why, when it can actually be helpful? Using chatgpt for searches is already much faster than denying 10 cookie banners. Some more privacy focused AI would be great, if firefox does that.
You miss the point. This takes away from them actually building and maintaining a browser. If they want to do this as a plug in or something, that's fine. I do not want it and am not interested in anything AI related being built into the browser. Its that toxic, VC mindset around everything having to be everything and following the hype of the moment. Its dumb and the announcement was a redflag warning.
Why should it be relegated to a plug-in? It is a feature everyone would find useful because no-one speaks every language. Also, since Chrome has this feature, new users would expect to have it work without having to research which plug-in to use.
You might not want to use it, but some people don't use Firefox bookmarks and you don't hear them demanding that bookmarks be moved to a plug-in. It's been a very long time since a browser was solely an HTML renderer, and while people were also against CSS and scripting at first, we've moved on.
"AI" has been used for many things for many years. The fact that the news is full of machine learning and generative AI doesn't mean that it's sensible to condemn anything using it.
“AI” has been used for many things for many years. The fact that the news is full of machine learning and generative AI doesn’t mean that it’s sensible to condemn anything using it.
Maybe you should move onto Edge. I hear they are big on AI in the browser.
Nope, I'm with Tropical here
Long time Firefox user, but if they start shoving AI stuff into it, I'm out
Servo is looking kinda fine there, might as well make my own browser
@joewilliams007@SvensKia@TropicalDingdong@firefox I disable searching in the browser. When I wish to search, I go to a specific search engine web page.
When users have to work to DISABLE stuff in a browser, it reduces the trust people have in the browser. Consider why people have been switching back to Firefox, and now you are making the same mistakes that drove people away fro their former browsers.
Are you saying you've disabled searching from the address bar and instead load up whatever.com and then type your search into there? I don't understand what you think you're gaining by enforcing this extra step.
I have some genuine hope for this plus the semantic web. Have quick general answers be answered by the LLM, and use it to also generate vector (or a knowledge graph from wikidata) results of the other content on the internet so that if want to dig deeper it can ingest a specific sources data (or route to an models with that info already embedded) or just return it to you for your personal reading.
Pretty exciting tbh, and hopefully all open source, open data, on local or distributed systems!
At least all the pieces are moving to make that possible.
I cannot stress this enough: LLMs are not, have never been, and quite likely never will be search engines. You may as well ask your a auto-complete questions.
I figured, but this thread wasn't specifically about helping with code. It was about replacing search engines. Common uses of search engines include finding information about ongoing events, e.g. the COVID pandemic or an upcoming election. Your comment implied that you thought this information is better obtained from ChatGPT than from Google.
Thing is you don't know if the answer is good. The answer looks good, yes, but how could you tell if ChatGPT just pulled the thing you looked for out of it's digital behind?
Exactly, without knowing the source(s) you can’t determine the quality of the answer. Although in saying that, I imagine people who use ChatGPT as a search engine might also not be super critical of search results.
But LLM can and do use search engines and create summary of the search with references to actual pages, where they found this information. ChatGPT saves tons of time by doing search for you. Want to double check? Go and click the reference. But so far it was very reliable in my experience.
Are you informed about what they're using AI for? One example is in-browser translations, which allows it to work offline and be privacy-respecting (no calls to Google, etc).
@Deebster@TropicalDingdong@firefox Why is translation marketed as AI?
You are hurting Firefox's image by trying to label raw logic as AI which a buzzword accompanied by a lot more baggage you don't want to have.
They are the ones who made the announcement and I could give two wiffs of stinky piss about their image. I only care if they make a good product that serves my needs. The microsecond their product ceases to to be a 'good' product from my personal viewpoint, I'm dropping them and calling them out on it. They've got the best browser in the game rn (imo). Its their game to lose but it looks like they are starting to make some dumb ass choices.
@jfmezei@Deebster@TropicalDingdong@firefox Maybe because Mozilla and their partners actually did use machine learning to build the model that powers the local translation feature?
The wide public has accepted calling all LLMs as "AI". LLMs are probably the best tool to create quality, native-sounding translators. Since LLMs are called "AI", modern translation engines which are made using LLMs will be called AI.
The other guy is just being a prescriptivist with language. It's a sentiment I sympathize with and which others have made very coherently, but at some point we have to just accept that the "buzzword" is the way it is going to get used and stop getting bent out of shape about it.
If they want to build extensions that do those things, I fully support them.
I don't need or want my engineering team building my browser doing things other than building my browser. I want them working on the browser. I need AI I didn't ask for in my browser like I need an additional hole in my head.
I want my browser to be a browser and if Firefox isn't focused on just building a great browser, I'm leaving.
I don't really know any real alternative if you value privacy and the openness of the web?
There are 3 companies making (as in putting the real work into it) browsers, Apple Google and Mozilla. Apple is only serving its own interests, and Google has clearly indicated where they want the web to go (more ads, less blockers, more tracking).
There are plenty of forks of Firefox, many of which will likely not follow Mozilla’s lead when it comes to adding non-essential functionality to the core product.
Ok, but you're still using firefox under the hood. Same story as the chrome reskins: you are beholden to what upstream does. You can go around limitations and do stuff differently, but pretty sure it's not the forks that will keep eg. the extension manifest V2 alive in chrome to be able to have a real ad blocker.
Every fork has a reason it exists, which is to do something specific differently from the upstream. There's almost guaranteed going to be a fork whose whole shtick is going to be “no built-in AI”.
In 2016, Opera, developed in Norway, became a subsidiary of an investment group led by a Chinese consortium.
The company released a gaming-oriented version of the browser, Opera GX, in 2019, and a blockchain-focused Opera Crypto Browser into public beta in January 2022. In April 2023, the company announced a major overhaul to the browser called Opera 100 and code-named "Opera One", adding a new user interface and several AI-related features.
Opera users also have access to Opera News, a news app based on an AI platform.
Add comment