Developer argues apps should favor on-device AI over cloud APIs to improve privacy, reliability, and user experience

The article criticizes the growing habit of adding AI features through cloud APIs from providers like OpenAI or Anthropic, arguing it makes apps fragile, costly, and privacy-invasive. The author says many tasks, such as on-device summaries, can run locally on modern phones using increasingly powerful chips and neural engines.

The piece reflects a broader push toward local AI as developers reassess the tradeoffs of cloud-dependent features. If more AI workloads move onto devices, apps could become more private, resilient, and cheaper to operate, while cloud providers may face pressure to justify when remote inference is truly necessary.