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

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

thatprivacyguy.com

Post reply on HN
Print view View original

951–960 of 1001 posts

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

#951

Post by bjelkeman-again »

McGlockenshire wrote: post #733
ddtaylor wrote: post #579

Earlier quoted context omitted.

The problem is that some of us are still on connections that charge per GB in rural areas. Here in Montana it's very common to pay about $0.25 per GB regardless of how much you use, so this is a $1 additional cost per desktop device. Places like public school districts have hundreds of computers and this will be somewhat significant for them.

Google's updater service also currently ignores the windows 11 metered connection hint. It will gladly download that model over your cell connection even if you have a data cap. This is infuriating behavior. Silicon Valley must wake up and understand the entire world does not live like them.

They live in a bubble and not a lot of the surrounding world makes it in to them. I know it is hyperbolic, but I lived there for a while and I stand by that opinion.

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

#952

Post by arendtio »

scriptsmith wrote: post #31

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…

/dev/mapper/vg_system-arch 207G 192G 4,7G 98% /

Just don't keep free space around :-D

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

#953

Post by ainiriand »

lobito25 wrote: post #760
ainiriand wrote: post #327

Earlier quoted context omitted.

Sometimes I marvel at how nice it would be to have such a narrow view of the world and other's perspectives and contexts. Life would be so much easier!

Ironically, your comment seems to perfectly reflect the viewpoint you're criticizing.

In my humble opinion, the parent of my comment seems to paint a picture that is quite drastic. I see why many people would have Chrome installed and I do not think something like that should be impose on to anyone.

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

#954

Post by arendtio »

IG_Semmelweiss wrote: post #894
bean469 wrote: post #824

Earlier quoted context omitted.

> most alternatives, until recently, were not great Curious to know, what viable alternatives did you discover?

Gmaps WV while using VPN. Download from Fdroid. Its miles ahead of something like Osmand , which i really, really, tried to like for a year, but its a UX disaster (i could never ever figure it out)

Better use it now, because it has a best-before date of 2026/2027 (depending on when Google will enforce the developer identity registration).

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

#955

Post by hzwanip »

scriptsmith wrote: post #31

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…

I think it's great, LFG Chromium OSS

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

#957

Post by alsetmusic »

cachius wrote: post #896
anigbrowl wrote: post #785

Earlier quoted context omitted.

You just described 95% of the parts of all software, especially in this era. Yes, that's the problem

The problem here is that the on-device model is old news packed as clickbait without any research beyond his file system. https://news.ycombinator.com/item?id=48034889 And all news outlets spreading it w/o any further research of their own. Policy GenAILocalFoundationalModelSettings disables and removes the local model without any flag hacks since 2024. In Canary since January behind Settings > System > On-device AI…

> The article doesn't mention Chrome version, release channel, whether on fresh vs existing install an if settings were altered.

Actually, it does claim this is the stable release channel. And it's reasonable to assume that if the author is documenting on a fresh profile / user account on the Mac, they probably downloaded the current (that day) release, though we're inferring when don't don't truly know for sure.

> 10. Code-signed, shipped through the normal release channel. This is not test build behaviour. It is Chrome stable.

I think it's poor form to run cover for one of the biggest corporations in the world like this. Don't let them off the hook. As the author correctly points out, metered connections are being abused. Hell, last month I somehow hit ATT's bandwidth limits on my mobile and got throttled for five days. It made my phone so unusable that I turned on hotspot on my work phone and connected to that over wifi when I went to lunch.

Post reply on HN