Photo by Omid Armin on Unsplash

Hey, Siri, I want a taco.

Thoughts from the “Apple Intelligence” WWDC presentation.

Bret Waters
2 min readJun 11, 2024

--

This morning, Apple held its big annual WWDC presentation in Cupertino, and, of course, it was all about AI. Generative AI has been consuming all the oxygen in Silicon Valley for the past 18 months, so Apple needed to wave that flag today. And wave it they did.

They announced new AI functionality across all the various Apple platforms. To me, the big deal is that most of it runs on Apple’s own device silicon (not in the cloud), which allows them to tell a privacy story that they hope makes AI seem a lot less scary to consumers.

But as I watched Craig Federighi’s demo this morning of what Siri can do now with this new enhanced Apple AI, I thought about how Siri finally looks like it’s living up to its potential — and what a long, strange trip that’s been.

Siri began life in 2002 as a project within the artificial intelligence center at SRI International. The project was called CALO — “Cognitive Assistant that Learns and Organizes,” and the idea was to create a personal assistant that, well, learns and organizes.

Meanwhile, a completely different group at SRI was working on a speech recognition engine with natural language processing (later spun-off as Nuance Communications). Someone at SRI said “Hey, maybe we could put this voice recognition technology on top of this personal assistant software and see what happens!”.

Eventually the venture group at SRI got interested (shout-out to my friend Alan Herzig) and in 2007 they spun-off a new startup named “Siri, Inc”. The founding CEO was Dag Kittlaus, and “Siri” was a Norwegian name he liked.

In 2010, Dag Kittlaus got a call from an “unknown number.” He hesitated, deciding whether to answer. It’s a good thing he picked up because it was Steve Jobs calling, and he wanted to buy the company. A deal was done, and the technology was incorporated into the iPhone.

So, although Siri was shining bright on the stage in Cupertino today, it took more than two decades of twists and turns for it to get there.

Here’s the main takeaway: truly disruptive innovations take a long time to finally get mass market adoption. The term “Artificial Intelligence” was coined by Stanford professor John McCarthy way back in 1955, and Large Language Models go back to the 1980’s. So while AI has been dominating the 2024 headlines, it didn’t all just happen overnight.

As I said last week, software has been getting smarter and smarter for 40 years. More great things ahead.

--

--

Bret Waters

Silicon Valley guy. Teaches at Stanford. Eats fish tacos.