Don’t get me wrong I hate LLMs being shoved down our throats, but I think the “Mozilla using AI” stuff is overblown. The few LLM features they’ve implemented are useful and non-intrusive. I actually think it’s a rare example of the tech being used intelligently (no pun intended).
I’ve been distro hopping lately, and I’ve discovered that if you don’t turn the sidebar off before you log into your Mozilla account, it will turn the sidebar back on on all of your other Firefoxes (as well as other settings like the studies).
I could not disagree more. a browser is a local piece of software to me that I use to connect to a variety of servers. It having an ai server its communicating with all the time and with access outside the add on system is not something I want. I have looked at and was not put off by other firefox “controversies” but this one is making it where it is not simply a browser anymore. which is what I want.
this does not make me happier. I can turn it off but if im going to run a local llm I rather do that as its own piece of software. Its funny because I have mentioned I would like to see a linux operating system using a local llm where its acting a bit like the change from text based to gui based operating systems. I don’t want every piece of software operating a local llm to do its job or a part of its job on my computer. Its great I can turn it off but I don’t want to be downloading an llm model with every piece of software and just turn it off for my browser and my office suite and anything else im running. Its actually worse with the browser that already has taken over so much of the application space as a kind of universal tool.
I wonder how long before someone says “Having to wait for security patches to be made only when new exploits are discovered is inefficient, we need an AI agent running all the time, inspecting every command, to respond immediately if an exploit is found.” and then they just stop security patches and you need that AI crap consuming ram and energy like hell while sending all your data to the mothership.
Don’t get me wrong I hate LLMs being shoved down our throats, but I think the “Mozilla using AI” stuff is overblown. The few LLM features they’ve implemented are useful and non-intrusive. I actually think it’s a rare example of the tech being used intelligently (no pun intended).
These things have been incredibly intrusive. They even added a whole sidebar to take up more space like they were fucking Yahoo and this was IE.
Weird, I don’t have a new AI sidebar and I’m running the latest version.
I’ve been distro hopping lately, and I’ve discovered that if you don’t turn the sidebar off before you log into your Mozilla account, it will turn the sidebar back on on all of your other Firefoxes (as well as other settings like the studies).
I could not disagree more. a browser is a local piece of software to me that I use to connect to a variety of servers. It having an ai server its communicating with all the time and with access outside the add on system is not something I want. I have looked at and was not put off by other firefox “controversies” but this one is making it where it is not simply a browser anymore. which is what I want.
Well you’re in luck because the translation feature is local https://support.mozilla.org/en-US/kb/website-translation
this does not make me happier. I can turn it off but if im going to run a local llm I rather do that as its own piece of software. Its funny because I have mentioned I would like to see a linux operating system using a local llm where its acting a bit like the change from text based to gui based operating systems. I don’t want every piece of software operating a local llm to do its job or a part of its job on my computer. Its great I can turn it off but I don’t want to be downloading an llm model with every piece of software and just turn it off for my browser and my office suite and anything else im running. Its actually worse with the browser that already has taken over so much of the application space as a kind of universal tool.
I wonder how long before someone says “Having to wait for security patches to be made only when new exploits are discovered is inefficient, we need an AI agent running all the time, inspecting every command, to respond immediately if an exploit is found.” and then they just stop security patches and you need that AI crap consuming ram and energy like hell while sending all your data to the mothership.