Apple and a Hacker's Future
3 hours ago
- The author's always-on Mac Mini was hacked via a macOS screen sharing vulnerability (CVE-2026-65400) exploited when port 5900 was exposed, allowing root access and a Monero crypto miner.
- The author's AI agent (Claude) detected the breach, diagnosed it, and helped remove the malware, demonstrating the value of persistent agent monitoring.
- Apple's Transparency, Consent, and Control (TCC) permission system is problematic for headless Macs running agents because prompts are invisible to software, causing silent failures and requiring manual intervention via screen sharing.
- Apple's security update labeling is misleading: automatic security update settings do not cover point releases, leaving systems vulnerable despite the user's intent to stay protected.
- The author is frustrated with Apple's restrictive approach to agents and full disk access, preferring open systems like Linux for server use and envisioning a future without Apple by default.
- The author is unenthusiastic about Apple's smart home push, preferring to build their own agent-controlled systems rather than rely on Apple's constrained ecosystem.
- Agents make anyone a hacker, and Apple's walled garden feels like a prison, leading the author to consider alternative platforms.