Apple reportedly sends 200 Siri engineers to AI bootcamp ahead of Gemini‑powered relaunch

A rapid, risky retraining push

It has been reported that Apple (苹果) plans to send roughly 200 engineers from its Siri team to a multi‑week AI programming bootcamp to teach systematic use of AI coding tools. Reportedly, after the training the core Siri development team will be trimmed to about 60 people, another roughly 60 will be moved into an evaluation group to monitor live Siri performance and safety, and the fate of the remainder is unknown. The push comes less than two months before Apple’s Worldwide Developers Conference (WWDC) where the company is expected to unveil a next‑generation Siri reportedly powered by Google (谷歌)’s Gemini model.

Background and recent reorganizations

Siri has long been viewed inside and outside Apple as a lagging product. It has been reported that internal bloat and infighting have hampered progress over the last 15 years, and the arrival of ChatGPT and other large language models has only widened the gap. Apple’s repeated reorganizations — most notably the March 20, 2025 shift that moved Siri out of John Giannandrea’s AI organization into Craig Federighi’s software engineering group with Mike Rockwell overseeing the product — were prompted by a delayed relaunch that failed to ship on its original timetable.

Tooling, vendors and technical friction

Parts of Apple’s engineering organization are said to have widely adopted external AI coding assistants — Anthropic’s Claude Code and OpenAI’s Codex among them — and have budgeted accordingly. The Siri team, by contrast, is reportedly behind in adopting these tools, which is one rationale for the rapid training. It has also been reported that Apple is still negotiating server‑hosting arrangements with Google for Gemini integration, underscoring how technical and commercial logistics remain unsettled even as a June WWDC reveal looms.

Can short‑term retraining buy long‑term competitiveness?

The immediate picture is clear: Apple is combining skills retraining, team slimming and a deep technical pivot under time pressure. But large model capabilities iterate far faster than traditional software cycles. Can a few weeks of bootcamp and a reshuffle turn a longstanding underperformer into a sustainable, AI‑native product? Or is this a one‑off remediation that will leave Apple exposed to faster‑moving rivals and broader geopolitical and data‑sovereignty tensions that now shape AI partnerships and cloud hosting decisions? Only ongoing structural change will tell.