AI cuts Josh Comeau's course sales to a third
Josh W. Comeau sells his third course at a third of the usual and blames AI: less incentive to pay for training and models that absorb his work without compensation.
Josh W. Comeau, one of the best known front-end course authors, has just launched his third course, Whimsical Animations, and it is on track to sell roughly a third of the copies a launch of his usually moves. It is not an isolated case: his two earlier courses are also declining, with revenue that he himself puts at more than 50% below last year.
He tells the story in a thread on Bluesky that Simon Willison picks up. And his diagnosis points in one clear direction: AI.
The double penalty
Comeau describes what he calls a "double whammy". On one side, many people wonder whether developer jobs will still exist in a few months, so they hesitate before investing time and money in learning new skills. On the other, those who do want to learn no longer depend on a paid course: an LLM offers personalized tutoring on demand, so the incentive to buy packaged training drops.
On top of that comes a third layer, the one that stings most: the models, he says, "slurp up all of our work and regurgitate it, without consent or compensation". In other words, the very material he charges for ends up feeding, for free, the tool that eats into his sales.
Not an isolated case
It is worth placing who is speaking. Comeau is not just any infoproduct seller, but an author with a reputation built on free, carefully written technical articles about CSS, React and web animation. That someone like him sees a launch sink to a third of the usual is not explained by a single bad month alone. He says he has spoken with other course creators and they all see the same trend: revenue down more than 50%, fewer people engaging with their content, an audience migrating to LLMs.
The signal is uncomfortable because it touches a business model that until recently seemed solid: the expert who documents a topic well and charges to teach it. If the bulk of that demand shifts to a chat that answers instantly, a good part of the economics of independent technical training is left up in the air.
The first factor he cites, the fear that the job will disappear, deserves a separate reading. The profession does not need to actually vanish for sales to fall; it is enough that enough people believe it might happen and put off the decision to train. Uncertainty, even if it never comes true, already slows spending. It is an expectations effect that no single tool causes on its own, but that the discourse around AI feeds every day.
Who this matters to
For anyone who makes a living selling packaged knowledge, courses, books, tutorials, Comeau's anecdote is a warning with numbers on the table. For developers, it reflects a climate of uncertainty that goes beyond the tool: if the students themselves doubt that the trade will exist in a few months, the problem is not only pedagogical. And for anyone building AI products, it raises the thorny question of where the training material comes from and who keeps the value.
There are nuances Comeau half acknowledges: market saturation for courses or changes in his own channels may also play a part. But it is hard to ignore the coincidence of three different products falling at the same time and in the same proportion.
At ElephantPink we live both sides: we use LLMs daily to speed up work and, at the same time, we depend on there still being people who document things well so that these models have something good to learn from. Comeau's testimony does not prove a global trend on its own, but it puts figures to a real tension between those who produce knowledge and those who automate it. If the sector does not find a way to compensate the former, the well the latter drink from will slowly run dry.
Sources
Read next
Synthesia moves from corporate video to avatar roleplay
Synthesia launches AI Roleplay Sessions: employees rehearse tough conversations with avatars that score and give feedback. What changes and for whom.
AI mania is degrading decision-making at large companies
Nik Suresh collects anecdotes from his consulting work: AI strategies signed off by executives who never used the tools, plus internal token consumption leaderboards. Simon Willison recommends it.
Spotify removes 75 million spam and 'AI slop' tracks
Spotify says it pulled 75 million fraudulent tracks in a year and rolls out a spam filter, AI use disclosure and a tougher rule against voice impersonation.