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

Apple's accidental moat: How the "AI Loser" may end up winning

adlrocha.substack.com

Post reply on HN
Print view View original

341–350 of 402 posts

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#341

Post by dwaite »

lern_too_spel wrote: post #310
crote wrote: post #118

Earlier quoted context omitted.

The hard part is doing so without completely ruining the existing app ecosystem. Rosetta 2 is genuinely impressive.

Vs. FEX and Prism?

Yes. Apple put custom hardware support in the M series chips based on the needs of Rosetta 2. The x86_64 performance on Rosetta 2 was often higher at launch than the prior generation of Intel chips running those same binaries natively.

Microsoft and Qualcomm already knew the performance of x86 app emulation on windows was killing the ARM machine lineup, so Qualcomm was working on extensions to their chips and Microsoft on having Windows support them already, but ARM64EC and Prism didn't launch for two years after the M1 shipped.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#342

Post by alwillis »

Gigachad wrote: post #58
colechristensen wrote: post #41

Earlier quoted context omitted.

Local models seem somewhere between 9 and 24 months behind. I'm not saying I won't be impressed with what online models will be able to do in two years, but I'm pretty satisfied with the prediction that I won't really need them in a couple of years.

We still aren't going to be putting 200gb ram on a phone in a couple years to run those local models.

> We still aren't going to be putting 200gb ram on a phone in a couple years to run those local models.

You can already buy an iPhone with 2 TB of storage. The CPU, GPU and Neural Engine all share the same pool of RAM and the SSD is directly connected to all of this. You won’t need 200 GB of RAM to run local models when you essentially have 500 GB of virtual memory.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#343

Post by dwaite »

maplethorpe wrote: post #181
grtteee wrote: post #2

This is the classic apple approach - wait to understand what the thing is capable of doing (aka let others make sunk investments), envision a solution that is way better than the competition and then architect a path to building a leapfrog product that builds a large lead.

Didn't they rush to integrate ChatGPT into their OS back in 2024? Reality doesn't seem to align with your description.

I wouldn't describe it as 'rushed'. Its integrated pretty much exactly the way they said it would be, as a fall-back from Siri when you ask world knowledge questions.

The part that doesn't work is having Siri locally smart enough to use it as a tool.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#344

Post by alwillis »

adrian_b wrote: post #139
jurmous wrote: post #98

Earlier quoted context omitted.

We don’t need 200gb of RAM on a phone to run big models. Just 200 GB of storage thanks to Apple’s “LLM in a flash” research. See: https://x.com/danveloper/status/2034353876753592372

Yes, I agree that this is the right solution, because for a locally-hosted model I value more the quality of the output than the speed with which it is produced, so I prefer the models as they were originally trained, not with further quantizations. While that paper praises the Apple advantage in SSD speed, which allows a decent performance for inference with huge models, nowadays SSD speeds equal or greater than tha…

> While that paper praises the Apple advantage in SSD speed, which allows a decent performance for inference with huge models, nowadays SSD speeds equal or greater than that can be achieved in any desktop PC that has dual PCIe 5.0 SSDs, or even one PCIe 5.0 and one PCIe 4.0 SSDs.

Apple’s advantage is their unified memory architecture where the CPU, GPU and Neural Engine share the same memory and the SSD is directly connected to the SoC--less latency than PCIe. Memory bandwidth starts at 300+ GB/s.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#345

Post by ascagnel_ »

ajdegol wrote: post #195
Fricken wrote: post #170

Earlier quoted context omitted.

All the king's horses and all the king's men couldn't come up with a killer app.

I think it was $1000 too expensive to take off. It’s also too heavy, they should drop the front screen and ruthlessly save weight. Chicken and egg problem, if no-one buys it, no-one will develop any killer apps. Whether it’s pleasant to have any screens that close to your eyes - or ever will be - is maybe the bigger question for VR.

> Chicken and egg problem, if no-one buys it, no-one will develop any killer apps.

Disagree on this. Going back as far as VisiCalc, it's about a device making space for a killer app, and that killer app selling devices. Apple has torched so much developer good-will that even a lower price wouldn't make the space for a killer app.

When was the last time a new, mobile-first killer app came out?

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#346

Post by dwaite »

dangus wrote: post #6
grtteee wrote: post #2

This is the classic apple approach - wait to understand what the thing is capable of doing (aka let others make sunk investments), envision a solution that is way better than the competition and then architect a path to building a leapfrog product that builds a large lead.

It’s even more superpowered than previous implementations of this strategy. When they made the iPhone, iPod, and Apple Watch they had no specific hardware advantage over competitors. Especially with early iPhone and iPod: no moat at all, make a better product with better marketing and you’ll beat Apple. Now? Good luck getting any kind of reasonably priced laptop or phone that can run local AI as well as the iPhone/Ma…

All three of those products launched with custom hardware made by partnered manufacturers.

At iPhone launch, I seem to remember Apple still having quite a bit of the flash ram market tied up from their exclusive iPod contracts - Apple basically helped finance new factories to be spun up in return for exclusive access to their production.

The Apple Watch had the S1 system on package, which included an Apple custom CPU. There were a number of miniaturization techniques and custom parts Apple used which I remember competitors lagging on being able to replicate due to the broader market tendency to integrate off the shelf products (but I don't have more part examples or timelines).

Since they try to stay secretive about upcoming products, competitors may only get hints about what Apple is doing through your typical industrial espionage channels until the product comes out. That creates quite a bit of lag then you are starting a new product design cycle based on a product your competitor just hit the market with.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#347

Post by carlosjobim »

smt88 wrote: post #219
tokioyoyo wrote: post #77

Earlier quoted context omitted.

Yeah, that just doesn't pass the simplest sniff tests. I barely use iMessage, and yet I'm an iPhone user. Basically everyone around me is the same.

> Yeah, that just doesn't pass the simplest sniff tests. I barely use iMessage, and yet I'm an iPhone user. A single anecdote isn't data. You're not a typical consumer. The only major market where iPhone outsells Android (number of handsets) is the US, and it's because of iMessage. Android is 70% of the world market and dominates LatAm, Africa, and Asia.

But now you've argued that people are buying Android instead of iPhone just because of social pressure from their peers.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#348

Post by codeptualize »

crote wrote: post #118
moondev wrote: post #91

Earlier quoted context omitted.

How is changing the architecture of a platform that only you make hardware for doing the impossible? They could change the architecture again tonight, and start releasing new machines with it. The users will adopt because there is literally no other choice. Every machine they release will be fastest and most capable on the platform, because there is no other option

The hard part is doing so without completely ruining the existing app ecosystem. Rosetta 2 is genuinely impressive.

Exactly this! Rosetta + the whole app developer community who really quickly released builds for M chips (voluntary or forced, but it did happen).

I had the initial m1 air, and it was remarkable how useable it was. You'd expect all sorts of friction and issue but mostly things just worked (very fast). Even with some Rosetta overhead it was still fast compared to intel macs.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#349

Post by planb »

mdorazio wrote: post #267
planb wrote: post #197

Earlier quoted context omitted.

Services revenue is mostly just 30% from App Store Sales. This means every time a user clicks a pro account for ChatGPT or Claude on their phone, Apple makes more money than they could make with a self deployed model.

Not true. App Store is less than 1/3 of services revenue.

Source? But it‘s true that I forgot about the Google Search deal.

Re: Apple's accidental moat: How the "AI Loser" may end up winning

#350

Post by lern_too_spel »

dwaite wrote: post #341
lern_too_spel wrote: post #310

Earlier quoted context omitted.

Vs. FEX and Prism?

Yes. Apple put custom hardware support in the M series chips based on the needs of Rosetta 2. The x86_64 performance on Rosetta 2 was often higher at launch than the prior generation of Intel chips running those same binaries natively. Microsoft and Qualcomm already knew the performance of x86 app emulation on windows was killing the ARM machine lineup, so Qualcomm was working on extensions to their chips and Microso…

FEX uses TSO on M series chips.
Post reply on HN