If you want to understand how far behind Apple is on AI, consider this: the company is sending hundreds of its own Siri engineers to a multi-week coding bootcamp to learn how to use AI tools. Not to build them. To use them.

The Information reported Wednesday that Apple plans to pull a significant chunk of its Siri team, a group numbering in the hundreds, off active development and into what amounts to an emergency crash course in AI-assisted coding. The timing is remarkable: WWDC 2026 is roughly two months away, and Apple is expected to unveil a dramatically overhauled Siri. Instead of building that product, a huge portion of the team is learning the tools that OpenAI and Anthropic engineers have been living in for years.

That leaves about 60 people on the core Siri development team and another 60 focused on testing and evaluation. For context, Anthropic has more people working on a single model release.

The Bigger Problem Nobody Is Saying Out Loud

This is not a story about retraining. This is a story about a company that spent fifteen years building a voice assistant with intent-based heuristics and now needs to rip it out and replace it with LLM architecture. The engineers who built the old Siri are not the engineers who know how to build the new one. Apple is betting it can bridge that gap with a multi-week bootcamp.

Here is what makes this so painful: Apple already struck a deal with Google to license Gemini for Siri integration. It already admitted it blew a five-year lead on AI, according to former insiders who spoke to CNBC. It already hired Google’s former AI marketing chief. And now it is sending its own engineers to learn the basics of the tools their competitors mastered years ago.

The bootcamp is reportedly focused on vibe coding techniques using AI assistants like GitHub Copilot and Claude Code. The goal is to make Siri engineers dramatically faster at writing the code needed for the LLM-powered overhaul. But speed is not the problem. Direction is the problem. Apple still does not have a competitive foundation model. It is still dependent on Google for its core AI capability.

The Competition Is Not Waiting

While Apple retrains its workforce, Anthropic is about to drop Opus 4.7 and an AI design tool that has already crashed Adobe and Figma stocks before it even launched. OpenAI just released a cybersecurity model. Google put Gemini inside Boston Dynamics robots. The AI race is not slowing down for Apple to catch up.

The most telling detail: Apple is pulling people OFF active Siri development to attend this bootcamp. That means fewer people building the product that is supposed to save Apple’s AI credibility, so that those same people can eventually build it faster later. It is an investment that makes sense over twelve months. The problem is that Apple does not have twelve months. It has two.

When the richest company on Earth has to send its engineers to bootcamp to learn the tools that twenty-person startups already use every day, you are not watching a retraining program. You are watching an admission that the old way of building software is dead, and Apple knows it.