Hacker News Reader

Comment by jandrewrogers | Hacker News Reader

That article is a bit incoherent.

I think it is pretty clear at this point that Apple has gone all-in on being the ideal edge silicon for AI. This leverages their core competencies, requires only modest investment, and will likely pay out no matter how the AI market eventually shakes out. They are in one of the only parts of the obvious future AI market where there isn't really a fight for greenfield turf with other big companies.

Staying in their lane is arguably the optimal business decision for Apple and they lose nothing by it.

View on HN

Here's an interview with the senior product manager of Apple Silicon that more-or-less says the same thing: https://www.macrumors.com/2026/07/06/apple-silicon-exec-expl...

> He also described a shift toward running AI locally rather than in the cloud – a move motivated by privacy, security, and the rising cost of inference as agents consume more tokens. However, Brooks envisions a hybrid future in which agents decide what runs on-device and what gets sent to the cloud.

And coupled with their efforts in auditably-private cloud computing, that's a strong pitch.

> auditably-private cloud computing

Do those of you who have read the white paper have confidence in this claim (as I’d hope)?

Blog: https://security.apple.com/blog/private-cloud-compute/

The top threads here have commentary: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu... especially https://news.ycombinator.com/item?id=41937664

IMO - while nothing is a guarantee against a state-level actor that can compel both hardware-level bypasses and put such bypasses under a gag order, it's a well-thought-out system under those constraints. The idea of a self-auditing and self-healing fleet that requires simultaneous compromise of software and enclave hardware is a meaningful level of defense in depth against literally everything else.

Just feels like the Apple press/podcast circuit just keep trying to spin ways to frame Apples fumble of the AI era as some sort of savvy move.

AI is overhyped regardless of whether Apple exists or not. They’re getting attention only because they didn’t jump on the bandwagon, and from a purely investment standpoint their fiscal discipline is honestly astounding.

When your competition is currently burning billions of dollars left and right, Apple refunding Google $1 billion and giving OpenAI zero dollars seems pretty smart to me, Apple‘s largest acquisition in their entire history is only three billion dollars that’s planning.

Apple's competition is high-performance compute. Nvidia is eating their lunch to the tune of $4 trillion, and Apple just let it happen. They left billions on the table in revenue that Nvidia eats up with their own ARM servers that aren't arbitrarily restricted by the OEM.

marketing has always been one of their strong suits

If they've gone all in on that, it looks even worse for Apple. Qualcomm's Hexagon has been a better NPU since the beginning. Now even Mediatek is better at it than Apple. Blogs like macrumors and daringfireball are notorious for not having any clue about what is going on in the broader industry.

I generally agree. Xserve didn't work out for Apple and they dropped it pretty quickly--think I still have an A/UX coffee cup. Wasn't really in their DNA even if there was an argument before Linux took off. For better or worse, they're basically a client company and barely a software (OK and basically a software services) one--although they spend a fair bit on (roughly) user experience.

They sold the Xserve for ten years, across the PowerPC and Intel eras.

Still, in the scheme of Apple stuff it was a failed experiment--even if a worthwile one.

We were one of the first Xserve customers (Xserve and Xserve RAID).

I do agree it was an experiment. Unless someone has firsthand knowledge, we won’t know if it was someone’s pet project, a market they wanted to be in to dominate Dell, IBM, etc.

However, I would not call it a failed experiment. One of the largest supercomputers (at the time) was an Xserve cluster. Nothing about the experiment failed.

I will tell you that Steve Jobs was not a fan of the Xserve product line. He did not support it, which meant its days were numbered. He stated Apple was a consumer company. Not a business enterprise company. The Xserve teams were soon disbanded not soon afterwards.

Without any inside knowledge--and I'd have to revisit a piece I wrote at the time during my prior analyst stint--it always felt like a bit of a sideline/experiment. There were a lot of dynamics going on at the time. Linux was really exploding (and improving). There was a lot of consumerization (especially on mobile) happening within enterprise IT; Apple didn't need to become an enterprise IT company to effectively reap a lot of the benefits both directly and indirectly. Servers were increasingly commoditizing onto dual socket x86 (with some outliers).

It was worthwhile because in the end it taught them that they needed to do something like Apple Silicon. And hopefully the current memory fiasco will convince them that they need to take memory in house if they want to build what they need to build going forward.

I really don’t see any alternative long-term to not engineering and designing memory in house and collaborating with TSMC onshore in America.

Hopefully having an engineer from the Apple Silicon design group as the next CEO will pay off.

Maybe. Intel got out of the memory game baecause it became a commodity.

That said, I was involved in some research years back where there was the potential opportunity for memory optimized for specific purposes. Which didn't make sense for the vendor at the time. Apple's already doing it to some degree on-chip. But you can imagine it making sense for a vendor owning the whole stack which wasn't really the case for the vendor we were doing the research for.

A/UX predates Xserve by 14 years.

And how much did A/UX sell? I Have an A/UX coffee mug in my kitchen. It even has a phone number on it.

They've been investing in NPUs for a long long time. Since the first iPad Pro at least (2018). Apple Silicon is a natural fit for tensor/matrix operations due to unified memory, so if anything them going "all in" on that has been a nice side effect.

> Apple Silicon is a natural fit for tensor/matrix operations due to unified memory,

This is a common misconception. This is NOT true: memory designed for CPU performs horribly for GPU tasks. Mac/Strix Halo/NVIDIA Spark are performing horribly compared to a desktop video card with GDDR.

I spent quite a few weekends optimizing ROCM kernels for strix halo, i WISH I had GDDR instead of high-frequency CPU RAM; the bandwidth would be SO MUCH better.

I'm quantizing weights not to make computation faster, not because I'm out of memory, but because memory cannot move fast enough to be processed and it's cheaper to load quantized version, convert into bf16 compute, and discard; This happens every single time a token is generated for the whole model over and over.

The whole point of Apple silicon isnt speed, it is capacity. You can grab a Mac Studio with 192 gb of unified memory and shove a model in there that would otherwise require building a rig with multiple rtx 4090 s. For local dev just being able to fit the weights in memory is often way more important than token generation speed

I don’t think they need to go all-in on anything related to AI.

> This leverages their core competencies, requires only modest investment, and will likely pay out no matter how the AI market eventually shakes out.

This looks more like a call or a small raise or whatever (I don’t play poker).

Not really. Their new Siri is powered by Gemini.

Reply on news.ycombinator.com