What stood out is that people were not rejecting AI as a category so much as rejecting forced integration. The clearest pattern was familiar features getting replaced by AI-only flows or wrapped in AI prompts.
CarPlay was the main example. Several people said basic tasks like search or map entry now push them toward
Siri, while others pointed to Amazon review search, Google search summaries, Jira’s
Rovo buttons, Adobe Reader, and Office-style rewrite features as the same product move. The objection was less "AI exists" and more "I lost the old control surface and got a noisy shortcut to a cloud service instead."
A second theme was that this is not primarily a developer failure. It is a business decision. People called out a deliberate pattern of using popular features to drag adoption of weaker AI products, both to satisfy investor narratives and to defend incumbents from being turned into back-end systems for external assistants like ChatGPT or Claude. That framing also explains why opt-out often feels incomplete. Companies want the metric, the lock-in, or the future pricing lever, so fallback states stay underbuilt.
The practical advice that got traction was mundane and telling. Use
uBlock Origin or custom filters to hide AI panels. Add `udm=14` or CSS filters to suppress Google AI summaries. Swap browsers, office suites, or operating systems if the defaults keep regressing. A few people have already moved to Linux,
GrapheneOS,
LibreWolf,
VSCodium, or plain Bluetooth audio instead of accepting the bundled assistant path. The mood was annoyed but not helpless. People are assembling their own anti-AI stack because the mainstream products are no longer reliably preserving pre-AI behavior.