Siri co-creator Adam Cheyer says generative AI will spur an assistant voice renaissance; Satya Nadella says Cortana, Alexa, Siri, Google Assistant “don't work”
Teams working on the ecommerce giant's voice assistant have been hit hard by the largest cuts in company's history
Context & Ripple Effects
This debate lands at the exact moment the assistant era cracks. Siri co-creator Adam Cheyer bets generative AI will revive the category he helped create, while Satya Nadella declares the four flagship assistants simply don't work. Both positions have long roots: former Apple employees traced Siri's stunted progress to unambitious goals and shifting strategies, and Microsoft had already pulled Cortana back from consumer markets to focus on workplace conversations.
What changed is the money: the description notes Alexa teams were among the hardest hit in the ecommerce giant's largest-ever layoffs, and days later the New York Times documented how all three assistants fell behind despite their decade head start. The reset has since compounded — Google cut hundreds of Assistant staff to restructure around newer AI, and Apple executives privately concluded personalized Siri may need a from-scratch rebuild.
First-order effects
- Amazon's Alexa organization absorbs deep cuts in the biggest layoff round in company history, shrinking the team behind the most widely deployed consumer assistant precisely as its technical foundation is publicly dismissed.
- Nadella's "don't work" verdict completes Microsoft's exit from the consumer-assistant race it began when Cortana was refocused on workplace interactions, while Cheyer reframes the category's failure as architectural rather than permanent.
Second-order effects
- Google follows the same playbook, laying off hundreds of Assistant employees on the stated logic that restructuring will free the product to integrate newer AI — trading incremental feature work for a generative rebuild.
- Pressure lands hardest on Apple, where Federighi and other executives voiced doubts that personalized Siri functioned properly and some argued it must be rebuilt from scratch — turning the incumbent's installed base from an asset into legacy-code liability.
Third-order effects
- If Cheyer's renaissance thesis holds, the assistant layer gets rebuilt on generative foundations rather than patched, erasing a decade of accumulated command vocabularies and device integrations that defined how users learned these products.
- Competitive advantage migrates from whoever controls distribution on phones and speakers to whoever owns the underlying model — a structural inversion of how Siri, Alexa, and Assistant originally won their positions.
The trend: Voice assistants are entering a generative-AI rebuild cycle in which the first movers' decade head start converts into a liability rather than a moat.