261–270 of 782 posts
Does anyone want AI in anything? I can see the value of navigating to an LLM and asking specific questions, but generally speaking I don't want that just running / waiting on my machine as I open a variety of applications. It's a huge waste of resources and for most normal people is an edge case.
I also wouldn’t want to go back to only web search for finding things out. Search engines are generally inferior.
Heck yes, we want jt, can't imagine browsing without it. It's like going from YouTube to Tiktok, for most content we consume, you could cut 90% of it without losing anything of value.
Earlier quoted context omitted.
And will that work permanently, or will I have to hunt down another setting in another month when they stuff it into another workflow I don't want it in?
Yeah, if YouTube Shorts or Games are any indication, it'll be back soon! The AI Mode in Google Search comes up nearly every time I use it no matter how many times I hit "No"
Earlier quoted context omitted.
In firefox yeah! I use it often. I have it connected to a local Gemma model running in ollama and use it to quickly summarize webpages, nobody really wants to read 15 minutes worth of personal anecdotes before getting to that one paragraph that actually has relevant information, and for finding information within a page, kinda like ctrl-f on steroids. The machine is sitting there anyway and the extra cost in electric…
The default AI integration doesn't seem to support this. The only thing I could find that does is called PageAssist, and it's a third-party extension. Is that what you're using? https://addons.mozilla.org/en-US/firefox/addon/page-assist/
So grab ollama and your prefered model, install openwebui.
Then open about:config
And set browser.ml.chat.provider to your local openwebui instance
Google suggests the you might also need to set browser.ml.chat.hideLocalhost to false. But i dont remember having to do that
Does anyone want AI in anything? I can see the value of navigating to an LLM and asking specific questions, but generally speaking I don't want that just running / waiting on my machine as I open a variety of applications. It's a huge waste of resources and for most normal people is an edge case.
Unfortunately got to meet those KPIs.
I think "nobody i know" and nobody are different things. Mozilla wants more firefox users. there are "AI browsers" and ai integrated browsing is becoming more and more the norm. Mozilla is doing the right thing here, the features are there but unobtrusive. But down the road, I fully expect Mozilla to do whatever they have to do to remain in the game. Their small market share is hurting the entire internet, they can't…
Not really, outside influencers looking to capture the next hot thing (like Mozilla) and tech-bros, there is no living soul on this planet that wants or is trying to normalise AI browsers.
In particular I'd also love agentic AI so I can quickly automate tasks on shitty web sites that can't be reasonably automated otherwise.
But even a free, no-signup "summarize this wall of text" would be useful.
I think the adoption of AI browsers shows that there are people who find value in this, and I think a lot more people would be interested if it wasn't getting relentlessly forced on them at every corner, making them refuse it out of principle.
Earlier quoted context omitted.
I use Whisper running locally for automated transcription of many hours of audio on a daily basis. For the most part, Whisper does much better than stuff I've tried in the past like Vosk. That said, it makes a somewhat annoying error that I never really experienced with others. When the audio is low quality for a moment, it might misinterpret a word. That's fine, any speech recognition system will do that. The proble…
Try Parakeet, it's more state of the art these days. There are others too like Meta's omnilingual one.
You don't actually state whether you believe Parakeet is susceptible to the same class of mistakes...