Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why Apple believes it’s an AI leader–and why it says critics have it all wrong

arstechnica.com

Post reply on HN
Print view View original

41–50 of 217 posts

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#41

Post by gnicholas »

jointpdf wrote: post #34

Setting aside all the other work Apple is doing in this area, can someone explain why iMessage autocorrect is unacceptably bad (to put it diplomatically) in the era of weapons-grade language models? Its behavior is totally bizarre. It’s like an underpowered chess engine that makes flagrant blunders: capitalizing random words in the middle of a sentence (rock -> Rock), the context sensitivity of an actual rock, forcin…

Yeah I recently moved to Menlo Park and for some reason after I type “Menlo” (which it correctly capitalized automatically), it always suggests random words like “this” or “it”.

Sure, some people would say “I went to Menlo for lunch today”, but it should always have the option of “Park” (which when I type it, it always knows to capitalize).

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#42

Post by jolux »

aspenmayer wrote: post #21
jolux wrote: post #19

Earlier quoted context omitted.

I think a lot of people have underrated the benefits Apple reaps in this area from being so far ahead in chip design and controlling their hardware. As regards to successfully productizing ML I’ve recently become convinced that Apple has been more successful at it than Google so far. Translation was already well within Google’s wheelhouse and they had been doing it for years, and their work with cameras is going to i…

Pixel is Google being aware of their handicap in this in-house hardware prototyping and research area, and ironically and obviously and predictably for Google, they apply ML to it too. Apple can come at this same problem from the hardware and software sides all at once, with their own internal dev cycles aligning with the yearly iOS/iPadOS software drops, and iPhone/iPad hardware drops timed to a month or two of each…

How is this at all related to discussing Apple’s position in AI/ML?

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#43

Post by plutonorm »

Veedrac wrote: post #26

Apple's AI work is decent for what it is, and their AI hardware is fine too, but so much of this article is just weird. It starts by talking about how until recently Apple wasn't doing AI work where it needed to be. Then there's the weird excerpt where he claims Google is “not known for shipping consumer experiences that are used by hundreds of millions of people.” The author then raises the legitimate point that AI…

Was it written by gpt3?

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#44

Post by canofbars »

enjeyw wrote: post #10
f2322323ffff3 wrote: post #5

One thing where apple is shining is brainwashing. It looks very funny when caravans of people are waiting store opening to hysterically buy an overpriced smartphone, even if it's technologically 2 years behind cheaper Samsung S20 (except chip performance, but it doesn't matter anymore, because there is no tasks that needs so much computing power)

> because there is no tasks that needs so much computing power I think your own point explains why people are more than happy to use a more expensive iPhone despite technological differences. I'm currently developing with a relatively new android phone, but I'd never use it as an actual day-to-day phone in lieu of my circa 2014 iPhone (believe me I've tried). The extent to which I prefer the overall iPhone experience…

I just switched from a pixel 2 to an iphone 11. For me it seems that iOS and Android are pretty level quality wise, they both have minor things that they do better but the number one thing that makes the iphone better to me is they are able to do all the awesome stuff android does without sending your data away for processing. Having image recognition without the privacy loss is worth the price tag to me.

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#45

Post by jolux »

Veedrac wrote: post #26

Apple's AI work is decent for what it is, and their AI hardware is fine too, but so much of this article is just weird. It starts by talking about how until recently Apple wasn't doing AI work where it needed to be. Then there's the weird excerpt where he claims Google is “not known for shipping consumer experiences that are used by hundreds of millions of people.” The author then raises the legitimate point that AI…

Apple’s chips are way better so whether they’re doing more locally or not they have the headroom to do a lot more.

I would agree with that characterization of Google. They have a very small handful of successes and an enormous pile of failures that they’ve discarded. They give every impression of not really knowing what to do.

I think Giannandrea was also referring specifically to the kinds of experiences Apple can provide on the iOS and iPadOS platform because of their high end hardware and deep software integration. Google has yet to replicate that, and even seem to be bored with Android recently: https://daringfireball.net/linked/2020/08/05/wear-os-music

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#46

Post by nfg »

jointpdf wrote: post #34

Setting aside all the other work Apple is doing in this area, can someone explain why iMessage autocorrect is unacceptably bad (to put it diplomatically) in the era of weapons-grade language models? Its behavior is totally bizarre. It’s like an underpowered chess engine that makes flagrant blunders: capitalizing random words in the middle of a sentence (rock -> Rock), the context sensitivity of an actual rock, forcin…

The one the drives me mad the most is when I type a word which it incorrectly autocorrects to a proper noun (which gets capitalised), I then use backspace to delete that word and type the - completely different - correct one, but it insists on upper-casing the first letter to match the original casing of the incorrect autocorrect. It really should keep track of the fact that it introduces the case-change and invalidate it when the automated action itself is invalidated. This happens frequently to me.

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#47

Post by majestik »

> After a brief pause, he added: "I guess the biggest problem I have is that many of our most ambitious products are the ones we can't talk about and so it's a bit of a sales challenge to tell somebody, 'Come and work on the most ambitious thing ever but I can't tell you what it is.'"

JG, I don’t think that’s your “biggest problem” - Siri is. Your privacy centric on-device strategy limits your view of user feedback, Google gets a lot of shit wrong but they know how to transmit user data and understand their feedback.

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#48

Post by llampx »

moandcompany wrote: post #23
Despegar wrote: post #3

This piece is interesting because Apple was saying this all along but no one really believed them because it sounded like excuse making. But here JG is basically saying the same thing: >Yes, I understand this perception of bigger models in data centers somehow are more accurate, but it's actually wrong. It's actually technically wrong. It's better to run the model close to the data, rather than moving the data around…

When it comes to machine learning, the subtlety here is that there are at least two sides or facets to machine learning: (1) training and (2) inference. It's fair to say that there are multiple areas for AI leadership. It is generally believed that: Model Creation (1) Those with the access to the best (which is not necessarily the most, but often believed to be) data have a strong starting point for training models;…

Re: Prediction at the edge, I would think that it's better if there aren't going to be any updates to the model. Or if internet access is limited. Correct me if I'm wrong, but most of the ML inference actually takes place on the cloud nowadays, not on-device.

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#49

Post by travbrack »

jointpdf wrote: post #34

Setting aside all the other work Apple is doing in this area, can someone explain why iMessage autocorrect is unacceptably bad (to put it diplomatically) in the era of weapons-grade language models? Its behavior is totally bizarre. It’s like an underpowered chess engine that makes flagrant blunders: capitalizing random words in the middle of a sentence (rock -> Rock), the context sensitivity of an actual rock, forcin…

Same experience with gboard on iOS.

Re: Why Apple believes it’s an AI leader–and why it says critics have it all wrong

#50

Post by llampx »

jolux wrote: post #19
skybrian wrote: post #18

Earlier quoted context omitted.

It's not my area, but I read Google's AI blog and they seem to be doing machine learning research that ends up in its phones, such as various camera improvements and offline Google Translate. It seems like it's the sort of thing you do when you can, but often it comes as a second phase after getting it to work in the data center. I didn't see anything in this article that was obviously unique to Apple.

I think a lot of people have underrated the benefits Apple reaps in this area from being so far ahead in chip design and controlling their hardware. As regards to successfully productizing ML I’ve recently become convinced that Apple has been more successful at it than Google so far. Translation was already well within Google’s wheelhouse and they had been doing it for years, and their work with cameras is going to i…

Even there, Google is hindered by the fact that it is so bad at hardware. The Pixel's camera lead is not as much as it used to be, and now Apple is ahead on video. The newest phone from Google uses the same sensor as the Pixel 2, whereas others have moved onto bigger sensors with quad-bayer filters.
Post reply on HN