Proton Users Revolt After CEO Calls Security Caution 'Loser Thinking' Amid AI Push

What it claims

Decrypt reports that the pentesting dispute began in August 2026. A Proton employee asked for "a few days" to have specialists try to break new code before release. CEO Andy Yen replied: "That's bullshit and loser thinking. Of course it's possible to be fast and secure." Yen also told staff to stop writing code by hand and said he would re-interview employees about their AI use. Staff feared those answers could decide whether they kept their jobs. Yen cancelled the re-interviews on 30 October 2026.

CTO Bart Butler's 1 November 2026 manifesto goes further than the summary's "direct LLMs". It says: "Your job is to help the LLM write better code faster." It also says "'Programming' is dead. Engineering is not." It frames avoiding AI as an "existential threat" to Proton's mission. It calls for company-wide automation of approvals and other development steps, on the reasoning that faster code generation achieves little if those steps stay slow. Butler describes that overhaul as the next phase of adoption. The manifesto also states that expertise and testing remain essential, and that engineers stay accountable for work models cannot reliably handle.

Evidence

The 404 Media investigation drew on nine current and former employees, internal messages and meeting recordings. On Proton's feedback forum, a proposal calling for Yen to step down had 2,020 votes when Decrypt reviewed it. Proton told 404 Media that management had sometimes communicated too bluntly and that employees were anxious. It did not retract the AI push.

Limits

The manifesto commits to ending most human code review only "eventually", and the article gives no timeline. It also does not say which release gates would be automated, or whether pentesting stays in the process. The forum vote count measures sentiment, not churn. Decrypt notes that cancelling a paid plan does not mean deleting an account. Claims of a toxic culture come from Reddit and social media posts by current and former employees and are unverified. Proton's Lumo chatbot has protections against training on user conversations, but the article does not connect those protections to the internal coding changes.