenth century, an “avalanche of numbers” produced by state statisticians bred similar unease, with the populace worrying that statistics constituted not just probabilities but ironclad predictions—not merely possibilities, but strict limits placed on free will. Ralph Waldo Emerson joked in 1849 about the “terrible tabulation of the French statists” which proved that if “one man in twenty thousand, or in thirty thousand, eats shoes, or marries his grandmother, then, in every twenty thousand, or thirty thousand, is found one man who eats shoes, or marries his grandmother.” In Dostoyevsky’s Notes from Underground, published in 1864, the tormented Underground Man (himself a former civil servant) decries attempts by bureaucrats to “deduce the whole range of human satisfactions as averages from statistical figures and scientifico-economic formulas.” Instead, he declares, the “best and greatest good” is to follow “one’s own free and unfettered volition, one’s own caprice, however wild.” To just do things, you might say.

This valorization of agency is more than just a cultural matter too. As select denizens of the tech world have been promoted to spheres of political influence, the concept of agency has become similarly important in political action. The clearest example of this dynamic comes from Elon Musk and his recent tenure cutting federal budgets and head counts at DOGE.

Musk himself is an exemplary high-agency individual; someone whose personal will is integral to his success, who prefers to operate by diktat than consensus, and who shows steadfast belief in his own genius. Musk’s supporters acclaim his ability to just do things; to brush aside the red-tape, dead tradition, and safety regulations of NPCs (short for non-player characters, a term Musk uses frequently and which also distinguishes between the predictable masses and the few, agentic player characters like himself). The only time Musk doesn’t take responsibility for his company’s actions is when they fail, at which point he tends to blame the meddling and jealousy of others, rather than address his own incompetencies.

With DOGE, Musk attempted to transplant his business acumen into government, conscripting a number of ambitious, intelligent, high-agency tech workers with the aim of slashing wasteful spending. The results of this experiment are broadly known: DOGE failed on its own stated terms, making far fewer cuts than Musk promised while senselessly slashing humanitarian programs and reducing American soft power. (As one former DOGE employee told the media: “[I was] pretty surprised, actually, at how efficient the government was.”) The writer and journalist Max Read has pointed out that the failure of DOGE challenges one of the “fundamental beliefs of the new Tech Right”—that “a sufficient number of sufficiently ‘cracked’ programmers can solve any problem put in front of them.” Here, the “cracked programmer” is a sort of Silicon Valley folk hero; an unorthodox worker whose erratic and sometimes slackerish behavior is excused by their high agency and genius insights.

When viewed from a political vantage, the deeper attraction of “agency” becomes apparent. In our current political climate, it’s presented as a solution to the problems of stagnant government, appealing to a broad range of the political spectrum. On the left, it accords with the complaints of “abundance liberals,” who advocate for governments to ignore obstacles like NIMBYism and environmental protectionism to build, build, build: build houses, more public transport, more power infrastructure, and so on. On the right, agency combines with neoreactionary politics that advocates for a corporate dictatorship as the ideal form of government. In February, one conservative commentator even identified “you can just do things” as the guiding principle of Trump’s second term, reflecting his expansion of executive power at the expense of congress and constitution.

In a climate of deep political pessimism, this sort of behavior can be attractive even to would-be political opponents. For example, when chipmakers NVIDIA and AMD agreed to pay the U.S. government a 15 percent fee on sales of AI chips to China, one viral tweet responded: “I wish democrats had realized ‘you can just do things.’ You can extort American corporations if you want!” See also, the reception to Dan Wang’s Breakneck, a book published in August that contrasts America and China, characterizing the former as a lawyerly society, predisposed to block action both good and bad, and the latter as an engineering state, inclined to plough forward with new enterprises without much introspection. Commentators on both left and right have regarded the Chinese model Wang describes with some envy (usually ignoring or at least dismissing the warnings he makes of an engineered state that treats people as numbers in a spreadsheet).

More broadly, the valorization of “agency” is also an adaptation to a crumbling social system which no longer offers support or meaning to the individual. This usage is a rebranding of established neoliberal thought: if you are struggling in life, if you’re anxious or lonely or can’t afford the rent, it’s because you are simply not being agentic, you’re not trying enough. In this framing, social challenges become the responsibility of the individual, and collective responses are framed as a form of moral infirmity. (“The fundamental weakness of western civilization is empathy,” as Musk put it.) The only real change is that the stakes have increased. Instead of pulling yourself up by your bootstraps, walking round local businesses with a pile of résumés till you get a job, you should bootstrap your own billion-dollar business. After all, the tools are all there, so why not show a little bit of agency?

The irony is that while political and technological developments are encouraging people to “just do things,” these same developments are making human agency harder to exercise, particularly with regard to AI. In the cultural realm, the replacement of artistic choice by AI tools means removing a level of intent and decision-making. A filmmaker might, in the past, have carefully considered how to light, frame, and block a certain shot; in the future, AI will make these decisions instead, drawing its decisions from a weighted average of its training data. In online spaces, arguments are now ceded to AI adjudicators. On X, the cry of “Grok, is this true?” is slowly taking over from human debate (even though it’s clear that one highly agentic individual, Elon Musk, has his finger on the scales when it comes to Grok’s opinions). Even worse, people trying to make a specific point don’t even bother to formulate the argument themselves. They just get machines to justify their thoughts for them.

And AI doesn’t need to live up to the promises of superintelligence in order to be installed into powerful systems previously governed by humans. Already we have chatbots dealing with customer complaints; algorithms that decide whether to grant parole to offenders; and AI models being used to generate military targets. Everywhere we look, the world is ceding control to automatic systems that cannot be reasoned with like humans and whose decisions are often inscrutable, interrogated only after the damage has been done. The delegation of such responsibility is often embraced in the name of efficiency or neutrality, but can also be manipulated by those in power while disenfranchising the masses. This is, perhaps, the real reason that so many people are keen to tell one another that “you can just do things.” It’s a reaction to a world in which, instead, things are just done to us.