That's one part of the piece that I disagreed with as well. That said, it's true that Apple is not a AI-first/data-first company. As a result, they have less data and data culture inside the company. While these acquisitions may change their culture incrementally, it's unlikely that their company DNA catches up to Google/Facebook's.
Culture is a powerful force in shaping products. For better and worse, Google/Facebook have the "let's use customer data to build products" DNA ingrained into their company DNA. Of course, this has pros and cons. At times, their thirst for customer data can be creepy/controversial while Apple's steadfast support for customer data privacy is assuring.
Ultimately, it comes down to this: How willing are we to exchange our own behavior and demographic data for convenience? If the answer is yes, then the future is as Marco predicts it.
Google has many things going on for it other than search (and you dont have to go for dinosaurs, basic search in some Google properties, like YouTube, is quite terrible), that can count as killer features. Many people use it to organize their lives (calendar) or as comminication platform (hangout, email...) not to mention Google also has its own version of HomeKit as well.
My first insctinct when reading this article, though, is to question wether we put too much faith in the success of any service launched by Google and downplay any other competitor. Weirdly absent from Marco's list is IBM's Watson, quite a success story from a company that does not collect personal information on a daily basis yet somehow has been able to advance significally in the field of AI and Machine Learning.
Aside from that, maybe its me but I don't feel like Siri is so far behind serevices like Alexa or Cortana and until Google Assistant shows up in a more defined state, I wouldnt dare to put it ahead of the pack as well. There are a couple of things Alexa does better, but it mostly has to do with being open to third party services (which im assuming Siri will have soon) and better mic technology on the Echo, a hardware problem.
>But if Google’s right, there’s no quick fix. It won’t be enough to buy Siri’s creators again or partner with Yelp for another few years. If Apple needs strong AI and big-data services in the next decade to remain competitive, they need to have already been developing that talent and those assets, in-house, extensively, for years.
This is so true. It's not easy to out-accrue data over your competitors, especially if your competitor has multiple services used by millions of people everyday. This has an interesting side effect when it comes to hiring: The companies with "best" data have a better chance of hiring the best data talent because they can provide most challenging problems (for systems and data engineering people) and better equipment (There's no substitute for a vast amount of high-quality data for data scientists). This, in turn, means that best algorithms will come from data-rich companies.
What Marco is hinting at is that smartphones will increasingly become a conduit for data-intensive algorithms. If this is indeed the case, it means that data-rich companies can create deeper engagement and hence acquire more data, and the virtuous (or vicious?) loop will continue.
The question Marco was posing is different: What if hardware mattered less to the point that consumers will switch over to different hardware with better/1st class software support. If that happens, the future doesn't look too bright for Apple (or so posits Marco).
Apple might as well continue to be a first-class hardware-first business. However, if their hardware advantage is eclipsed by their data/AI disadvantage, Apple might lose a big chunk of their market share.
AI for consumers is not a logic problem, it's a data problem. Whether strong or weak, consumer AI needs to be personal. It needs access to personal data.
Certainly better logic will create the foundations of whatever system is developed, but this only needs to be solved once. Smart people will solve it, and Apple could very well buy this answer in an acquisition if no one beats them to it.
But the main problem is still with feeding the system food for thought, so to speak. Data is the fuel. Apple barely has any data.
So in this sense, Apple could never win because it doesn't have a fuel source. And consumers will have a hard time allowing Facebook or Google to sell their data to Apple. For even now, the whole premise of this data collection is that their privacy is still being protected and is mostly for ad targeting. But this fuel is their bread and butter. It's no coincidence that ads are the main revenue for both companies. And only they are in a position to leverage their data for personal AI.
Ultimately though, neither Google nor Facebook have all the important stuff, and their data is always a product of their other services, which will always make it partial. For example, Amazon may have data on your shopping habits on Amazon, but won't have data for other things.
What is needed is a personal data mining system. This system needs to be owned, and not sponsored. It needs to be open source, so there are no ulterior secrets. And it needs to be able to gather fuel from Google and Facebook and everyone you give data to. The future will be more about us tracking ourselves than corporations tracking us in secret.
Maybe call the system "Introspection". It will happen one way or another, because there is immense value in such a system. And it can happen now, because it's about the data. Logic can be updated when better systems are developed. The data just needs to be collected, and already, there are enough tools to make data about ourselves insightful.
Oddly enough I find Apple's approach to AI more to my liking than Googles. When I do a voice query on Google and all of a sudden every ad on the device is related to what I just did, I feel abused.
I think the confusion here is the difference between data mining and AI. Google is great at data mining and less convincing when it comes to actual learning (Look at Alpha GO, it is an algorithmic exploit of a game, not a breakthrough).
I've been thinking recently that Google is the one with the losing approach. Personally, I have been using Bing a lot more because Google has been pushing the good results to page 2-3. Whatever their big AI is doing seems, at least for me, to not be working.
The "Breakthrough" for AlphaGO was that Go is well mapped by a Monte Carlo tree search. Once they figured out it became a domain specific problem and totally irrelevant as a general AI breakthrough.
The article makes a great point about Apple's susceptibility to being overthrown in the phone market. The watch hasn't really done well and iTunes DRM theft of people's music libraries is becoming a thing. It's ok if Apple dies like Blackberry. They've died before. This time, though, there's no Jobs to resurrect them. They might become the IBM of their generation of companies, still around, still hoarding cash, but not very relevant.
There is no conceivable metric that shows the Apple watch as a success. Apple, which routinely leaks units sold to business media, has not provided numbers for the watch. According to analysts sales dropped 90% a month after launch and 66% of watches sold were the low end model. Apple is a remarkably successful hardware company but (so far) their watch has fallen flat.
Also, I think you're underestimating how heavily machine learning plays into the services that you use every day (e.g.: spam filters, web search, movie recommendations, autocorrect on your phone,...). The best AI services are the ones that you don't even think about using.
I think he's overstating the possible negative effects of this shift for Apple.
Facebook and Google especially wouldn't dream of excluding Apple devices (read, very lucrative Apple customers) from their ecosystems. Apple make their money from devices and will continue to do so for a very long time.
A common pattern in new industries is that they have tons of competition when they're immature. Then, one or more players figure out what people really want. The first to release gets most of the market with the First Mover advantage. The rest mostly disappear with final result being maybe 2-6 competitors getting almost whole market to themselves depending on industry.
Sorry, but Machine Learning and statistical AI is way over-hyped in my experience working in the industry. There are definitely very cool things coming from it. But the hype implies it will solve way more cases than it currently can.
Modeling complexity/cost is worse than exponential. The low hanging fruit is already taken. This doesn't mean we should stop doing it. But we should curve our expectations.
Apple has been taking a nuanced approach in blending public and on-device private analytics. The advantage of this privacy-oriented approach is that Apple gains better on-device metadata than Google, because it is based on explicit consent via structured APIs for data interchange. Over time, this can create implicit contextual search queries, with all the economic value that comes from deriving user intent.
In other words, since they got permission once in the past by means of some impossible to read ToS, they'll be better at reading our minds so they can monetize it? And we'll all feel better about it than Google's method of sucking up everything that's available and having their AI figure out how to monetize it?
Yes, I was wrong, and I see your point. The article says AI could be Apple's undoing, though, if it is able to deliver something far beyond what Apple can do.
Are you saying that even if this is the case, users will stick by Apple instead of abandoning their platform? Or that Apple will still be able to make enough on a diminished user base because of their privacy concerns?
The latter seems a little like RIM's strategy of marketing secure communications until they had no more market share.
On the AI front, Apple of course needs to compete. But AI/statistics relies on a combination of signal and compute. If Apple can convince customers that they will honor customer consent signals (on-device permissions), they can gain cultural permission to mine data on the device, then send only a computed result to their servers for combination with public data/signals. They can add another custom processor on the device for user data analytics (there is already a motion co-processor).
Apple has a good track record of using marketing to sidestep technical limitations, although it remains to be seen if that can be maintained without Jobs. They were relatively early in marketing Siri (long before Echo), even if the execution has been weak.
It's a fallacy that AI requires loss of privacy. It's easy to forget that the most powerful general purpose computing object on the planet is the human brain. Any AI strategy must leverage this in some manner. Harvesting historical data is one tactic. There is no reason why we cannot create AI business models around consent and 1:1 delegation from AI to human, e.g. explicit permissions.
Apple seems well aware of the technical challenges in protecting privacy while training algorithms, e.g. Apple Maps only retains query data for 15 minutes, https://thestack.com/cloud/2015/09/07/apples-privacy-policie... . Past technical breakthroughs have come from constraints, let's hope Apple can recruit smart people up to the challenge of balancing privacy & AI.
His Apple concerns are merited. His Blackberry story I somewhat disagree with. Relevant quote:
"No new initiative, management change, or acquisition in 2007 could’ve saved the BlackBerry. It was too late, and the gulf was too wide."
He's saying Blackberry couldn't make it vs Apple at what Apple did. That's true. Yet, Apple wasn't doing what Blackberry did: great at battery life, enterprise integration, business-grade apps, and (relative to others) security. Apple might have knocked out half to three quarters of Blackberry's user-base with some that still care about those above features.
I kept saying Blackberry should buy a better microkernel, reinvent their image for reliability/security, catch up where they can, double down on enterprise stuff, and go from there. They started doing some of that with QNX and new Blackberry OS's. The Playbook smoked the iPad in comparisons with incredible responsiveness that's still inconsistent on iOS and Android. It's security potential at OS/component layer is still higher given it's a proper, fast microkernel.
Also, the author noted that the reason Apple was unstoppable was that they built on 10 years of Mac-related research and tooling to produce the phone. The reason that's wrong is that their strategy is exactly the method Google used to fight them: building on Linux, Java, and otherwise over a decade of proven, constantly-improved software. Blackberry could've done something very similar with their own interface and API on top of it. The OS would shift where more of it was free and open but their key benefits would be the differentiator. Their app ecosystem would differentiate similarly by leveraging those benefits.
So, I think there was an exit strategy available for RIM that they just didn't capitalize on properly. I about wish I could've been in there to help them as we really need privacy/security/reliability-focused platforms at mass-market prices. Especially smartphones. Apple is luxury with inconsistent, security results. Android is open enough that many harden it but its owner runs surveillance platforms. Blackberry still had potential to be different there. Still does although I wouldn't place any bets at this point...
Instead, we just get to watch it implode trying to be a better Apple than Apple and now a half-arsed Android. All I hope is that something good happens to QNX's as RIM's acquisition of it is suddenly a bad thing. Lots of tech disappears or withers because it's absorbed by idiotic bureaucracies in downward spirals.
Hmm, I've never directly given Google a dollar, but I've spent thousands on Apple stuff primarily because they do not have big-data AI services. No company is perfect, but it's easy to see their priorities.
Thats the scary thing about Google. They've become so ubiquitous that you don't need to pay them directly for them to make money off of you. They make money off of the data that they are able to glean from your interactions on the web, your devices etc. Its strength might also be its weakness if the mass of consumers shift towards a more privacy oriented mindset.
Its strength might also be its weakness if the mass of consumers shift towards a more privacy oriented mindset
I'm guessing you mean consumers will stop using the internet because otherwise your comment makes no sense. You don't have more privacy on the internet because you don't use Google services.
I'm not sure which part of these spells "Apple’s apparent inaction" in AI field stated in the article?
[1] https://sg.finance.yahoo.com/news/exclusive-apple-ups-hiring...
[2] http://iphone.appleinsider.com/articles/15/09/07/apple-finds...
[3] https://jobs.apple.com/us/search?#specs&ss=machine%20learnin...