Try Orbit by Mozilla: a new AI productivity tool

[sarath]

Hi folks,

We’ve got something exciting to share first with our community! Introducing our beta version of Orbit by Mozilla, now available as a Firefox add-on. This is a new AI tool to help summarize emails, docs, articles, and even video transcripts.

Designed to boost productivity, Orbit provides concise summaries right in your browser. Whether you’re wading through long emails, dense documents, or informative videos, Orbit helps you get to the heart of the content quickly.

Features include:

• Quick Summarization: Get summaries of what you’re reading or watching across sites like Gmail, Google Docs, YouTube, and more.
• User-Friendly: It’s simple to use – just add it to your browser and you’re good to go.
• Interactive Q&A: Ask questions about page content and quickly find specific information.
• Customizable Summaries: Choose your preferred summary length and format.
• No AI Training: Unlike other services, Orbit’s AI does not train on your data, keeping your information private.

We invite you to check it out and download Orbit. We’d love to hear your feedback once you try out Orbit, and we’re open to all your ideas and thoughts.

[doodlemancy]

someone must have offered them a LOT of money to hawk this garbage, more than they're afraid of losing in user support, and i don't think it's going to work out for them in the end lol

[Gabi]

Hallo malogan 😀 from Germany, I don't agree with your comment! You have spoken "very honest" and "too hard" in my opinion, because the Moz. Employees, Admins + developpers do a very good job in my opionion! 

It is a difficult time - all over the world and finding constructive solutions should be the main aim. I am a completly new member, from Germany and I use Firefox from the beginning - 2002! It is the same in Germany and Europe. Everything depends on market share. In Germany 13 percent of the users use the Firefox Browser (as it is Gecko engine and has excellent data securities and transparancy)!
Don t forget this.

The future is (unfortunately) AI and really believe me, I reject it also - as many people on the globe.
But, rejecting it completly is not the solution, because then, there is REALLY NO ALTERNATIVE where transparancy is given. Therefore:
1.) Should somebody d' ont need an (ai) add on, the possibility must be given to remove it.
2.) Should somebody would like to have more data secure - I mean an alternative to i.e. chatgpt
or Copilot- then it would be given with "Orbit" or whatever.
I think, it is important to find solutions, how the Firefox Browser comes back " to its roots in 2002" (with settings....which can be done by the users) and at the same time goes with the time - finding data secure alternatives - perhaps with add ons and/or with strong partnerships (companies) in order to earn more money (developpers are not for free).
In addition , I think Mozilla is under great pressure - particularly in the US - to be a good alternative to the Global players like Microsoft-Open AI and Google, which definitely have much more money.

So, MALOGAN - think about it and perhaps you have good ideas for this problem? Then let us know as community! Everybody would appreciate it. Thks.

[doodlemancy]

Can you guys try reading the room? Or reading anything that isn't ChatGPT output? This is so pathetic. Seriously just sad that you're wasting time and resources on this scammy garbage. I guess I'm not donating to Mozilla or paying for Mozilla VPN anymore.

[BlazeHedgehog]

So you had a thread asking about feedback for an AI service in Firefox, and it garnered more than 3000 replies. Unfortunately, the Mozilla forums are also beholden to the gimmick of endless scrolling instead of the much smarter paginated view (for a forum especially), so it's actually harder than you'd think to look at all 3000+ replies.

But I decided to take at least a snapshot of who was for/against adding any AI features to Firefox at all, and my criteria was largely single replies to the original post, not factoring replies to replies.

I looked at 688 single replies, and out of those 688 single replies, 655 (95%) were a strong and clear "No" to any AI being added to Firefox in any capacity, while 33 (5%) were positive or at least neutral to its inclusion.

It really feels like you already made up your mind about including AI features in Firefox and the idea of letting users share their feedback was mostly performative, given you ignored pretty much everyone. The idea that this is an add-on you can either choose to install or ignore is moot; it's a defense against the overwhelming majority that told you not to do this and an excuse to do it anyway.

Who are you really serving here? Because it's not your userbase, that's for sure.

[doodlemancy]

yes. chiming in to say i also watched this thread closely. a lot of people responded with a lot of well-reasoned feedback. they linked sources and listed clear, provable reasons that they don't want AI features in Firefox. it was not just an unthinking reactionary response to the tech buzzword of the moment-- although "AI" is such a sham in general that no is a perfectly fine place to both start and stop.

the overwhelming majority of the comments in that thread were negative on AI. one of the rebuttals i saw was "well, some popular anti-AI people posted about this elsewhere so the feedback is skewed that direction" and like... okay, but that's still genuine feedback wherever it came from-- people heard Firefox was implementing AI garbage and they went to the trouble to show up, sign in or make an account, and tell Mozilla, in no uncertain terms, that that they didn't want "AI" anywhere near Firefox. the notion that this was an "inorganic" response and thus worthy of dismissal is, frankly, hilarious coming from the people who think that have successfully fooled themselves into thinking predictive text is intelligent. also: consumers are actively turned off by AI.

the staff stopped responding, but early responses from staff were in a tone of like... "ok, we're still doing this though-- how can i convince you to be less mad about it?" from the pro-AI side, i saw basically no attempt to address anyone's concerns, besides "trust me bro" on the privacy side and "well if you don't like it, you don't have to use it", which misses the point of the outcry entirely.

i'd like to bring up again that Mozilla stopped accepting bitcoin donations a few years ago over environmental concerns, back when NFTs were the hot new scam bandwagon, and they haven't even quietly reversed course on that yet.

AI presents not only environmental concerns, but serious ethical and privacy concerns. this has been brought up over and over. Orbit runs off mistral 7b, who won't disclose where their data comes from, meaning: it's probably unethically scraped just like everyone else's and they don't want to admit it. if you cannot tell me where the data to train the AI came from, you have no grounds to claim that it was trained ethically. so what's going on here? why is Mozilla not being ethically consistent here? what convinced you guys to go in on this garbage?

by the way: AI can't even summarize text correctly. even this "innocuous" and seemingly helpful feature that Orbit is supposedly going to do is a sham. i'm sorry, but in order to obtain information from text, you do, in fact, just have to read it. i have severe ADHD and i understand that this is a struggle. but you have to read, or at least have a trusted human summarize it for you. "AI" is literally making more work for you because you have to fact-check it anyway! would you rather read something for 20 minutes and have all the correct info the first time, or read something for 2 minutes and then repeatedly have to go back and unlearn false facts/fix mistakes you've made because Microsoft Sam fed you a bunch of misinformation? a stitch in time saves nine. ask chatGPT to explain that one to you, i guess.

and for god's sake, paginate long threads in your forum, it's a mess.

[AndyDroid10X]

Don't you guys think that people who use AI features and find them useful just use them and don't share their outrage because everything is alright?

[SoliTheFox]

Quick question: do you use any kind of ad blocker?

How ethical are ad blockers in terms of freely available knowledge? Especially for smaller websites.

How much does a server, domain and time does it cost to maintain a website up and running?

How are open models unethical and why does wget offer a recursive feature and why doesn't it get any hate? What about Heretrix from The Internet Archive? Is The Internet Archive ethical? Have you agreed on having snapshots of your website stored publicly?

I wonder what your opinion is about all of these topics.

[EcoHound]

Are you people that dense? That other AI thread has hundreds of individual comments outright saying NO. Myself included

[Gabi]

Hi everybody,
Summarized: Everybody is disappointed and angry about the AI-Extensions - and writes endless text.😲

But nobody considered about a reason ? and nobody has constructive Ideas to solve it for the worldwide community.

I am from Europe- Germany, I am really shocked about this and a better cooperation and understanding of economical reasons and pressure (through the global players) to the non-profit organization Mozilla within this community would be appreciated.

ACTS AND FIND Solutions.
In German we say: nicht quake - maake...! this is a joke

[PirosThe3rd]

Gabi, you say you don't understand why people are against AI - did you read the thread about the AI feedback at all? Many people explained it very very clearly.

[SoliTheFox]

I still don't quite see why adding more features is something bad really, especially if it is and optional feature.

I wonder how many of these people claiming "ethical high ground" actually don't use unblock origin or other ad blocks that are unethical towards people hosting websites. Servers are expensive as hell. Complete paywalls suck, and are unethical too, but stealing income from freely available knowledge is as much unethical as people keep arguing about "unethical scraping".

[lackey]

Adding features: can slow browser or introduce new places to break. Wastes time better spent elsewhere. 

Refusing to consent to seeing things you don't like: not unethical

Taking people's things without consent: unethical

Do two wrongs equal a right? No

I hope this clears things up for you. 

[SoliTheFox]

I mean, if adding an OPTIONAL feature, as an extension, slows the browser down, then you should buy a new PC.

If someone uses ad blockers on someone else's website, it is someone else's  website, it is their revenue you are taking away. If you use ad block you are taking people's things without consent, after all, the person is providing information freely as long as you consent to help him with the revenue by using ads. This is such a huge problem that Google is trying really hard to block people from using ad blockers. If this is not, by your own definition, unethical, then you are being hypocritical to say the least.

Hope this clears things up for you.

Also, i wonder what your opinion about The Internet Archive is, and more over, what your opinion about Heretrix (The Internet Archive crawler) is.

[lackey]

Are you against taking things without people's consent?

[SoliTheFox]

On the contrary. I use ad blockers. I believe in free knowledge, open science, open source, open access, whatever the source it comes from. Free of charge.

Being against GenAI without also being against The Internet Archive, scihub, zlib, Heretrix, recursive wget, and other websites or tools like these, that have been around for many years, is hypocritical to say the least.

[lackey]

Your list is not exhaustive: it extends to taking people's possessions and bodies without consent, and the people and corporations with the most power can take the most.

I don't want to live in your world. 

[SoliTheFox]

Sure mate, definitely. By saying i'm in favor of open science, open source, open knowledge, etc. and taking it from big techs and industries that want to put paywalls on stuff that should be freely available, i'm definitely saying i want to steal your gf, wife, mom, etc. I'm also taking your car if you don't take care...

Sure mate, good thinking. 

[lackey]

Mate, I asked you whether you were against taking things with other people's consent. You said you supported taking things without people's consent.

If you wanted to only support some taking of things, you should specify that, because in this thread everyone has been talking about a corporation that takes from people, and you are on the side of that corporation. 

[SoliTheFox]

"Being against GenAI without also being against The Internet Archive, scihub, zlib, Heretrix, recursive wget, and other websites or tools like these [...]"

"If you use ad block you are taking people's things without consent, after all, the person is providing information freely as long as you consent to help him with the revenue by using ads."

People these days are against GenAI but at the same time can't even read or understand a basic text.

Also, Orbit uses a Mistral model. Mistral is not a large corporation. https://mistral.ai/company/

[lackey]

"not a large corporation" is valued at billions of dollars. 

See, why did you even get upset about how I understood your answer, when I was right: you endorse the corporation taking from people without consent?

[wutongtaiwan]

Let's make an extension of nightly's AI function, don't build it into Firefox, it's too loud to oppose it. After the extension is made, it is placed in the "Recommended Extensions" section to facilitate users to install the extension

[technoxenoholic]

mozilla SHOULD NOT UNDER ANY CIRCUMSTANCES include generative ai as a base firefox feature. moving it into an optional addon instead IS a step up (you are removing it from the browser's base features, aren't you? you had better be). the fact that you promise not to train your ai on our data is also, obviously, a step up from most similar services. (assuming you actually mean it, that is.)

however, mozilla should also not be wasting time and resources on building a generative ai addon. there are, unfortunately, PLENTY of developers who can and will make a hundred generative ai addons themselves regardless of what firefox includes. stick to vetting those addons for whether or not they are upholding firefox's privacy standards (which are becoming very questionable, frankly, in the face of this insistence on endorsing generative ai), and don't expend the dev time on making your own.

[Thullin]

Neat but please don't make it mandatory like you just did with the list all tabs feature in the latest update.

[AndyDroid10X]

I really don't get all of that people. I guess people who are good with AI like me just use AI features happily, while others could just don't turn those features on, but instead they're commenting around that Firefox is not about AI, AI is bad, they collect their data etc. While I understand the concerns about privacy, I feel OK with that features as long as they're not turned on by default or work as a part of core features like in windows.

I love Orbit and I love that side panel, it's just useful.

[Erwan]

I haven't tried Orbit, but all my interactions with an AI on the internet (help service, support chatbot, etc.) ended with: "how do I talk to a real human?"

Needless to say, AI chats irritate me to no end.

Question 1: How is Orbit hosted? what are the energy and cooling costs? How much physical space does it take up?

Question 2: Being a member of an NGO with a lot of pressure from the State or other large structures, can I be sure that AI will not be a good way to leak my data?

[Erwan]

I still wanted to test...

I uninstall it immediately.

He doesn't even bother to answer me in my language, I'm not going to spend an hour translating his answers each time.

[Uberprutser]

I'm not the biggest fan of the whole "ai" fad, but I am giving Orbit a try 🙂

Some feedback:
I'm using a low end laptop, and Orbit feels a bit slow.
I would like to move Orbit to the left, so it no longer overlaps with the vertical scroll bar.
I'm using a screen reader extension, and parsing an Orbit reply to the screen reader could be better.

Orbit a nice gimmick, certainly no IA. But I prefer this above having to login to OpenAI 🙂

[hazing3991]

It would be less distracting to have it as a browser button in the toolbar, rather than a floating button that covers the web pages.

[Zoutpeper]

Instead of investing in a **bleep**ty machine learning algorithm you could also invest in making your product better. Adding "AI" (which is not AI) does not make your product better, it makes it worse.
This is neither exciting nor well thought out.

[Carolly]

Painfully obvious at this point that Mozilla sold out, because they have refused for months now to address any negative feedback about AI.

I don't really have anything to add because it's really clear you're not listening to your (now former) users. Just sucks. I switched to Firefox to get away from this garbage and now I've had to find an even more niche browser - and that's incredibly difficult when they're almost all built on the big browser frameworks.

Shame on you.

[PirosThe3rd]

Please check if it's allowed to say before replying, of course, but could you recommend alternatives to Firefox? I want to leave now too.

[Entropic_Decay]

Piss off with this garbage. I'm moving to a fork as soon as I can. You've had your chance to backpedal.

[NightmareSwag]

Get that **bleep** outta here, most of us use Firefox to avoid trash like this.

[November]

So you're all only pretending to listen when you make these "feedback" posts? I know you didn't miss the overwhelming negativity on your last AI post.

[PirosThe3rd]

This is genuinely embarrassing. After a record-setting number of negative replies on AI in your feedback forum, y'all are still not seeing that this will cost you significant money and reputation as a product and corporation? Did any of your staff even read them? If not, what was the point of asking for feedback?

The replies in that feedback thread were, for the most part, not reactionary. I read at least half the pages myself. We did our best to be polite and explain thoroughly why AI is not an ethical investment for the future and why you will lose business if you did this.

Your continuance to ignore us is clear. Firefox is no longer a browser for the people like it started as, it's for the corporations just like Chrome and Edge. Incredible job.

[onionella]

It does not soothe me a single bit that your Orbit does not train on _my_ data, as long as it is based on any of the currently available efficient LLM's, since they all, every single one of them, are based on stoled copyrigthed content. You, Mozilla, are normalizing the laundering of intellectual property disguised as artificial intelligence.

[Entropic_Decay]

Yknow, figured I might as well reply to this again, make my feelings on this move of yours incredibly clear.

The Mozilla I remember from when I was growing up - and even up until a couple years ago - is dead in my eyes. You're not listening to the outcry of your user base, you're ignoring the moral and ethical concerns around this stuff you're trying to push (which, if you'll recall, you stopped accepting donations through crypto a few years back specifically due to moral and ethical concerns over it), you claim you're not harvesting user data as you've always avoided and yet in the last year you've put in a data-harvesting function on your Firefox browser that users were opted into by default. You've completely eroded what trust and loyalty I had for you, which was quite a bit. From the responses to this and your last generative model post, it seems that's the case for many others as well. It's going to be a massive uphill battle to win that trust back, if you even bother, and you'll have lost many permanently - myself included, very soon.

So, I'll end this off with a couple questions. Honestly, given the lack of response to the negative comments, I doubt an answer will be forthcoming, but I figure I'll ask regardless.
Who's paying you to shove this LLM garbage into your flagship offering, and how much was the price for your users' trust?

[SoliTheFox]

Awesome! Can you guys consider using Llama3.1 instead? Or maybe 3.2 3b? They are really good multilingual models. In my experience they are one of the best models available and surpass mistral in day to day use. Consider Gemma2 9b and qwen2.5 too.

[eliasaj]

Looking forward to having this on Android. Summarize is a great tool!

[haingdc]

I'm still experience with this addon. These are something that we might need: In many sites, Orbit can hide content in the page, It's a need for collapse/expand feature to resolve it and still persist chat session.