I said no and Apple said yes
Apple’s rollout of on-device “Apple Intelligence” and related AI processes in macOS has renewed fears that users can no longer meaningfully say “no” to major features, even when settings suggest they are disabled. Commenters highlight dark patterns such as nagging prompts, re-enabled options after updates, hidden background daemons and large, undeletable model files on small SSDs as signs of eroding user consent and growing value extraction. Many see this as part of a wider industry trend toward locked‑down, ad- and AI-driven platforms, prompting calls for regulation, stronger opt‑out controls, or a shift to Linux and other open systems for those who want real control over their machines.
Dark patterns, consent, and “no” that never sticks
- Many see Apple’s prompts (Siri, Apple Pay, Focus, trials, upgrades) as dark patterns: only “Yes” or “Maybe later,” never “No, never,” and nags that reappear.
- Several generalize this to the industry (Microsoft, Google, newsletters, cookie banners), describing a “ratchet of consent” that only tightens and never fully reverses.
- Others argue wording like “Not now” can reduce user anxiety if changes remain reversible, but critics counter PMs routinely repurpose that ambiguity to justify more nagging.
Apple Intelligence, storage, and telemetry behavior
- Big concern: macOS 27/“Apple Intelligence” installs ~20–22 GB of models on local disks (especially painful on 256 GB devices) and runs multiple AI‑related daemons even with Siri/AI toggles off.
- Some report options to disable Apple Intelligence were present in earlier betas and removed, or that updates silently re‑enabled Siri/AI and related “phone home” behavior.
- A specific “Apple Intelligence Report” setting is disputed: one side says it controls how often personal data is sent to Apple; another insists it only enables a local transparency log, not extra uploads.
- Apple docs are cited showing multiple per‑feature AI toggles; critics reply this is the opposite of a single global “off” and does not clearly reclaim disk space.
Privacy, value extraction, and shifting perception of Apple
- Many feel Apple has shifted from “user‑centric, privacy‑first” to “services and upsell”: iCloud storage nags, Apple Music/TV promos, OS badges, and now AI by default.
- Some still see Apple as relatively best‑in‑class on platform security and UX, arguing missteps may be temporary and not part of a foundational “hostility.”
- Others call this classic “enshittification”: once an ecosystem is dominant and sticky, the vendor can prioritize monetization over consent.
Alternatives: Linux/BSD/Asahi vs practicality
- A large sub‑thread debates fleeing to Linux/*BSD/Asahi Linux on Apple Silicon to regain control and avoid forced AI.
- Advocates emphasize ownership, no ads, and the ability to actually say “no,” but acknowledge rough edges: driver issues, upgrades that occasionally break, and higher maintenance.
- Several admit they stay on macOS (or Windows) because laptops are more reliable “just work” tools, especially for non‑technical users and corporate environments.
Regulation, AI attitudes, and broader consent worries
- Some call for regulation: explicit rights to refuse features at install/update time, with “refuse” as the default.
- There is strong anti‑AI sentiment in parts of the thread, focused less on model quality and more on consent, surveillance, and resource waste.
- Others argue AI is inevitable and useful if managed with better guardrails, and that personal boycotts alone won’t change industry behavior.