The Cost of Critical Thinking and Why We Must Pay It

The argument

The essay opens with a small research disaster. Omonijo, working on her dissertation as an intellectual historian, hit writer's block after chasing an uncited source referenced in 1990s scholarship. She spent two days with librarians and archivists trying to locate it. It did not exist.

That is the setup, and it does two jobs. It demonstrates the labour, two days and multiple specialists to establish a negative, and it makes the case for what she calls good friction: the difficult, unavoidable processes that genuine thinking requires. The cost in the title is that friction, and the claim is that we are increasingly unwilling to pay it.

Cognitive annihilation

She offers a working definition of the term in the subtitle. The risk, as she frames it, is not that AI outthinks humanity. It is that thinking itself stops being fundamentally human, that we retreat from cognitive difficulty far enough that the capacity, and with it a sense of self, goes with it.

It is a deliberately unsettling phrase and she treats it as provisional rather than settled, which is the right register for it.

On registers of writing

A secondary thread contrasts the different modes of writing the same person does, a dissertation chapter versus a Substack essay, and what each demands. The through-line is that ease on the page is bought with difficulty behind it. She reaches for Maya Angelou's 1983 line: _"Easy reading is damned hard writing."_

Other figures appear as company for the argument rather than as evidence: Teju Cole (_Known and Strange Things_), Claudia Rankine (_Citizen: An American Lyric_), and Audre Lorde: _"Even if you are afraid, do it anyway... Silence never brought us anything."_

A note on access

The piece is paywalled beyond the opening. The framing, the archive story and the definition of cognitive annihilation are readable free. The full development of the argument is not.

Why it matters

The current attention discourse mostly runs on brain-scan headlines and screen-time guilt. This essay does something more difficult: it argues from inside a discipline whose entire method is slow verification, at exactly the moment when the tools on offer promise to remove the slowness.

The uncited-source anecdote is the sharpest part. A system trained to be helpful would have produced something plausible about that reference in seconds. Two days of human work established that the correct answer was _nothing is there_. That gap between an answer and the truth is the whole problem in miniature, and no amount of interface polish closes it.

For neurodivergent readers the "good friction" framing needs handling with care. Not all difficulty is productive, and difficulty that is imposed rather than chosen is often just an access barrier. The useful reading is about not outsourcing the strain that is doing the thinking, rather than romanticising struggle as such.

Key takeaways

- Two days of expert archival work to prove a source did not exist. The labour of verification is the argument, not a preamble to it.

- "Cognitive annihilation" names the risk as losing the capacity to think, not being outthought.

- Good friction is the cost of real intellectual work, and the essay's claim is that avoiding it costs more.

- Partially paywalled. The opening and core framing are free.