The Singularity May Already Be Human - in a Neurodivergent Way

Tech companies are hiring neurodivergent minds as a strategic advantage. The awkward question is whether their platforms helped create them.

In late 2025, Palantir Technologies announced a Neurodivergent Fellowship. The defense and data-analytics firm, whose platforms support government and enterprise decision-making at scale, framed the program explicitly as strategic rather than philanthropic. Neurodivergent individuals, the company stated, “will have a competitive advantage as elite builders of the next technological era.” Their pattern recognition, non-linear thinking, and capacity for hyperfocus were presented as assets precisely suited to an AI-driven world. A video message from chief executive Alex Karp, filmed while cross-country skiing, invited applicants who “can’t sit still” or who “think faster than they can speak.” The fellowship was not, Palantir insisted, a diversity initiative. It was an effort to recruit minds that see patterns others miss and that will help secure Western technological leadership.

From Inclusion to Instrumentalization

The announcement crystallizes a broader development. Leading technology companies increasingly treat certain cognitive profiles once regarded as deficits or eccentricities as competitive resources. At the same time, diagnosis rates for autism spectrum disorder and attention-deficit/hyperactivity disorder have risen sharply. U.S. Centers for Disease Control and Prevention data now place autism identification at roughly one in 31 eight-year-olds, a marked increase from earlier decades. Parent-reported ADHD diagnoses among children aged three to seventeen hover around 11 percent, with further growth in recent national surveys. While expanded awareness, broader criteria, and reduced stigma account for a substantial portion of these figures, the acceleration of diagnoses since the early 2010s coincides with the smartphone era and the restructuring of childhood around algorithmic platforms. The question is whether technology is simply revealing cognitive diversity that always existed or actively participating in its amplification and reconfiguration. The language was revealing not because it resolved the ethical questions around Palantir’s work, but because it made explicit how cognitive difference is being folded into strategic competition. This is not to claim that digital technology simply causes autism, ADHD, or neurodivergence. The more careful question is how technological environments make certain traits more visible, more burdensome, more rewarded, or more economically exploitable.

Platforms and the Remaking of Cognition

Psychologist Jonathan Haidt has documented how the “phone-based childhood” that became widespread after 2010 correlates with pronounced rises in adolescent anxiety, depression, and attentional difficulties. Constant notifications, infinite-scroll feeds, and the displacement of unstructured play fragment attention and reduce opportunities for the face-to-face interaction through which children historically developed regulatory capacities. Nicholas Carr, in The Shallows, argued that the internet’s interrupt-driven architecture, hyperlinked pages, constant context-switching, and optimization for rapid scanning, reshapes neural pathways away from sustained, linear concentration toward shallower, more associative processing. Sherry Turkle, in Alone Together, observed a parallel social consequence: digital platforms multiply the volume of connection while often diminishing its depth, producing the paradox of being “alone together.” Individuals report preferring the controlled, low-stakes nature of mediated interaction to the vulnerability and unpredictability of embodied presence.

These shifts point toward the production of what might be termed atomic individuals, persons whose primary cognitive and relational environment is a personalized stream of algorithmic stimuli rather than shared physical spaces, durable institutions, or unmediated social bonds. In such conditions, certain traits become newly legible as advantages. The capacity to detect signal amid informational noise, to maintain intense focus on complex systems for extended periods, and to tolerate rapid shifts in context aligns with the demands of software development, large-scale data analysis, model training, and the sensemaking work performed by firms like Palantir. It is therefore unsurprising that a company operating at the infrastructure layer of Western security and governance would seek to recruit such profiles directly.

Yet the same technological environment that appears to reward these traits may also intensify the conditions associated with them. Early and continuous exposure to variable-reward interfaces, the erosion of boredom as a generative state, the quantification of performance across work and social life, and the reduction of embodied, multi-sensory interaction all exert pressure on developing attention systems. Byung-Chul Han has characterized contemporary society as an “achievement society” in which individuals internalize the imperative to optimize themselves continuously. The pathologies that result, burnout, depressive exhaustion, and states of hyperarousal without stable direction, arise less from external prohibition than from the absence of limits on self-exploitation. Digital platforms operationalize this dynamic at unprecedented scale and speed, turning attention itself into a site of constant extraction and measurement.

The Singularity in Human Form

Peter Thiel, Palantir’s co-founder, has long argued that technological stagnation, not acceleration, poses the greater civilizational risk. His emphasis on “definite optimism”, the belief that the future can and should be deliberately built, underpins a broader Silicon Valley orientation toward rapid deployment, scaling, and competition, particularly in the domain of artificial intelligence. This orientation does not treat human cognition as fixed. It treats it as substrate. Something to be augmented, selected, and, where necessary, reshaped in service of capability. The Neurodivergent Fellowship is one expression of that logic. It assumes that certain atypical minds will prove disproportionately generative within the systems now being constructed and that recruiting them is therefore instrumentally rational.

The deeper issue is whether this recruitment is occurring within a feedback loop that also increases the prevalence or salience of the very traits being sought. If sustained attention, tolerance for ambiguity without immediate stimulation, and the capacity for unmediated social attunement are partly cultivated through developmental experiences now in decline, then the technological acceleration celebrated by its proponents may be altering the distribution of cognitive styles even as it claims merely to harness existing variation. The distinction between revelation and production becomes difficult to maintain cleanly. Both processes can operate simultaneously, and their relative weights remain empirically contested.

This returns us to the technological singularity, conventionally understood as the emergence of machine superintelligence that triggers exponential, uncontrollable progress. A more immediate transformation may already be visible in the human side of the equation. As AI systems become infrastructural, embedded in search, writing, analysis, and operational decision-making, human cognition increasingly operates in symbiosis with them. Interfaces optimized for fluid interaction with probabilistic models and agentic tools may privilege certain styles of thought (iterative prompting, comfort with partial or provisional outputs, rapid pattern completion) while de-emphasizing others. The atomic individual, whose attention is continuously partially occupied by personalized feeds and whose primary collaborators are non-human systems, inhabits a different cognitive ecology from predecessors whose primary interfaces were other minds, physical environments, and slow media.

The relevant question is therefore not only what machines will become, but what kinds of minds the emerging technological order will select for, accommodate, and perhaps inadvertently produce. Palantir’s wager is a bet on comparative advantage within a particular civilizational project. Yet if that project’s substrate also contributes to the conditions that make such profiles both more common in diagnosis and more functionally variable in outcome, then the celebration of neurodiversity and the production of attentional strain are not easily disentangled. The same infrastructure that identifies and elevates certain minds may narrow the range of developmental pathways available to others.

None of this implies technological determinism or a simple reversal of progress. Human institutions retain agency in the design of platforms, the regulation of childhood environments, the structure of education, and the metrics by which performance is defined and rewarded. The uncertainty lies in whether these levers are being exercised with adequate attention to the cognitive and social consequences of acceleration itself. The Neurodivergent Fellowship, read as symptom rather than solution, suggests that the terrain of cognition is now explicitly political-economic. Who thrives, who struggles, and which forms of attention are treated as resources or liabilities are increasingly shaped by the architectures of information and the incentives of those who control them.

The singularity, if it arrives, may not announce itself first in the surpassing of human benchmarks by machines. It may already be legible in the quiet restructuring of what kinds of minds find the contemporary world legible, habitable, and productive, and in the growing distance between those minds and the developmental conditions that once produced a wider distribution of cognitive styles. Whether this constitutes adaptation, selection, or unintended transformation remains the open question that any serious account of technological acceleration must now confront.