Plymorph

ETH Zurich researchers successfully demonstrated twenty-five security flaws across three major cloud password managers: Bitwarden, LastPass, and Dashlane.

The takes are sharply split on digital safety. A segment dismisses cloud managers entirely, with ThunderComplex calling them 'scams' due to ceded data control. Meanwhile, others defend these services out of practical necessity, citing the need to accommodate non-technical family members, as noted by DahGangalang and MalReynolds. On the LLM front, the discussion centers on risk: while markz sounds the alarm over the widespread, unvetted practice of using AI for passwords, Ephera counters that LLMs aren't doomed, stating they require explicit tool-calling, like calling `pwgen`, to achieve genuine randomness.

The core consensus points to extreme user caution. For maximum security, users are leaning toward decentralized, high-maintenance personal solutions like KeePassXC/Syncthing. The established vulnerability of major cloud players, combined with LLMs' predictive text bias, creates a clear mandate: users must bypass mainstream tools for genuine protection.