Apple’s AI and I

I’m not going to lie. Siri has been a huge disappointment. It’s so bad when I really need it that I can’t trust it at all. The only times I use it now are when I’m just too lazy to type and I don’t care if it fails or not. Using it to do something in the car, such as message my wife or even just launch Waze without being stupid about it is pretty much a waste of time at this point. I still try it from time to time, but generally after the third incorrect voice-to-text translation of what it is that I’m trying to say, I rage quit it and start racing around corners as fast as I can to let out the anger.12

Apparently I’m not the only one who’s noticed that Siri fails at both comprehension and contextual awareness, because the default narrative right now is that Apple is doomed to fail against companies like Google and Amazon that have virtual assistants backed by the power of search. Doomed, I say! Usually default Apple is doomed narratives are laughable at best, and throat punch worthy at worst, but in this case even some notable Apple diehards are getting in on the fretting.

Where Apple suffers is big-data services and AI, such as search, relevance, classification, and complex natural-language queries.1 Apple can do rudimentary versions of all of those, but their competitors — again, especially Google — are far ahead of them, and the gap is only widening.

And Apple is showing worryingly few signs of meaningful improvement or investment in these areas. Apple’s apparent inaction shows that they’re content with their services’ quality, management, performance, advancement, and talent acquisition and retention. And they may be right — they may be fine.

– Marco Arment, May 21, 2016

Marco Arment always takes heat anytime he opens his mouth, but most of that is just idiotic overreaction and the fact that he inevitably backtracks when people start yelling at him. If he’d just tell them all to shove it, he’d be fine. Some people just need to be told to shut up before they respect you, especially on the internet. In this case, I actually agree with Marco’s post, based upon Siri’s lack of visible progress and the seemingly small amount of effort Apple is putting into it. Siri was first unleashed upon the world in 2011. Five years later, it’s no longer funny that Siri often doesn’t have a clue what people are talking about and that it handles almost everything by performing a generic web search and dumping the results on the screen for the user to deal with. It really does appear to those on the outside (i.e, everyone) that Apple is not meaningfully advancing the technology of Siri.

Brian Roemmele disagreed with Marco a couple days ago on Medium, outlining a more positive take on the situation:

The “Apple is not ready for AI” narrative is quite uninformed and does not align with the facts. There is quite alot going on with the Siri team on what I am calling Siri2. Apple has also acquired rather important and vital AI startups. The results of these acquisitions will begin to surface in 2016.

He then goes on to list some of the acquisitions that he believes will bear fruit for Apple’s intelligent assistant, and predicts a coming smarter Siri, which he dubs Siri2.

Brian expands upon this theme in a separate Medium post, published yesterday:

It is not a secret that Siri has not kept up the pace that just about all of us expected, including some of the Siri team. The passion that Steve had seemed to have been waning deep inside of Apple and the results were Dag and Adam Cheyer moved on and formed Six Five Labs (A play on V IV in Roman numerals) and Viv.

Tom Gruber, one of the original team members and the chief scientist that created Siri technology, stayed on and continued his work. During most of 2016 and 2017 we will begin to see the results of this work. I call it Siri2 and am very certain Apple will call it something else.

Apple has always been a vital mix of internally created technology and acquired technology. From iTunes to TouchID Apple has been spectacular in identifying young and smart companies and integrating it into the very core of Apple.

Late in 2015 Apple approached a small Cambridge, England Voice AI company called VocalIQ and made a pitch to Blaise Thomson that he could not refuse. As a University of Cambridge spin out, VocalIQ had already been around for about 2 years and I had become very familiar with thier amazing technology. VocalIQ built astounding technology that no doubt you and I will use every day, some day soon.

The key to VocalIQ’s importance, according to Brian, is its ability to learn through interaction, something that is fundamental to the usefulness of any intelligent assistant, especially going forward in an increasingly competitive landscape. Also included in VocalIQ’s list of abilities is noise robustness, or the ability to accurately understand what a person is saying in less than pristine audio environments. This is something Siri flunks every time. Step one in communicating with Siri while driving, for example, is turning off the stereo and HVAC. If you don’t, you may as well assume “Hey, Siri” doesn’t even exist.

If Apple utilizes just a small subset of the technology developed by VocalIQ, we will see a far more advanced Siri. However I am quite certain the amazing work of Tom Gruber will also be utilized. Additionally the amazing technology from Emollient, Perception and a number of unannounced and future Apple acquistions will also become a big part of Apple’s AI future. I wrote about how the Voice First, Voice Commerce and Voice Payments world will play out here [1], [2], [3], [4]. As I have asserted in my 1989 Voice Manifesto, there will not be advertising in Voice First devices, there will be Voice Commerce and Voice Payments. The push mechanisms of advertising give way to Intelligent Agents pulling ontologies.

Apple has a mixed resume in terms of results with acquisitions. Not everything it buys turns to gold. On the other hand, Siri plays a big role in several of Apple’s key platforms, and is rumored to be expanding onto at least the Mac, and possibly onto an Echo competitor as well. If this is true, Apple needs to up their game with Siri exponentially. Giving them the benefit of the doubt and assuming they fully understand just how true this is, then they no doubt do intend to make good use of VocalIQ.

For Apple’s sake, and ours3, let’s hope that’s the case. I like Brian’s vision of Siri’s future better than mine.

  1. Not really. 
  2. Yes, really. 
  3. By which I am referring to Apple product users.