DHH backs Proton CTO over AI memo leak
The adoption fight has moved from whether agents work to whether a boss can order you to use them.
Joseph Cox posted what he calls the Proton CTO's AI manifesto, a memo he says he obtained multiple copies of. DHH read the same document and came away with the opposite reaction to the coverage.
"There is nothing scandalous in this memo," he wrote, calling it hilarious that 404 is treating it as such. He says the memo is 100% correct, and that if employees at Proton are in revolt over this, they are in for a rough ride as the age of agents unfolds.
There is nothing scandalous in this memo. It's hilarious that 404 is treating it as such.
His actual argument
DHH's case for the memo is not that mandates are pleasant. It is that the alternative is worse. If you are stuck with an org refusing to adapt to reality, he says he has complete sympathy for some shock therapy, because otherwise the business becomes uncompetitive, customers eventually leave, and everyone is fired.
He also went after the reporting rather than just defending the memo, saying 404 is running the same playbook as The Verge, scandalizing routine business behaviour by giving an outlet to disgruntled employees and sensationalizing the age-old phenomenon that some people do not like their boss. His sign-off was a direct message to @andyyen not to let them fold him.
In a follow-up he went further on where he thinks the real delusion sits. The mistake, in his telling, is treating AI as just another tool, or a stochastic parrot, or something where humans still need to review every line.
The business read
None of this settles whether the memo is good management. It is one founder's opinion about another company's internal document, not a result, and the employees on the other side of it have not had their say in these posts. What it does mark is where the argument has moved. A year of debate about whether agents actually work has turned into a fight about whether a CTO can tell staff they have to use them, and that fight is about managers, not models.
The real AI psychosis is thinking "this is just another tool
Steal the numbers, not the vibes
One email a day. Under three minutes. Real MRR, real pricing, real launches from founders building with AI.