Apple Foundation Models • Hacker News Board

Re: Apple Foundation Models

#241

Post by scotty79 » Thu, Jun 18, 2026, 8:38 PM UTC

scosman wrote: post #233

scotty79 wrote: post #198

Earlier quoted context omitted.

But models aren't universally best, especially small ones. For text Gemma is great. For vision qwen3.6 is amazing.

I think the point is: if every app chooses the best model for their use case, my phone is hosed (disk, memory swap, memory). A good-enough default might be better for the user than the each app having the best possible.

Ability to substitute designated model in a running app with another one or some abliterated version would be amazingly cool. Which means Apple won't allow that.

Re: Apple Foundation Models

#242

Post by throw10920 » Fri, Jun 19, 2026, 8:16 AM UTC

wahnfrieden wrote: post #237

throw10920 wrote: post #235

Earlier quoted context omitted.

This has got to be satire. Everyone, especially Singaporeans, know what "Chinese ingenuity" really is.

It’s merely descriptive of ingenuity required to distill models back into reasoning models without having any of the chain of thought. You underestimate the original work required because of biases

It's a uniquely Chinese cultural feature to laude finding more clever ways to cheat and steal and deceive as "ingenuity".

Re: Apple Foundation Models

#243

Post by malachi22 » Fri, Jun 19, 2026, 7:40 PM UTC

nate wrote: post #88

VadimPR wrote: post #45

How can you practically use this in software if you're to deploy this to users? Asking a user to create and enter their own API key is a bar too high for good UX.

Ugh. It really is. I have allihat.com which is the only safari extension (i think still) that talks to claude. And it's well sought for. But you as a user have to enter a friggin claude api key. :( And I still don't grok their TOS around this. Like you can still type: ```setup-token Set up a long-lived authentication token (requires Claude subscription)``` but this seems like a trap? :) Whose using this? Doesn't this…

[flagged]

Re: Apple Foundation Models

#244

Post by 5701652400 » Sun, Jun 21, 2026, 9:55 AM UTC

Tagbert wrote: post #222

5701652400 wrote: post #220

Earlier quoted context omitted.

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.

"... technically, I did not have sex with her!"

this is a blurry line. when factual statements are so confusing (deliberately? we well never know), there is not much difference between lie and technically-turth ful misdirection. harm is done. ill-intent is likely present.

you can burry "facts" into thousands pages long T&C (apple is infamous for) nobody reads. and claim "you agreed to that!". nobody does, nobody did. technically this is correct and truthful, but harmful ill-intent against user pulling power to corporate entity as much as it can.

same here, apple would "technically" not lie, but annouce their products and build their API confusing to their advantage (upselling features that do not exit).

when you have to fight over word definitions, it is a red flag.