Hacker News new | ask | show | jobs
by hdevalence 369 days ago
It’s odd that the article identifies Apple’s hardware as a limitation for AI. I don’t think this is the case. If anything it’s the opposite, and makes Apple’s lack of execution more mysterious.

I was running Stable Diffusion on my iPhone two years ago. You can get quite good open weights models running on-device today. What’s going on over there?

7 comments

I don't think there's much mystery here. Apple has blocked their researchers from publishing and has a very siloed approach. Researchers don't like this, so they work for companies that allow them to publish and engage in the research community. As a result Apple can't hire the talent needed to execute in this space.
> Apple has blocked their researchers from publishing

Counterexamples: https://machinelearning.apple.com/research/

Sort by oldest. They had two papers in 2017 and three in 2018. It was only in 2019 that they started opening up more, but you can't change decades of culture quickly. The AI researchers I know (and I know a lot, having been VPE at multiple AI companies) would rather work with groups that are more open.
They did start this but they were a few years late to the game, so they're probably having to play catch up.

Also they still have a streak of secrecy so I wonder if any researchers are skittish thinking that Apple might just arbitrarily stop them from publishing certain papers?

As if xAI publishes a lot. But even they have models that work.
As far as I can tell xAI doesn't publish because they don't have anything to publish. It seems to me like they aren't doing anything novel or state of the art, they're just consuming other people's research and implementing it.
This invalidates your thesis that Apple is behind on AI because they can't hire talent. And that they can't hire talent as they don't innovate and publish papers.

xAI doesn't innovate, doesn't publish papers, and started literally a decade later than Apple into AI, and yet they're ahead of Apple in AI.

So the high order bit is something else. Apple seems consumed by its legacy dependencies, structures... people even. There's no one on top with focus. I'm not even saying "with vision" or "with competence" or "with expertise". I'm just saying "with focus". If Tim Cook wanted an LLM he'd have a top of the line LLM running in 6 months with the resources he has. He doesn't.

His focus is on making lots of the things he wants to sell, and he doesn't care, or understand what those things are. This is why they have warehouses full of unsold Vision Pro inventory now. He's a logistics guy. Blind to the rest.

And Apple was doing great with someone like him at the top for years. But now the backlog of products is over and they keep shipping duds, while their software work is nosediving.

They need someone who understands "product" stat. Someone with holistic instincts. Someone like Jobs. I mean at least a bit like him.

I wonder if AI is just not up to Apple's standards yet. It's very, very amazing in some ways, but also, deeply flaky in others. Remember all the Apple news summarization memes? That sort of thing bothers any tech company but it really bothers Apple. "Very amazing but also deeply flaky" is fine for the Android ecosystem but it's not on brand for Apple.

I suspect Steve Jobs would be very aggressively driving them internally but also not necessarily releasing much yet either.

This is much less of a Jobs Apple and more of a Sculley Apple
> The problem there is twofold. One is that so-called on-device AI hasn’t yet proved to be a major selling point for products such as PCs and smartphones. The other is that Apple’s lack of its own cloud-based AI capabilities leaves the $3 trillion company still in need of powerful allies. The company struck a deal with OpenAI last year to effectively back up its own AI capabilities, and analysts have widely been expecting similar partnerships involving other major AI services, including Google’s Gemini.
I think the challenge is battery. Frequent inference draws a ton of power.
Yes. If I have to trade off a stupid phone with 100% determinism and good battery life versus an intelligent phone with 70% determinism and shit battery life then I'm going to take the stupid phone.
I get 4 or so days of battery life. If I lost a day I don't think I'd notice.
My phone 15 pro is almost always dead by the end of the day. Yesterday it actually died and I had to plug it in. last night it was at 5%

and, I’m not a major user. I don’t TikTok, X, Instagram, or any other social media. Nor do I do much else on my phone that I see heavy users do around me

You can see which apps are eating battery in settings. I do 80% charging doing about 6 hours of screen time a day and I usually end the day around 30-50%.
Did you notice battery life tanking a few weeks ago or am I imagining things?
noticeable decrease after updating ios. yay
15 pro here. I noticed it with 18.5
That’s where PCC comes in. But I still don’t understand how their device margins are going to fund all these AI data centers

Especially now that their App Store revenue is collapsing

Meanwhile, the main attraction of the upcoming iphone 17 lineup is a thin and light "Air" model that has a bigger screen and a smaller battery.
Thinner with a smaller battery, yes. Not a larger screen though, actually slightly smaller than the "Plus" iPhone which it will supposedly be replacing.
I wouldn't say that's a Unique Apple Problem though.
Wrong or not, it's a danger for Apple when this sort of thing starts popping up in the media.

It wouldn't take much for a "Apple have taken a wrong turn, and have left themselves unable to participate in the race for AI" narrative to take hold. If that happened, it would severely damage their position at the premium end of the market.

All I know is that my iPhone 13 pro had better battery life than my iPhone 15 pro. Is this Apple Intelligence chip just draining my battery?
iPhones have less ram than android though, Android phones often have 12gb or 16gb, apple only recently upped it to 8gb.

https://ioshacker.com/iphone/how-much-ram-does-my-iphone-has