It wasn't AI that flipped Proton's charge

A longtime Proton customer questions whether the company can still be recommended after reporting described CEO Andy Yen berating an engineer who requested several days to penetration-test new code before release. Yen reportedly called the request “loser thinking” and urged staff to stop carefully crafting code and rely more on AI. The central argument is that AI-generated code is not inherently the problem: it can be used responsibly when humans review it, test it, and take responsibility for what ships. The deeper concern is a workplace culture in which raising security concerns is treated as disloyal or inefficient. Staff were also reportedly told that their AI use could affect whether they kept their jobs; after that plan leaked and was abandoned, Yen said it had served its purpose.

Proton’s business depends on users trusting it to protect their privacy, so speed cannot be prioritized at the expense of careful security work. The writer argues that a release exposing users’ mail could destroy that trust, making a few extra days of testing a meaningful safeguard rather than needless delay. An employee’s public defense - that Yen supported penetration testing so long as it did not slow releases - is seen as reinforcing the problem: security is acceptable only when it does not impede delivery. The writer also views supportive Reddit comments from employees as clumsy crisis management, noting that none challenged the company’s direction or response. Proton’s leadership, the argument concludes, now seems to regard the caution that once justified users’ confidence as an obstacle.