Technology

Not working for me, OpenAI still exists :D

Been saying that since the start, AI kill switch should be a standart in every Application. Did that first long ago in HugstonOne, AI intelligence and speed in my own terms.

Mozilla has released so many self-described AI features in the past few years, but this is the only one that has:

• been requested by the community
• received broad critical acclaim

I hope Mozilla learns their lesson. I doubt they will, but I hope.

"Kill switch" – oh the drama. Let's call every simple toggle 'kill switch' from now on.

Browsers don't need LLMs

I personally don't HATE ai but I don't want it in my browser or email or anything like that. I have a local llm I use for random stuff all the time but I don't need or want a company viewing everything I'm doing, adding buttons in places I'm likely to accidentally push, or training their shit on my dumb behavior. ai has destroyed much of the Internet already to the point that you almost need to use an llm in order to get any useful information during a search. Otherwise you're just filtering through ai generated webpages with the highest seo possible.

I've already switched over to LibreWolf a month or two ago. Clean, simple, and it just works.

all you have to do is click on Settings > AI Controls. You'll then see a very bold and prominent option called 'Block AI Enhancements.'

I don't see it on mobile though.

Remember when they had a "kill switch" for javascript?

The only useful thing ive found for AI is its ability to read text from an image. Which is good for taking serial numbers from a photo, and copying from an app that otherwise doesnt allow copying on phone. Thats it. A tool.

And immediately blocked.

I'm not against AI, I use it, but I want to be using it on my terms, not have it shoved into everything I use.

Step 1. Add AI. Step 2. Add (broken) switch. Step 3. Pretend to fix switch. Step 4. Hide switch in sub-menus. Step 5. Remove switch.

... And all they actually need to do is make "AI" an extension. Let the users install it if they want to, or don't. That's the whole point of extensions. But they would never dream of that, hell no.

Also, the kill switch does not fully remove the AI slop. Remember to uncheck perplexity from the search engine list, and also uncheck AI suggested tab group name.

The Translation feature seems to be classified under AI. Idk what technology does it actually use, but it's done locally on device

@avidamoeba@lemmy.ca @technology@lemmy.world

The problem still remains: why's this thing "opt-out" and not "opt-in"? Why not make it an official, totally optional (as in voluntarily wanting to have it and, only then, proceeding to have it) plug-in or extension that the user (let us remember the meaning of "User Agent": an agent acting on behalf of the user, not a piece of software who's become "the user") could install at any moment, out of their own will?

I'm far from being an anti-AI person, I myself use those clankers on a daily basis. However, I use them because I want to, while I still want to, not because they were pushed unto me.

Mechanisms of "opt-out" where there should be an "opt-in" is a form of dark pattern.

In fact, the very concept of "opting-out" is a dark pattern per se, because it implies something pushed unto a person, something from which they were "allowed" the "right to leave".

Yeah, it's awesome to have means of "opting-out" from something, but having an "opt-out" mechanism in place doesn't mitigate the very fact that it was coercively pushed unto the person beforehand and didn't require explicit consent from the person unto which the thing was pushed.

Speaking of "consent", situations like these are not that much different from the dark pattern "Yes / Not now" we've been seen everywhere: in certain scenarious, this insistence and disregard for explicit consent would verge the criminal (e.g. harassment), but suddenly it's "okay" when corporations (and the State itself) do it.

If, say, a situation where someone is being harassed and, only after having started to harass, the harasser offers the harassed a means to leave the harassment, does this make the harasser less of a harasser? Because that's the same absurd logic behind the corporate advocacy whenever it's said "oh, but Mozilla is offering an opt-out, you can always turn off 'sponsored shortcuts' (that is, after having been faced by the shortcut from a Jeff Bezos corp as you proceeded to open a new tab for accessing the opting-out settings, but that's totally okay), 'sponsored wallpapers', and the 'Anonym tracking', and now you can, check this out, you can turn off the clankers, too! Wow, isn't that such a cute corp, the corp with the cute fiery fox mascot?".

Not to say how it's gonna end up cluttering the upstream with (more) binary blobs, adding to the Sisyphean struggle that WaterFox, IronFox, LibreWolf, Fennec, among other Firefox forks, have been experiencing upon trying to de-enshittificate the enshittificated and de-combobulate the combobulated.

"Mozilla needs to make money". Yeah, yeah, because the very fundamental, immutable principle of cosmic existence boils down to "there's no such thing as a free lunch", amirite? After all, "money" is clearly within the table of elementary particles alongside quarks and gluons, isn't it? And Mozilla needs to make money... We had a tool for that: it's called donations.

So, there's a "bug", though I expect to FF it's a feature: If you individually block all of the AI features, then click on the master switch to block all AI, everything's great. But if you revert that master switch suddenly it "forgets" all of your settings and shit is activated again.

It seems by design. And since it's opt in, if FF "accidentally" disables the master switch (I'm betting it will eventually) you lose that extra layer of protection. OH, and I had disabled EVERYTHING in registry (about:config) before this and translations were still available. I guess it's time for me to explore other FF-core options....