Project Zero will disclose vulnerabilities 90 days after reporting

Google’s Project Zero is tightening its vulnerability disclosure timelines, moving from 90 days after reporting to vendors down to 30 days by 2028, in response to how quickly AI tools can now uncover exploitable bugs. Commenters debate whether shorter deadlines will genuinely improve security or simply punish smaller vendors and volunteer-run open source projects that struggle to patch and test quickly. The policy also raises questions about power imbalances, potential “bullying” of dependents like ffmpeg, and whether large companies that benefit from these fixes should help fund remediation rather than just expose flaws on a fixed schedule.

Reaction to the 90 → 60 → 30-Day Policy

  • Some find the move to 30 days “needlessly harsh,” especially for inherently hard-to-fix bugs or complex ecosystems (Android, distros like Ubuntu).
  • Others argue shorter deadlines are necessary because attackers can now find and weaponize bugs faster, especially with AI.
  • There is support for clear, fixed deadlines; without them, vendors may delay indefinitely.

AI and the Changing Vulnerability Landscape

  • Many comments see LLMs as shrinking the time between discovery and exploitation: if Project Zero can find it, so can attackers’ tools.
  • Disagreement over whether AI implies “same-day remediation” or even “zero-day disclosure”; critics note LLMs are error-prone, non-deterministic, and expensive to run thoroughly.
  • Some imagine AI-driven loops where AI finds, fixes, and validates patches, marginalizing human developers; others insist human understanding and testing remain essential.

Disclosure Timing, Fixing, and Testing

  • One view: disclosure should be soon after a fix/workaround exists, with a maximum window to pressure vendors.
  • Another: even if a fix is hard or slow, users should know so they can avoid or isolate vulnerable software.
  • Multiple commenters stress that patch rollout, canaries, and testing (citing major outages) are often harder than coding the fix.

Effects on Vendors, Open Source, and Devices

  • Concern that 30 days is unrealistic for small companies, hobby projects, or embedded devices that barely get updates.
  • Some argue such software shouldn’t be on the public internet; if it can’t be updated, it should be isolated or effectively “smashed.”
  • Others worry this expectation will harm open source and low-income users who rely on cheaper, poorly maintained devices.

Power, Incentives, and “Extortion” Debate

  • Debate over whether time-limited private reporting plus guaranteed disclosure is “extortion” or just transparent scheduling.
  • Critics note Google benefits financially from fixes in widely used OSS (e.g., media libraries) without necessarily funding them, calling this a form of coercive pressure.
  • Defenders respond that vulnerabilities exist regardless; public disclosure after a set time lets users protect themselves and is not a demand for money or control.

Regulation, Liability, and Market Effects

  • Some expect stricter timelines to push markets toward vendors capable of fast patching and to encourage regulation (e.g., EU cyber rules, minimum maintenance, right to repair).
  • Others doubt users will “vote with their wallets,” especially at the low end, and see government intervention as the only realistic lever.