Google Chrome is downloading a 4 GB Gemini Nano model onto users' machines without consent, with no opt-in, no opt-out short of enterprise tooling, and an automatic re-download every time the user deletes it. The pattern is identical to the Anthropic Claude Desktop case I wrote about last month, but the scale is between two and three orders of magnitude larger. This article does the legal analysis and, for the first time, the environmental analysis. The numbers are not small.
AI runs in the cloud because it needs a powerful server to run the biggest (i.e. “smartest”) models.
The cloud servers are doing nothing special that another powerful enough computer could do, just a huge amount of data processing.
You can run an ai chat on a steam deck or directly on a phone, if it’s not too demanding (“smarter” models are bigger data files, so won’t fit in the memory of a small device).
Today, for instance, I had a phone call from “Spectrum Internet support” and part-way through the call my phone blared an alarm and said “possible scam” on screen.
The phone itself interpreted the conversation as sus.
The cloud being a bunch of computational power (servers). A bunch of phones in a network also can be utilized for said computational power. Passing the savings on to you! ;)
What I want to see is throttleable models, kind of like progressive JPEG, where the default model is “nano” and it has a watch function that analyzes if more tokens might be needed for a certain task and scales up as needed — identifying if the resources are too much for the device and offloading to the cloud (with explicit permission) only if (but always if) needed. Over time as the technology improves, larger models move to the endpoint.
And then people could have a basic set of sliders: on-device only, on-cloud only, or somewhere in between, based on the user’s preferences.
That’s basically model routing, and has existed a while. Open AI’s GPT-5 and llama-swap do that, for example. If the task is simple, it uses a smaller, less intensive model, and only uses the slower, larger one of the task is more complex.
Though most tend to operate with models on the same device/service, rather than a model run elsewhere.
Yeah, even there. A page loading is one thing, but browser features are somewhat independent of the content. There’s also a good chance this is being used as a hook for other Google products like Drive or Docs (which are basically websites under the hood) to allow offline file management, creation, etc.
It’s a bad choice, but it wouldn’t be the first bad choice Google has made.
Can someone ELI5 why they are doing this? I thought all the AI shit was in the cloud?
AI runs in the cloud because it needs a powerful server to run the biggest (i.e. “smartest”) models.
The cloud servers are doing nothing special that another powerful enough computer could do, just a huge amount of data processing.
You can run an ai chat on a steam deck or directly on a phone, if it’s not too demanding (“smarter” models are bigger data files, so won’t fit in the memory of a small device).
Today, for instance, I had a phone call from “Spectrum Internet support” and part-way through the call my phone blared an alarm and said “possible scam” on screen.
The phone itself interpreted the conversation as sus.
https://support.google.com/phoneapp/answer/15654065?hl=en
The cloud being a bunch of computational power (servers). A bunch of phones in a network also can be utilized for said computational power. Passing the savings on to you! ;)
I recall this. Pretty much the same idea, but this time it is opt-out(or is it actually?)
They need their features to work offline too probably.
It’s also cheaper, if they can offload a portion to the user’s computer.
Cheaper for them, that is.
What I want to see is throttleable models, kind of like progressive JPEG, where the default model is “nano” and it has a watch function that analyzes if more tokens might be needed for a certain task and scales up as needed — identifying if the resources are too much for the device and offloading to the cloud (with explicit permission) only if (but always if) needed. Over time as the technology improves, larger models move to the endpoint.
And then people could have a basic set of sliders: on-device only, on-cloud only, or somewhere in between, based on the user’s preferences.
That’s basically model routing, and has existed a while. Open AI’s GPT-5 and llama-swap do that, for example. If the task is simple, it uses a smaller, less intensive model, and only uses the slower, larger one of the task is more complex.
Though most tend to operate with models on the same device/service, rather than a model run elsewhere.
In an internet browser?
Yeah, even there. A page loading is one thing, but browser features are somewhat independent of the content. There’s also a good chance this is being used as a hook for other Google products like Drive or Docs (which are basically websites under the hood) to allow offline file management, creation, etc.
It’s a bad choice, but it wouldn’t be the first bad choice Google has made.
Well everything else is in it.
Shit, Chrome supports the use of COM ports. It’s an OS within an OS.