AI guru Tom Gruber, last of the Siri founders, retires from Apple
Tom Gruber, an artificial intelligence expert who cofounded Siri, Inc., prior to its purchase by Apple, has announced his retirement from the company, becoming the last of Siri's three cofounders to leave Apple.
According to a report Wednesday night by The Information that was confirmed by Apple, Gruber has announced his retirement, and will now devote his energies to pursuing "personal interests in photography and ocean conservation."
The site reported that Vipul Ved Prakash, Apple's head of search, is also out.
Gruber most recently had been working as head of Siri's Advanced Development group; new hire John Giannandrea recently assumed leadership of the overall Siri team.
A computer scientist who spent the early part of his career in academia, Gruber did a stint at Stanford in the 80s and 90s, before entering the private sector as Chief Technology Officer of enterprise software firm Intraspect Software.
In 2007, Gruber became the cofounder of Siri, Inc., which was acquired by Apple three years later, with the Siri feature soon incorporated into the iPhone and later other Apple products.
In a TED conference talk in 2017 called "Our Robotic Overloads," Gruber predicted that while some fear artificial intelligence replacing humanity, it's better for the technology to augment humanity.
"What if you could have a memory that was as good as computer memory and is about your life? What if you could remember every person you ever met," Gruber said. "How to pronounce their name?" Their family details? Their favorite sports? The last conversation you had with them?"
The same website, The Information, published an overview in March of the Siri team's long history of acrimony and infighting, which has been linked to Siri falling behind Apple and Google's counterparts.
Siri's other cofounders were Dag Kittlaus and Adam Cheyer. Kittlaus left Apple in 2011, while Cheyer exited in 2012. Kittlaus objected on Twitter back in March to statements made in The Information's piece about Siri.
According to a report Wednesday night by The Information that was confirmed by Apple, Gruber has announced his retirement, and will now devote his energies to pursuing "personal interests in photography and ocean conservation."
The site reported that Vipul Ved Prakash, Apple's head of search, is also out.
Gruber most recently had been working as head of Siri's Advanced Development group; new hire John Giannandrea recently assumed leadership of the overall Siri team.
A computer scientist who spent the early part of his career in academia, Gruber did a stint at Stanford in the 80s and 90s, before entering the private sector as Chief Technology Officer of enterprise software firm Intraspect Software.
In 2007, Gruber became the cofounder of Siri, Inc., which was acquired by Apple three years later, with the Siri feature soon incorporated into the iPhone and later other Apple products.
In a TED conference talk in 2017 called "Our Robotic Overloads," Gruber predicted that while some fear artificial intelligence replacing humanity, it's better for the technology to augment humanity.
"What if you could have a memory that was as good as computer memory and is about your life? What if you could remember every person you ever met," Gruber said. "How to pronounce their name?" Their family details? Their favorite sports? The last conversation you had with them?"
The same website, The Information, published an overview in March of the Siri team's long history of acrimony and infighting, which has been linked to Siri falling behind Apple and Google's counterparts.
Siri's other cofounders were Dag Kittlaus and Adam Cheyer. Kittlaus left Apple in 2011, while Cheyer exited in 2012. Kittlaus objected on Twitter back in March to statements made in The Information's piece about Siri.
Comments
In real-world, day-to-day use, it is my contention that Siri (especially on iOS) is competitive with the other voice assistants (though I acknowledge Google's is further along/better in many areas) . This is actually impressive when you remember that Siri is the only one not mining your entire life for sellable data about you (but this also means it is doomed to be third out of three forever, essentially).
I think Giannandrea and his team will nonetheless be able to make Siri even better despite the privacy restrictions. There are plenty of resources to help Siri be more predictive and learn without compromising security and privacy, and I'm confident he will find them.
Even more importantly is that Apple is one of the very few ethical companies that has sufficient clout to make a difference in the world.. That alone takes any regret away that the price tag of their technology would induce..
As to Siri, I know in time it will be able to do the same things as Google.. Now that the AI team is reorganized and the infighting stopped hopefully, I think it will be sooner rather than later and who knows maybe Siri will leap ahead.. Also, imagine what an advanced AI can do with full 3D perception of the world, instead of being restricted to clever 2D pixel analysis and manipulations..
The other area is search related as well - Siri is just not as good at providing information as Google Assistant. Perhaps not surprising. However, for me, I don't use Siri for this much (as I generally want to "see" information, sort through many options at once, not get it via voice, unless it is a very simple answer).
Yes, Siri still has voice recognition issues - and this is the most legitimate criticism of the service. It seems like there could be some relatively simple fixes here, as well as an opportunity for new h/w in new models:
- You can setup "Hey Siri" to only respond to your voice, and not someone else around you. But Apple have never expanded this to help Siri learn more about how you say things.
- Is it really that difficult to have a setup (that you can do, or update, at anytime), where you say perhaps 10 (set) phases to Siri, with results stored on device, that could greatly improve the experience. Perhaps pronounce the names of your family, or top 10 contacts, so that it won't mistake "Owen" for "oh when".
- Apple are the top dogs of mobile silicon, and the A11 had the neural engine, so hopefully a new model (soon) can have some dedicated signal processing, ML, and voice h/w that would make Siri's voice recognition service highly accurate. Perhaps it needs 2 mic's for this as well.
- With the advance of silicon, DSP, and mic technology, surely Apple should be the first to have "on-device" voice recognition, with Siri able to perform local tasks without cloud access (set a time, turn x-y-z on/off, raise volume, etc).
”Hey Honey”, “Yo, bitch!”, “HAL!” etc.
people should be able to teach the beast their personalized attention phrase.