Live data from Hacker News

Google Chrome silently installs a 4 GB AI model on your device without consent

thatprivacyguy.com

471–480 of 1001 posts

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#471
post #463

Earlier quoted context omitted.

Yes, which is why I use paid-for OSes and browsers, instead of free ones like Linux or Firefox. I don't want to be the product.

Why is nobody building a paid for browser with built in search engine and LLM assistant? Should probably make it open source for transparency. And before anyone says you would build/compile it yourself if it was open source, those ppl are already running their self compiled tools and are not the target market.

Who would pay for it when literally every other option has been free for a lifetime?

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#472
post #423

If Chrome has the #optimization-guide-on-device-model and #prompt-api-for-gemini-nano flags enabled, either because it's part of some Origin Trial / Early Stable Release or something, then web pages will have access to the new Prompt API which allows any webpage to initiate the (one-time) download of the ~2.7 GiB CPU or ~4.0 GiB GPU model using LanguageModel.create() https://developer.chrome.com/docs/ai/prompt-api Wh…

The more severe problem is that Google installs model weight files on a per-user basis, meaning Chrome occupies 4 more GB of space for every OS user on your device.

4GB, $0.10 (whatever the HD price) that is the equivalent of a High School level intelligent brain that can perform many cognitive tasks (and in the future even PhD level intelligence) for free?

Oh, the horror!!!

Wait, let me pay my HVAC guy $500 he deserved because he came all the way from his home to replace a fuse

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#473

"Silently installs" is misleading. They are including a file in the package which is presumably related to the functionality of the software. I don't use chrome for a long list of reasons but it is not standard or expected to get consent for that.

Then what is your definition of "installing" exactly? Are you going to split hairs about it not being a separate program being installed and running in the background, but weights being used by code that is run inside the browser? Because honestly, I don't think there's any significant difference from the user's perspective here. Other than the fact that doing the latter bypasses the need to get permission to install…

> A 4 GiB model has nothing to do with the functionality of a web browser.

Problem is, nor do half its T&Cs. What we thought was a web browser turns out to be a Google content delivery vehicle - and controlled by Google, not the target users.

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#474
This might be worth it if Gemma4 E2B were a good model, but honestly it's absolutely useless in all our testing without further training and finetuning, and those aren't usecases that are fit for normal web browser use such that one would care to support it by adding such overly broad and expensive infrastructure to make it happen.

Gemma 4 E4B is a much better model, but it's too large to simply download and run everywhere.

IMHO, this is jumping the gun. Google's going through a lot of effort to release a model that will give everyone a very poor first impression of what on-device models are capable of, souring it for everyone for a long time afterwards. It would be better to wait until a smaller, better model ships before doing this.

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#477
post #234

How hard would have been to add a simple message, warning people about it and offering to opt out? Most would have clicked OK without reading anyway, and Google could pretend they give a shit about users. Unless they expected blowback, and that kind of message is the "compromise" they want to eventually land on.

I think a local AI model is appreciated, but it being bundled and executed through Chrome, I expect that more or less all data get exfiltrated by Google. They simply read your mails, how would you expect there to be anything resembling decency in a company like that? It is the ad business. Bad thing is that people still use gmail.

Doesn't google already extract every single letter you put into the address bar as you type?

The idea that the model is local is just Privacy Washing. What's the chance they aren't capturing your prompts somehow? For "Telemetry" so they can triage bugs of course!

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#478

Earlier quoted context omitted.

0.04 to 0.1 kWh/GB is insane even for 2018 lol. I have gigabit internet (125 MB/s). This would imply when I'm downloading something I'm using 18 to 45 kW of electricity. Completely bonkers.

1.8 to 4.5 kW.

This was my math:

0.1 kWh * 3600s/hr = 360 kJ

360 kJ / 8s (time for 1 GB) = 45 kW

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#479
post #452

If Chrome has the #optimization-guide-on-device-model and #prompt-api-for-gemini-nano flags enabled, either because it's part of some Origin Trial / Early Stable Release or something, then web pages will have access to the new Prompt API which allows any webpage to initiate the (one-time) download of the ~2.7 GiB CPU or ~4.0 GiB GPU model using LanguageModel.create() https://developer.chrome.com/docs/ai/prompt-api Wh…

You can already trigger a 2 GB model download with the Summarizer API[0], which is already shipped in Chrome. Summarizer.create() [0]: https://developer.chrome.com/docs/ai/summarizer-api#model-do... I think this is a distinct model from the Prompt API, since the other shipped AI APIs use fine tuned models.

So now we're up to 6 GB

Re: Google Chrome silently installs a 4 GB AI model on your device without consent

#480

If Chrome has the #optimization-guide-on-device-model and #prompt-api-for-gemini-nano flags enabled, either because it's part of some Origin Trial / Early Stable Release or something, then web pages will have access to the new Prompt API which allows any webpage to initiate the (one-time) download of the ~2.7 GiB CPU or ~4.0 GiB GPU model using LanguageModel.create() https://developer.chrome.com/docs/ai/prompt-api Wh…

[dead]
Post reply on HN