Wiki › People & organisations

Proton

Swiss privacy company (Proton Mail, Lumo) whose 2026 AI push combined AI-code tracking, re-interviews, a "your old job is gone" CTO manifesto and a CEO who called pentesting "loser thinking".

Proton is a Swiss privacy company. It launched Proton Mail in 2014, after the Snowden revelations, and has since added an encrypted drive, a password manager, an authenticator and Lumo, an LLM chatbot it says neither logs conversations nor trains on user data (Lumo 2.0 shipped in June 2026). The CEO is Andy Yen, a former CERN particle physicist. The CTO is Bart Butler.¹

The 2026 AI push

In October 2026, 404 Media reported on Proton's internal turn to AI. The story draws on eight current employees and one former employee, plus Slack chats, meeting recordings and an internal manifesto. Proton's spokesperson did not dispute any of it.¹

• Timeline. Proton first banned AI internally. It later approved AI for code, financial data and customer-facing replies on Reddit and LinkedIn, with exceptions for competitively sensitive code and R&D. From late August 2026, Yen pushed for maximum AI use as fast as possible.
• Scale. Butler said in a meeting that 90% of Proton's code in October was written by AI, much of it with Anthropic's Claude.
• Mechanisms. Proton set up an email address for reporting anything that blocks "AI velocity", installed software on workers' computers to track how much AI code they generate, and announced that staff would be re-interviewed about how they use AI. Employees read the re-interviews as a threat to their jobs.
• Rhetoric. Yen called a request to pentest new code before release "bullshit and loser thinking". He said 2026 would be the last year Proton hires engineers and that it would hire people who build with AI instead. Butler's manifesto told engineers "your old job is gone" and set the goal that most code be reviewed by AI rather than people.
• Business motive. An internal document says Proton will miss its annual recurring revenue target "due to insufficient velocity" and calls it "an underperforming company". Yen denied that the push was financially driven.

Employees describe a culture of fear in which security concerns go unspoken. The case is the main example on AI Adoption Mandates. It also tests the review-free goal of Agentic Software Factories at a company whose product is trust.