Daniel Hook, chief scientific officer at Holtzbrinck Group, took a Waymo across San Francisco and noticed something uncomfortable: as a self-confessed introvert, he found the empty driver’s seat delightful. He was in town to collaborate with Susan Winslow, chief executive of Macmillan Learning. Two people who had crossed an ocean and a continent to talk to each other were quietly pleased to be in a car whose defining feature is that you don’t have to talk to anyone.

He names what they had stumbled into. The Waymo effect is what happens when a technology removes the friction of dealing with another human being, and we experience the removal as pure gain — because the costs of that friction were always visible to us, while its benefits were not. It is Tim Wu’s tyranny of convenience with a mechanism attached; Borgmann’s “device paradigm”, where the commodity keeps arriving and the practice that once earned it quietly disappears. The taxi driver, he suggests, was for many of us the last stranger we were obliged to encounter.

Then the turn: large language models are the Waymo of intellectual life.

  • A collaborator is available occasionally, between teaching, deadlines and time zones; a model is available at 2am on a Sunday
  • A collaborator arrives with their own agenda, their own framing, and an inconvenient conviction that your central assumption is wrong; a model arrives with no agenda beyond being useful to you
  • A collaborator must be persuaded; a model must merely be prompted
  • A model will challenge you as robustly as you request — and not one degree more. It critiques the argument you brought. It will not tell you that you are solving the wrong problem, or that a rival group tried this in 2019 and abandoned it

“The collaborator’s inconvenience,” he writes, “is not a bug in the collaboration; it largely is the collaboration.” His word for what thins out when every conversation is redirected from a colleague to a chatbot is decollaboration.

The part research leaders will not enjoy is his account of why this is not just individual weakness. Three forces compound: tightening budgets defund exactly the line items whose value is unmeasurable — travel, workshops, sabbaticals (“we defund the corridor and then wonder where the corridor conversations went”); evaluation systems still reward output and the speed of output; and the quietest force of all, which is that the model never argues about author order and claims no share of the credit. In a system where credit is the currency of survival, a brilliant interlocutor demanding none is not merely convenient. It is arbitrage.

He is careful with the evidence. The team-science literature finds small teams disrupt while large teams develop, and that atypical cross-field pairings produce the most novel work; early generative-AI evidence points the other way — individual productivity up, diversity of ideas narrowing. His phrase for the present is an uncontrolled global experiment in trading serendipity for throughput. And on writing, the point most relevant to anyone who has outsourced a first draft: writing takes time because thinking takes time, and the two are not separable stages. “Outsource the writing and you have not accelerated the thinking; you have skipped it.”

This is not a ban argument, and he says so. He cites Dashun Wang’s Nature comment proposing AI agents as aeroplanes for the mind, and Wang’s own prescription of pilot-in-command science — researcher as captain, agents as crew — along with Wang’s warning that individual gains can narrow collective diversity. He reaches back to Lisanne Bainbridge’s 1983 “ironies of automation”: the more reliably a system runs itself, the less practice its overseer gets at the skill they are kept in the loop to exercise. The problem, Hook concludes, is not the technology. It is that we have made the human conversation the expensive option and the machine conversation free, and then been surprised at what rational people choose.

His prescription is to fund the friction — treat collaboration as infrastructure, evaluate contribution rather than velocity, and ask “who did you think with?” with the same seriousness as “what did you publish?” The question he leaves on the table is the one worth carrying: who is in the front seat, and when did we stop noticing that nobody was driving?