How AI and Machine Learning Work at Apple • Hacker News Board

Re: How AI and Machine Learning Work at Apple

#71

Post by sangnoir » Wed, Aug 24, 2016, 9:38 PM UTC

c0g wrote: post #61

ThomPete wrote: post #54

Earlier quoted context omitted.

There is only one way to win and that is to publish. Innovation in this field doesn't happen in a vacuum but by multiple simpler components combine in ever more emerging complexity.

Apple can free ride on other people publishing and synthesise in house. The field might lose out (but there are always free riders) but does Apple?

> The field might lose out (but there are always free riders) but does Apple?

Yes - those in the field who prefer to publish will not be willing to join Apple. How much of an opportunity cost this is to Apple remains an open question.

Re: How AI and Machine Learning Work at Apple

#72

Post by tahoeskibum » Wed, Aug 24, 2016, 9:39 PM UTC

Nice article, but in my perception, Apple is way behind in AI vs. Google on mobile. Siri's current speech recognition is still far behind Google on my iPhone. Half of the time it Siri doesn't recognize something but Google recognizes it right way. Apple Maps continues to lag in the following features: biking and public transit and even things like which lane to take on a big interchange. As a result I end up using Google (Now) and Google Maps by default.

Re: How AI and Machine Learning Work at Apple

#73

Post by jomamaxx » Wed, Aug 24, 2016, 9:39 PM UTC

Smerity wrote: post #2

None of this really answers the overlying question that Jerry Kaplan and Oren Etzioni raised. The question raised by most in the field isn't whether Apple use AI/ML internally, the real question is why they avoid the research community so strongly. For me, the greatest thing about the ML/AI community is how open it is and how strong a sense of camaraderie there is between people across the entire field, regardless of…

"real question is why they avoid the research community so strongly."

I think it's within their DNA.

They are a very secretive company for a lot of historical reasons.

Company culture is what companies are ... they are the social rules built into the organization that are very difficult to change once established.

Apple is within their moral right to do what they are doing. There are different approaches taken by other companies so it's healthy to have that.

If this were weapons systems or cancer cures, I might have a different opinion.

Re: How AI and Machine Learning Work at Apple

#74

Post by throwanem » Wed, Aug 24, 2016, 9:41 PM UTC

Chronic9q wrote: post #65

BinaryIdiot wrote: post #32

Earlier quoted context omitted.

> If they were serious, they'd provide objective evidence for their claims. Papers, source code, numbers. Not "we have the biggest, baddest GPU farm" and refuse to say any more as policy. But why? Who cares about papers, source code or numbers beyond the tiny segment that HN caters to? Apple has always been about the user experience. When they discuss a vast majority of prior accomplishments they discuss them, many t…

> But why? Who cares about papers, source code or numbers beyond the tiny segment that HN caters to? I don't know, say, the thousands of future engineers and researchers, that, you know, make the product?

Yes. That tiny segment.

Re: How AI and Machine Learning Work at Apple

#75

Post by k-mcgrady » Wed, Aug 24, 2016, 9:42 PM UTC

jaegerpicker wrote: post #67

soperj wrote: post #49

Earlier quoted context omitted.

A lot of their products are lacking though. Maps is awful. So is Safari.

It's generally helpful if you don't trot out 2-3 year old claims. Both maps and Safari have had massive changes recently and both are if not on par with other companies it's really close. Apple has had issues with OS X reliability and music can be truly bad at times and iCloud (especially sync) is almost embarrassing. I'm a huge Apple fan and an iOS developer but they aren't without fault but if you are going to call…

Curious about your take on iCloud sync. Is that from a developer perspective (I remember the API initially sucked) or user perspective? I've never implemented iCloud sync in an app but at least with Apple's apps it works perfectly for me. I moved thousands of notes from Evernote to Notes which I use daily, I use calendars/contacts, I use Reading List, the credit card and password sync for Safari works really well. In the past there were definitely problems but I haven't come across any in quite a long time.

Re: How AI and Machine Learning Work at Apple

#76

Post by throwanem » Wed, Aug 24, 2016, 9:46 PM UTC

sangnoir wrote: post #68

atonse wrote: post #57

Earlier quoted context omitted.

I'm not so sure you actually read the article. It explicitly talks about how they've accomplished a lot of this without violating their users' privacy. And that the fact that they deny AI researchers big hoards of data, hurts their reputation in the AI community, because most AI researchers want mountains of data.

> It explicitly talks about how they've accomplished a lot of this without violating their users' privacy. My question are: A. who is violating user's privacy, B. How is Apple any different[1] from them, besides proclaiming they aren't violating privacy? 1. Some data is collected from the device and sent to a server for processing. Until Apple does something truly radical (like 100% encrypted information that is proc…

Well, to your point B: for a start, I've already paid Apple. They have much less financial incentive to use my data in ways inimical to me than, say, an advertising company would.

Re: How AI and Machine Learning Work at Apple

#77

Post by soperj » Wed, Aug 24, 2016, 9:47 PM UTC

k-mcgrady wrote: post #60

soperj wrote: post #49

Earlier quoted context omitted.

A lot of their products are lacking though. Maps is awful. So is Safari.

Safari is my browser of choice and I have no issues with it (I much prefer it over Chrome). Maps has occasional issues but it's a solid product now. I use it for walking and transit directions and can't remember the last time I had to fall back on Google Maps.

It's taken me way out of the way, and through a toll bridge when that was not even close to the best way to go. Safari is bad because it's behind the times, and doesn't keep up with standards. Makes it really annoying to have to wait to use certain features because IE and Safari can't keep up.

Re: How AI and Machine Learning Work at Apple

#78

Post by niels_olson » Wed, Aug 24, 2016, 9:49 PM UTC

throwanem wrote: post #74

Chronic9q wrote: post #65

Earlier quoted context omitted.

> But why? Who cares about papers, source code or numbers beyond the tiny segment that HN caters to? I don't know, say, the thousands of future engineers and researchers, that, you know, make the product?

Yes. That tiny segment.

In my mind, this article is squarely aimed at that segment. They are trying to tell one of the most competitive labor markets in the world, AI developers, that they're hiring, but more importantly, that they're doing interesting things.

Re: How AI and Machine Learning Work at Apple

#79

Post by throwanem » Wed, Aug 24, 2016, 9:50 PM UTC

soperj wrote: post #77

k-mcgrady wrote: post #60

Earlier quoted context omitted.

Safari is my browser of choice and I have no issues with it (I much prefer it over Chrome). Maps has occasional issues but it's a solid product now. I use it for walking and transit directions and can't remember the last time I had to fall back on Google Maps.

It's taken me way out of the way, and through a toll bridge when that was not even close to the best way to go. Safari is bad because it's behind the times, and doesn't keep up with standards. Makes it really annoying to have to wait to use certain features because IE and Safari can't keep up.

You blame Safari for not being Chrome? I mean, OK, I guess, but there are a lot of browsers that aren't Chrome. Why single out Safari for vitriol?

(And I'm not even going to touch the frankly invidious IE comparison.)

Re: How AI and Machine Learning Work at Apple

#80

Post by BinaryIdiot » Wed, Aug 24, 2016, 10:00 PM UTC

Chronic9q wrote: post #65

BinaryIdiot wrote: post #32

Earlier quoted context omitted.

> If they were serious, they'd provide objective evidence for their claims. Papers, source code, numbers. Not "we have the biggest, baddest GPU farm" and refuse to say any more as policy. But why? Who cares about papers, source code or numbers beyond the tiny segment that HN caters to? Apple has always been about the user experience. When they discuss a vast majority of prior accomplishments they discuss them, many t…

> But why? Who cares about papers, source code or numbers beyond the tiny segment that HN caters to? I don't know, say, the thousands of future engineers and researchers, that, you know, make the product?

> I don't know, say, the thousands of future engineers and researchers, that, you know, make the product?

Apple hasn't needed to do that with engineers and researchers pretty much ever; now with AI it's different? Maybe it won't help attract some small portion of people but I'm not sure why it supposedly matters now where it hasn't in the past.