On April 15, The Information reported a somewhat ironic development: Apple is sending batches of Siri team engineers to an AI coding bootcamp, specifically to learn how to write code using AI tools.
The timing is two months before WWDC. Apple had planned to unveil a new, more powerful Siri at its June developer conference.
Now, they are cramming.
How Bad Is the Siri Team's Problem?
Consider the numbers: fewer than 200 engineers are attending the training, about 60 remain to continue Siri development, and another 60 are dedicated to evaluating Siri's performance. This configuration is leaner than many startups.
The background: The Siri team carries an unflattering label within Apple — the laggard in AI coding tools.
Other Apple engineering teams have long been purchasing Claude Code in bulk, with some allocating a significant portion of their engineering budgets to it. But the Siri team has been writing code using older methods.
This is not just a matter of tool preference. The Siri team failed to deliver several promised Apple Intelligence features for iOS 18, leading to a major management shakeup: AI director John Giannandrea left in late 2025, Craig Federighi (SVP of Software Engineering, overseeing iOS and macOS) took over AI direction, and Vision Pro driving force Mike Rockwell assumed control of the Siri team.
The new management's first order of business: go learn AI coding.
The Ironic Core of This Story
There is a paradox: engineers responsible for developing AI features do not themselves use AI.
You can view this as an organizational management issue — Siri engineers are skilled in traditional voice system development and need time to transition to the large model era. Or you can see it as a mirror: the penetration rate of AI coding tools inside large companies is far slower than it appears from the outside.
Not every software team uses Cursor or Claude Code. Some core teams at major companies still run workflows from a decade ago — not because they are unaware, but because no one has pushed for change, and there is no clear organizational pressure.
Apple is now using mandatory training to solve this problem. For a company known for "slow and steady wins the race," this pace is indeed urgent.
Siri's June Exam
Given Siri's current state, this training is not just about improving efficiency; it looks more like damage control.
Apple has signed a deal with Google Gemini, reportedly worth around $1 billion annually, to use Gemini models to power some of Siri's capabilities. Apple's own Foundation Models team is also expanding, betting on on-device inference and privacy protection. With both tracks running, the Siri users see is still the old assistant that often gives irrelevant answers and cannot handle multi-step tasks.
WWDC is in June. Apple needs to deliver a truly capable Siri to save face.
Sending engineers to learn Claude Code is, in a way, an admission: for progress that cannot be achieved by manpower alone, use AI tools to accelerate.
Can they catch up in two months? Users will tell you the answer directly.
Incidentally, the fact that Apple's Siri team learning Claude Code could not have appeared as news three years ago. That it can now shows that AI coding tools have moved from "optional" to "required" — even within the world's richest tech company, some teams are struggling to keep pace.
Sources: Siri Engineers Sent to AI Coding Bootcamp as Apple Prepares to Deliver Siri Overhaul (MacRumors); Report: Apple to send Siri engineers to multi-week AI coding bootcamp (9to5Mac); CocoLoop; Apple Sends Siri Staffers to Coding Bootcamp in Latest Shakeup (The Information)