Reimagining the mouse pointer for the AI era
Google’s proposal to make the mouse pointer an AI-aware control surface—where you can point at something on screen and say “do this” to an assistant—elicits both curiosity and alarm. Commenters question whether voice + pointer is actually faster or clearer than existing keyboard/mouse workflows, worry about continuous screen capture turning into another vector for surveillance and data mining, and doubt its practicality in offices where people don’t want to talk to their computers. A minority sees promise for accessibility, non‑expert users, and design tools if such agents run locally and respect user control, but most remain skeptical that this will meaningfully improve everyday computing.
Privacy, Surveillance, and Trust
- Many see this as “Google Recall for the pointer”: effectively continuous screen monitoring, not a smarter cursor.
- Strong concern that sensitive activities (medical info, protests, personal planning) would be captured and become accessible to Google, advertisers, law enforcement, or via legal discovery.
- Some think Google will start with “only when invoked” but expect product pressure toward predictive, always-on capture.
- A minority say this could be acceptable if models run fully on-device, never phoning home; Google is widely viewed as poorly positioned on trust compared to a local-first OS.
Actual Benefits vs Gimmickry
- Multiple commenters say the demo tasks (copying text, changing a number, moving an image) are faster with existing mouse/keyboard or app UIs.
- Some argue this simply reimplements context menus/right-click with voice, adding latency, cost, and failure modes.
- Skeptics see it as hype-driven “slopfeature” designed more for PM promotion and data collection than user value.
- Others note potential for simplifying complex or hidden UIs, especially for non-technical users who struggle with copy/paste, filters, or search.
Voice Interaction and Social Context
- Heavy pushback on voice as a primary input: awkward in offices, coffee shops, shared spaces; many don’t want to “talk to their computers.”
- Some power users and remote workers report voice/dictation is already faster and more natural for them, especially with AI tools.
- Accessibility is mentioned as a real upside (injuries, disabilities, radiology-style dictation workflows), but not a general replacement for mouse/keyboard.
- Interest in alternatives like subvocal or visual speech recognition to avoid disturbing others.
Interaction Design and Usability
- Concerns about gestures like wiggles and zigzags being ambiguous, easy to trigger accidentally, and ergonomically poor, especially horizontal drags.
- Fear of losing precise control over selection in favor of fuzzy “AI guessing what you meant.”
- Some note that pointing is indeed superior to verbal description in dense visual contexts; combining pointing + language is seen as conceptually strong.
Broader Context and Alternatives
- Several note this idea echoes older research (“Put That There”, bubble cursor, agent UIs) rather than true reinvention.
- Some want the underlying protocols/APIs so others can build less intrusive agents.
- Overall sentiment: interesting research direction with real potential in narrow or future contexts, but current Google implementation feels premature, mis-scoped, and privacy-hostile.