@deadbeef79000 I'm going to bookmark this "conversation" so whenever someone asks me why I don't have serious conversations with AI bros, I can point a to perfect example. Thanks!

@deadbeef79000 I genuinely don't get your train of thought, dude.

First you say you use GenAI. Okay.
Then when provided with a counter-argument, you refuse to read past the first sentence because you didn't like the tone. Okay.
Then when it's pointed out to you that doing so is kind of a d!ck move, you resort to childishly parroting what that person said. Okay?
Then when that person calls you out for said childish behavior, you double-down and unironically say, "I know what you are, but what am I?'" a phrase known to be said by childish assholes. Okay???
Then you for some reason just go, "Your mother," as if that makes any logical sense as any counter-argument or anything logically coherent to the conversation. What???
And now you're just quoting what I said.

What the fuck is your goal? I can keep going, dude. You resorting to the most childish, trite, and quite frankly, troll arguments makes you look like a complete idiot that no-one's going to take seriously. All you're doing is proving that GenAI users aren't reasonable and act like children when you try to have a conversation with them. Is that what you want?

@deadbeef79000 Like, what are you hoping to achieve? All you did was make yourself look like an unreasonable, childish asshole. Like, if that's what you wanted, congrats, you did it.

@deadbeef79000 Yep. Totally helping your case by acting like a child who can't participate in a discussion.

@deadbeef79000 Yep, thinking you sound smart by parroting what I said totally does not make you look like a smug asshole, buddy.

@deadbeef79000 Sounds like a you problem, buddy.

@deadbeef79000 You... you know these things can make up their sources, right? LLMs are notorious for making sh!t up, and it's gotten worse in later models with seemingly little the companies running this tech can do because it's more of a feature than a bug of LLMs.

They weren't made to be search engines, they're not, they're as I said, probabilistic mimicry machines. When you use it like a search engine, you're only querying its database of scraped stuff with a good chance it'll pull something out of its ass to try and give you the answer you want to hear, not actually searching the internet. These models aren't built to be accurate, they're built to be plausible. They generate what the algorithm deems what you want to hear, not the truth.

So not only are you using a machine as a search engine that was not built to be one and can generate random bullsh!t, you're also consuming a lot more energy and power to do so.

The only reason people are trying to use LLMs is as you said, to bypass Google's deterioration (or at least one of its many factors of deterioration), which is an issue that's only cropped up in the last few years. If Google didn't enshittify, we wouldn't be in this mess, and using LLMs is not going to fix that, especially as these LLMs inevitably enshittify themselves since LLMs are super unprofitable.

@null_dot Even if that were true, that doesn't consider the fact that generative-AI is a money black hole and literally the only company type profiting from this bubble are ones like Nvidia that are producing the graphics cards these models use. Generative-AI is not and will not be profitable; Mozilla is losing money by shilling this tech.

So even if you do attract a high volume of users by shoving in generative-AI (which I DOUBT), I doubt that would offset the money burning from running these models. It's not a smart business move, either for users who despise the tech or users who love it. It's a lose-lose.

@null_dot If all of the not Firefox users are AI-shills, then I don't think we're missing much from excluding them.

And what value? Wasting a load of water just for the chatbot to tell you that "strawberry" has 2 r's in it? There is nothing generative-AI can do that other tech can't do better and costing far less resources.

The "value" of generative-AI outside of highly niche scenarios is an illusion. For example, people think generative-AI in the workplace improves productivity when it actually does the opposite: https://www.forbes.com/sites/carolinecastrillon/2025/10/02/ai-workslop-could-be-the-biggest-threat-to-productivity/

This is a good quote from that article: "A recent MIT Media Lab report found that 95% of organizations see no measurable return on their AI investments. Workslop helps explain why. When employees use AI to create low-effort output that shifts the burden downstream, any productivity gains are lost."

Generative-AI has also actively made search engines worse, especially with their "summaries" containing significant errors about 50% of the time: https://www.bbc.com/news/articles/c0m17d8827ko

And I can go on and on. There is an entire laundry list why generative is just... not it, and whatever "value" people get from these can be gotten elsewhere, usually in higher quality and more ethically than relying on this tech.

@pennomi Yeah... I highly doubt the models featured with this, well, feature are those kinds of models. The vast, vast majority of users don't use those kinds of models.

@jahtnamas Doing that now, thanks!