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

Apple Foundation Models

platform.claude.com

Post reply on HN
Print view View original

221–230 of 244 posts

Re: Apple Foundation Models

#221

Post by 5701652400 »

post-it wrote: post #120

> a Swift package that makes Claude available as a server-side language model in Apple's Foundation Models framework Ahh I was hoping for the opposite: all of the existing features of Claude Code but somehow running locally on my laptop's neural engine. A pipe dream on an M2 with 8 GB of RAM, but I had a flicker of hope there.

I would not mind if cloud was actually private users iCloud. users pay for it, and it runs in Apple servers next to where users store their iPhotos already. that would be really elegant solution.

..but instead we get Claude, hosted who-knows-where. maybe in X-AI datacenters? maybe in Amazon somewhere? who knows..

Re: Apple Foundation Models

#222

Post by Tagbert »

5701652400 wrote: post #220
WorldMaker wrote: post #172

Earlier quoted context omitted.

Apple's been trying to make the marketing appeal that "Private Compute Cloud" is also a hardware project. Given it seems to rely on low level details of device Hardware Security Modules, it's maybe even at least a little bit more than just "marketing spin".

looks like it is not "Private iCloud Compute" at all. Anthropic literally says "Requests go directly from your app to the Claude API; Apple is not in the request path and does not see prompts or responses." — Apple straight up lied

No, that post is about Claude for Foundation Models. That is not the same as Apple Intelligence.

the Swift package for Claude for Foundation Models is about sending calls to Claude. That had nothing to do with Apples models which do use local models and models on Private Cloud Compute.

Your accusation that "Apple straight up lied" is based on misunderstanding TFA.

Re: Apple Foundation Models

#223

Post by FuriouslyAdrift »

connicpu wrote: post #173
godzillabrennus wrote: post #132

Earlier quoted context omitted.

If only we could buy 1TB of unified memory in a Mac for $1k-$2k in total hardware costs. Apple would basically be able to extinguish the entirety of the market cap for Nvidia, OpenAI, Anthropic, and others all at once. In 10 years, I hope my MacBook Pro can run today's frontier models and has 1TB of unified Memory.

The Nvidia GB300 DGX Station, which isn't even going to hit 1TB total memory, is expected to launch at almost $100k. Bit of a pipe dream with memory prices where they're at.

There are multiple server systems available right around the $100k range that have 512B of GPU RAM right now (4x AMD Instinct MI300A)

GIGABYTE G383-R80-AAP1 for example

Re: Apple Foundation Models

#224

Post by tempoponet »

shadowpho wrote: post #155
godzillabrennus wrote: post #132

Earlier quoted context omitted.

If only we could buy 1TB of unified memory in a Mac for $1k-$2k in total hardware costs. Apple would basically be able to extinguish the entirety of the market cap for Nvidia, OpenAI, Anthropic, and others all at once. In 10 years, I hope my MacBook Pro can run today's frontier models and has 1TB of unified Memory.

Why can’t Apple launch a $50k product for $1k? Everyone would buy it!

To go further down this pipe dream - Anthropic / OpenAI would buy them all and still price out the consumer. There's no end-run in this scenario.

Re: Apple Foundation Models

#225

Post by willy_k »

5701652400 wrote: post #221
post-it wrote: post #120

> a Swift package that makes Claude available as a server-side language model in Apple's Foundation Models framework Ahh I was hoping for the opposite: all of the existing features of Claude Code but somehow running locally on my laptop's neural engine. A pipe dream on an M2 with 8 GB of RAM, but I had a flicker of hope there.

I would not mind if cloud was actually private users iCloud. users pay for it, and it runs in Apple servers next to where users store their iPhotos already. that would be really elegant solution. ..but instead we get Claude, hosted who-knows-where. maybe in X-AI datacenters? maybe in Amazon somewhere? who knows..

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

Re: Apple Foundation Models

#226

Post by throwaway85825 »

CuriouslyC wrote: post #99
tedggh wrote: post #72

Earlier quoted context omitted.

Benedict Evans may be right after all; frontier models look more and more like telecom companies in the 90s. Billions and billions of investment in infrastructure while others further up the stack captured all the value.

There will be frontier models that are non-commoditized, but they'll be kept guarded and hidden away, and you'll only get the final result, so that they can't be distilled and their harness can't be reverse engineered. They'll be billed like employees, rather than like a tool.

The economically useful frontier models will be fine tuned on data to make them useful for a specific project or task.

Re: Apple Foundation Models

#227

Post by lumost »

CuriouslyC wrote: post #144
yandie wrote: post #140

Earlier quoted context omitted.

I doubt that. What stops the Chinese labs from figuring it out? It’s not like these models are fundamentally different from each other

If all you have is the starting point and the finishing point, the lack of the path taken from one point to another limits your ability to train models that can efficiently recreate the work, and increases its cost enough that it's possible the US labs can progress capabilities faster than Chinese labs can distill that behavior.

This just looks like a capex problem. There is no evidence that Anthropic has secret sauce above and beyond access to capital. If there is secret sauce, it's unclear that it changes the required amount of capital by all that much.

China will spend all of the money required to catch up, Google and OpenAI will both spend money to catch up as well. NVidia and others will not allow a frontier lab to become the AI bottleneck.

Re: Apple Foundation Models

#228

Post by embedding-shape »

computerex wrote: post #207
embedding-shape wrote: post #82

Earlier quoted context omitted.

> I think Evans is completely wrong. There are only 2 truly frontier models. (at least for now). And Anthropic seems to be leaving OpenAI behind so there might be only 1 in the near future. (which is scary/dangerous) Truly fascinating ecosystem and community in general, as experiences differ so wildly. Anthropic's models seems far behind OpenAI to me, especially when you get into "Pro" territory, and there doesn't se…

For HPC/ai work opus blows gpt away, it’s no competition.

As someone who just spent the last three days (tried using both, ended up using mostly Codex) implementing DiffusionGemma in Rust, I think they're more or less equal when it comes to machine learning and AI. They get stuck at different points, but wouldn't say one is a clear winner over the other. HPC I have no idea so I'll take your word for it :)

Re: Apple Foundation Models

#229

Post by enos_feedler »

tedggh wrote: post #72
harrouet wrote: post #49

This is Apple commoditizing LLMs while keeping control of the UX. They are a hardware company and will keep selling the best machine for AI use. Well done.

Benedict Evans may be right after all; frontier models look more and more like telecom companies in the 90s. Billions and billions of investment in infrastructure while others further up the stack captured all the value.

He denies comparing them to telecom companies and even says at various points in his writing. Instead he compares their usage to the usage of mobile data.

Re: Apple Foundation Models

#230

Post by ls612 »

hedora wrote: post #129
CuriouslyC wrote: post #99

Earlier quoted context omitted.

There will be frontier models that are non-commoditized, but they'll be kept guarded and hidden away, and you'll only get the final result, so that they can't be distilled and their harness can't be reverse engineered. They'll be billed like employees, rather than like a tool.

The non-commodity network services of the early 1990’s and the non-commodity 3d graphics hardware of the mid-1990s made the same argument.

They didn’t have the security state backing up their business thesis at gunpoint.
Post reply on HN