The Techno-Humanist Compass
Hoffman calls his political philosophy techno-humanism. It is a deliberate position between two camps he finds incoherent. On one side are the AI doomers, who he believes confuse worst-case thought experiments with most-likely scenarios and who recommend interventions that would mostly succeed in transferring control of the technology to less cautious actors. On the other side are the unbounded accelerationists, who he believes confuse the desirability of technological progress with an abdication of responsibility for its direction. Techno-humanism, in Hoffman's framing, is the stance that the right question is not whether to build AI but how to build it so that it expands human agency rather than contracting it.
The compass metaphor is doing real work in this position. A compass does not tell you where to go. It tells you which way is north relative to where you are. Hoffman's techno-humanist compass orients decisions about AI by asking, in any given choice, whether the result expands what humans can do or contracts it. A system that lets a small business owner do their own accounting is agency-expanding. A system that locks a worker into a dashboard whose decisions they cannot inspect is agency-contracting. The compass does not produce a single answer; it produces a direction, and the direction is then translated into product, policy, and capital allocation.
The philosophical lineage here is liberal in the classical sense — Mill, Berlin, Dewey, and the broader tradition that treats human autonomy as the central value to be protected and expanded. Hoffman read these thinkers carefully at Oxford and has cited them throughout his writing. What he has added is the practical claim that autonomy is now mediated by technology so deeply that you cannot protect autonomy by leaving people alone. You have to actively build technologies that respect and amplify it. The libertarian instinct to minimize intervention does not work when the technology is doing the intervening.
The hardest question for techno-humanism is what happens when the technology that is supposed to expand agency is itself shaped by the technology it is supposed to orient. AI systems train on human data and influence human behavior. The compass and the territory are coupled. Hoffman acknowledges this loop. His response is that the loop has always existed in some form — humans have always shaped tools that shape humans — and the right response is not to deny the loop but to design it consciously. Techno-humanism is, in this sense, a practice rather than a position. You commit to keep checking the compass against actual outcomes, and you keep adjusting both the technology and your understanding of what agency means.
The position has critics on both flanks. From the doom side, it is accused of being insufficiently alarmed at scenarios where AI moves faster than human capacity to steer it. From the accelerationist side, it is accused of being insufficiently committed to letting the technology run its course. Hoffman is comfortable in this middle position. He has been comfortable in it for two decades, since his earliest writing on the internet. He treats the discomfort of the flanks as evidence that the compass is functioning — that he is being pulled in opposite directions roughly equally.
What techno-humanism rules out is fatalism. It refuses the claim that the technology has its own logic that humans cannot influence. It also refuses the claim that humans have unlimited capacity to influence the technology. The compass is a tool for navigating the narrow corridor between these two errors. Whether the corridor exists, and whether it remains navigable as the technology becomes more powerful, is the question on which the entire wager rests. Hoffman has bet his career on the corridor being real. So far, he is still walking it.
