Apple clamps down on AI apps
Apple has removed the first 'vibe‑coding' app from the App Store and is tightening policy enforcement around AI‑driven coding and agentic apps, a move critics say conflicts with democratizing developer tools. At the same time, reporting indicates Apple is shaping an AI strategy that emphasizes deeper App Store integration and LLM‑driven search experiences. (techzine.eu) (jagranjosh.com)
Apple pulled the vibe‑coding app Anything from the App Store on March 26 after blocking updates to the app since December, a timeline confirmed by Anything co‑founder Dhruv Amin and App Store correspondence. (macrumors.com) Apple told developers the removal centered on App Store Guideline 2.5.2 — the long‑standing rule that bars apps from downloading, installing or executing code that changes app functionality after review — and rejected an update that would have shifted live previews into an external browser. (macrumors.com) Anything raised $11 million and was valued at roughly $100 million in a September round, and the team said the tool had been used to publish thousands of apps to the App Store before removal. (macrumors.com) Apple has applied the same enforcement to other vibe‑coding tooling — blocking updates for Vibecode and Replit earlier this quarter — signaling a platform‑wide application of the self‑containment policy rather than a one‑off takedown. (forbes.com) Internally, Apple is simultaneously pursuing a search‑centric AI strategy: Bloomberg reports an internal project dubbed “World Knowledge Answers” that would use LLMs to deliver an AI “answer engine” inside Siri, Spotlight and Safari with a planned spring rollout. (bloomberg.com) Apple’s machine‑learning research notes a 2025 Foundation Models suite that includes a compact on‑device model of roughly 3 billion parameters and a mixture‑of‑experts server model, plus a Foundation Models framework that gives developers programmatic access to on‑device LLMs optimized for Apple silicon. (machinelearning.apple.com) Apple has already folded LLMs into App Store features such as automated review summarization and is piloting AI‑driven tagging and discoverability changes, underscoring an approach that couples stricter code‑execution enforcement with deeper App Store and search integration. (winbuzzer.com)
