404media.co via Hacker News

Proton CEO Scraps AI 'Re-Interviews' After Memo Leak

Jobs Enterprise AI ai-business

TL;DR

  • Proton CEO Andy Yen cancelled the planned company-wide AI 're-interviews' on Friday, one day after 404 Media published internal messages from the push.
  • In the all-hands, Yen conceded 'a communication failure on my part' but said 'I'm still going to push' on AI adoption.
  • Staff told 404 Media the metrics the re-interviews would have captured will instead become a component of performance reviews.

Proton CEO Andy Yen cancelled a planned round of company-wide 're-interviews' meant to assess every employee's AI usage, one day after 404 Media published internal messages in which he dismissed security-conscious staff as engaging in 'loser thinking.'

In an all-hands on Friday, Yen told staff: "I want to acknowledge myself, personally, that there was a communication failure on my part." He paired the apology with a signal that the direction was not changing: "Going forward, I'm still going to push." On the cancellation itself he said the exercise had, "in a, let's say, unintentional way," already served its purpose. CTO Bart Butler was blunter: "We fucked up, we know we fucked up."

Staff reactions, relayed to 404 Media, were split. "Some are happy as it's progress," one employee said. "Others believe that this won't change anything because there isn't any consequence for Andy." Another told the outlet that leadership "said in the next breath they want to continue pushing AI and that the things that were to be measured in the re-interviews would become a component of performance reviews," meaning the scrutiny the re-interviews would have delivered gets folded into regular reviews instead.

The reversal lands inside a wider run of workforce-AI coverage we have been tracking, including Goldman's finding this week that AI-displaced workers wait a month longer to find new roles. The original Thursday report quoted Yen responding to staff raising security caution with: "That's bullshit and loser thinking. Of course it's possible to be fast and secure."

Shared on Bluesky by 5 AI experts