How The Wall Street Journal Fell Behind in the ‘Apple Is Behind on AI’ Arms Race

Aaron Tilley, writing for The Wall Street Journal under the headline “How Apple Fell Behind in the AI Arms Race” (News+ link):

For those who saw them, the demonstrations inside Apple earlier this decade of a revamped Siri offered a showcase of the amazing capabilities a powerful AI voice assistant could have.

The famed assistant, one of the last projects Apple co-founder Steve Jobs worked on before his death, had been given a total overhaul. Capable of running on an iPhone and without an internet connection, the new Siri impressed people with its improved speed, conversational capabilities and the accuracy with which it understood user commands. Code-named Project Blackbird, the effort also imagined a Siri with capabilities built by third-party app developers, according to people familiar with the work.

Yet a competing project won out in an internal contest ahead of the 10-year anniversary of Siri’s launch. Known as Siri X, the more-modest upgrade involved moving more existing Siri software onto iPhones from remote servers to improve the voice assistant’s speed and privacy. The Siri X enhancement was unveiled in 2021.

Tilley is the WSJ’s Apple beat reporter, and one gets the feeling he was tasked with filing a report with the above headline already written. These opening three paragraphs are the only interesting ones in the entire story. But there’s nothing actually new in them.

Here’s Wayne Ma, reporting for The Information in April 2023 under the headline “Apple’s AI Chief Struggles With Turf Wars as New Era Begins” (archive):

In some cases, Giannandrea’s new hires have run into seemingly impenetrable bureaucratic obstacles. In one example, he in 2019 recruited another close friend, Arthur van Hoff, to explore a project to rewrite Siri from the ground up. Code-named Blackbird, the effort involved creating a lightweight version of Siri, which would delegate the creation of more functions to app developers, said five former Siri employees. The software would run on iPhones instead of in the cloud, which would improve Siri’s speed and performance while enhancing user privacy, they said. Demos of Blackbird prompted excitement among employees on the Siri team because of its responsiveness, they added.

But there was a problem. Blackbird competed with the work of two longtime senior Siri leaders: Alex Acero and Robby Walker, who were responsible for two important teams that helped Siri understand and respond to queries. Acero and Walker pushed for completion of their own project, code-named Siri X, for the 10th anniversary of the voice assistant, which aimed to move the Siri processing software onto the device for privacy reasons.

However, Siri X’s goal was simply to reproduce Siri’s existing capabilities without the more ambitious targets of Blackbird, the people said. Despite that, Acero and Walker won. They assigned hundreds of people to their effort, which subsumed and killed Blackbird. Most of their project was completed in 2021.

Same story, but Ma’s version from 13 months ago included the names of the engineers in charge of the dueling projects.

Back to Tilley’s report at the WSJ yesterday:

Apple has long prided itself on perfection in its product rollouts, a near impossibility with emerging AI models. While OpenAI systems have dazzled more than 180 million users with their generation of writing, images and video, they are prone to occasional errors, often called hallucinations. Apple has had limited tolerance for such issues.

“There’s no such thing as 100% accuracy with AI, that’s the fundamental reality,” said Pedro Domingos, a professor emeritus of computer science and engineering at the University of Washington. “Apple is not compatible with that. They won’t release something until it’s perfect.”

How does this square with the state of Siri as it works today? Does Tilley think today’s Siri, though limited in scope, is “100 percent accurate” and “perfect”?

Don’t know about you, but that’s not my experience.