Apple Intelligence Foundation Language Models Tech Report 2025

Post by astrange » Thu, Jul 17, 2025, 11:27 PM UTC

0x457 wrote: post #68

ApolloVonZ wrote: post #59

Despite all the “Apple is evil” or “Apple is behind” (because they don’t do evil). Well, what they made with the Foundation Model is great. The fact that they build a system within the Swift language that allows you to specify structured data models (structs) to be used like any other model in a modern programming language, and you actually get back generated data in that format is great. Unlike a lot of other AIs wh…

Guided generation is called "Structured Output" by other providers? Well partially generated content streaming thing is great and I haven't seen it anywhere else.

Post by astrange » Thu, Jul 17, 2025, 11:30 PM UTC

const_cast wrote: post #77

Uehreka wrote: post #52

Earlier quoted context omitted.

I wouldn’t go as far as GP, but yes, absolutely, they must compete with large models on the internet. Customers are now used to being able to ask a computer a question and get something better than “I just ran a web search for what you said, here are the uncurated, unsummarized results”. Yes, this is in fact what people want. Apple is the biggest company in the world (don’t quibble this y’all, you know what I mean) a…

This usecase is run of the mill for someone like Google, who used to store and show you your location forever, but it's not in Apple style. It's hard to be like "uhhh privacy" when you send all requests to a remote server where they're stored in clear text for god knows how long. As of right now, there is no way to run big LLMs in a privacy preserving manner. It just doesn't exist. You can't E2EE encrypt these servic…

Post by robotresearcher » Thu, Jul 17, 2025, 11:37 PM UTC

wmf wrote: post #58

EGreg wrote: post #46

Here is my question… This is the first time that millions of people will actually download and run a model on their own devices. The question is… will Apple be constantly tweaking these models, or only during OS upgrades? I for one really like local software. Call me old-fashioned, but I enjoy when a company doesn’t switch up software anytime on the server, or phone the results home all the time in order to extract m…

The model is gigabytes so I doubt they will push updates frequently.

Post by avianlyric » Thu, Jul 17, 2025, 11:46 PM UTC

Uehreka wrote: post #52

outworlder wrote: post #40

Earlier quoted context omitted.

Only if you think they _must_ compete with large models on the internet.

I wouldn’t go as far as GP, but yes, absolutely, they must compete with large models on the internet. Customers are now used to being able to ask a computer a question and get something better than “I just ran a web search for what you said, here are the uncurated, unsummarized results”. Yes, this is in fact what people want. Apple is the biggest company in the world (don’t quibble this y’all, you know what I mean) a…

Post by simonw » Thu, Jul 17, 2025, 11:47 PM UTC

astrange wrote: post #80

perfmode wrote: post #22

> We believe in training our models using diverse and high-quality data. This includes data that we’ve licensed from publishers, curated from publicly available or open- sourced datasets, and publicly available information crawled by our web-crawler, Applebot. > We do not use our users’ private personal data or user interactions when training our foundation models. Additionally, we take steps to apply filters to remo…

> Using our web crawling strategy, we sourced pairs of images with corresponding alt-texts. An issue for anti-AI people, as seen on Bluesky, is that they're often "insisting you write alt text for all images" people as well. But this is probably the main use for alt text at this point, so they're essentially doing annotation work for free.