Proton goes all in on using AI
Proton’s decision to “go all in” on AI—paired with leaked internal messages from its CEO pushing aggressive adoption and dismissing security concerns—has shaken trust among privacy‑focused users and employees alike. Commenters worry that replacing engineers with AI‑driven workflows, discouraging cautious security practices, and fostering a snitch‑like internal culture is incompatible with Proton’s privacy brand, prompting many to weigh alternatives such as Tuta, Fastmail, Mullvad, or self‑hosting. Others argue that AI is an inevitable competitive tool, but criticize the leadership style and fear that “moving fast” with AI in a security‑critical product risks serious data‑governance failures.
Reaction to Proton’s AI Push and Internal Culture
- Many are disillusioned that a privacy-focused company is “going all in on AI,” especially after reading about internal manifestos, re‑interviews, and language like calling cautious staff “losers.”
- Several see this as evidence of a toxic, fear-driven culture: disrespectful to employees, encouraging snitching, and prioritizing speed and image over well‑being.
- A minority argue that aggressively pushing AI is rational in the current market and that some employees may be resisting necessary change, but even they criticize the CEO’s tone and execution.
- Some view the move to hire “product managers who build with AI” instead of engineers as an early signal of a wider industry trend to replace implementers with AI-using generalists.
Security, Privacy, and AI
- Commenters stress that for a company selling privacy and security, AI introduces serious data governance risks and “move fast” attitudes are dangerous.
- There is particular concern about dismissing requests for pentesting or caution as “loser thinking,” given how hard it is to be both fast and secure.
- A few contrarians claim AI-written code may end up more secure than human code, but others counter that current AI systems still fail badly and unpredictably.
Trust in Proton and Similar “Good Guys”
- ProtonVPN is praised as “borderline irreplaceable” due to its no‑logs track record in court; losing trust in Proton is seen as a major blow.
- Some say this does not automatically make Proton “bad guys,” but others point to a pattern: harsh internal behavior, politicized sponsorships, and prior controversies eroding trust.
- Some note that most tech companies are going heavily into AI; Proton is not unique, but the way it’s handled matters.
AI, Coding, and Jobs
- Debate over whether we still “need humans writing code”: some claim AI is ~90% there and will surpass humans soon; others say this has been over‑promised for years and the last 10% is the hardest.
- Historical parallels (self-driving cars, evolution from machine code to high-level languages) are used both to argue that tools change “what coding means” and to caution that “almost works” can remain stuck for decades.
- Some argue demand for software is effectively unbounded, so AI will increase total output rather than eliminate work; others worry about displacement and the quality of mass‑generated “slop.”
User Responses and Alternatives
- Numerous users say they will move away from Proton or are now actively planning to, constrained mainly by migration complexity and long‑held/lifetime accounts.
- Alternatives mentioned: Tuta for end‑to‑end encrypted mail/calendar, Fastmail for reliable email without strong E2E focus, Soverin, and self‑hosting as the only way to avoid abrupt strategic shifts.