Apple wants to train AI on your private personal data | Hacker News Reader

Apple wants to train AI on your private personal data

machinelearning.apple.com

30 pointscroes19 comments

Changed again?

"We do not use our users’ private personal data or user interactions when training our foundation models. We also respect the rights of web publishers to opt out of foundation model training."

Further down. Point 4 under Responsible AI

Your private personal data and interactions are never used to train our foundation models unless you explicitly choose to help improve them.

That's practically the opposite of what your editorialized submission title says.

No, it’s exactly what I wrote. They want to train AI on your data. I didn’t say anything about without consent or secretly but it’s a change to their previous stance where they said, they don’t do it.

Now you‘re one click away to allow them to use your data

Title seems misleading if not outright wrong, having read the article.

Where is is misleading?

They changed it from we don’t do it to we want to do it.

Now you‘re one click (or maybe two if they ask for confirmation of your choice) away to let them use your data.

Regarding the architecture:

> Instead of forcing the entire model into DRAM, the full model is stored in flash memory (NAND). Because NAND-to-DRAM bandwidth is too slow to swap weights token by token, as standard MoE models require, AFM 3 Core Advanced makes routing decisions per prompt. A lightweight, dense block selects a fixed set of experts during initial processing, periodically reselecting them during generation. To minimize data movement, the model relies on a high percentage of always-active “shared experts” alongside input-dependent “routed experts” swapped into DRAM only when needed.

This is an interesting hybrid between MoE and managing entirely separate domain-specific models. Select the experts once, bring them into memory, and run inference for some period of time before re-evaluating. Saves having all experts in memory, but it's better than just selecting a whole model per query since you have a high number of small opaque experts that overlap and combine in interesting ways.

There is a probably a massive quality hit to doing this but it's interesting because it allows infinite scaling of model size.

They changed

> We do not use our users’ private personal data or user interactions when training our foundation models.

https://web.archive.org/web/20260829051311/https://machinele...

> Your private personal data and interactions are never used to train our foundation models unless you explicitly choose to help improve them.

https://machinelearning.apple.com/research/introducing-third...

> unless you explicitly choose to help improve them.

Is it opt-in or opt-out? Big difference.

Opt-in = explicit and opt-out = implicit. You can't have implicit opt-in, that wouldn't make any sense.

You can't have implicit opt-out either. Opting is an explicit act.

Explicitly choosing to help improve something by providing your data seems like an opt-in but what do I know these days

So far Apple has been using default-checked checkbox during onboarding process for OS-level “help us improve…” diagnostics collection. Not sure about this specifically.

I don’t think that counts as an automatic opt-in, given that it’s presented to the user directly during the setup process.

It means my mother will have it on, at least, if they use the same pattern.

The sentence prior states: "Privacy is the default, not something you have to manage."

It would be nice if we were somehow able to switch AIs on our local systems, ie., replace the Apple "brain" with some other "brain". I doubt we will be able to do that, but it would be nice.

what a coincidence, i thought we were going to slow down training (amodei, altman et al.)

Apple keeps taking potshots at their own feet.

What Apple consumers want from Apple is a premium product without all the enshittification found everywhere else. I don't know what Apple thinks it's doing by introducing so many ads into everything (maps now!) and shuffling their feet backwards on all their privacy positions.

In a world where every company steals all my data equally, and is ridden with the same crappy ads, why on Earth would I pay a hefty premium to Apple? It's just so stupid and short-sighted in terms of product differentiation.

Reply on news.ycombinator.com