Our coolest unreleased browser feature? (demo)
HTML-код
- Опубликовано: 26 июн 2023
- We're exploring AI in Arc... but is it truly useful?
Josh shares one of our early prototypes and wants to know what you think. But really. We're balancing a lot at the moment (Perf, Windows, etc) and only want to build what lights you up. 
Наука
My vote would be to nail the basics and add cool stuff iteratively, but only when the basics are bulletproof. In fact, put Nate on the basic stuff. The guy is awesome! I recently changed the BBC site colours and layout, which was great to do, but I still can't choose a default search engine in Arc for mobile.
I completely agree. For me it's Android support (or a standalone tool for accessing saved passwords at minimum) that's very much required.
There are several great browsers for the basics
Completely agree!
Speaking of default search engine, I'd really love to be able to use Perplexity. Not sure if adding this would be on Arc's court or on Perplexity's, though.
Totally agree with this comment. Basics first... (maybe) AI later.
I think it is a cool feature.Nevertheless I cannot think of frequent usecases in my everyday life. I think focussing on other important tasks (simplifying arc and making it more stable) is good idea.
they need them investor bucks
This certainly seems like a cool feature, but I don’t think it’s necessary for Arc to succeed. I would really focus on optimization and the Windows build, I think those are the two main issues right now when looking at the Arc subreddit.
This looks amazing! But I think adding more crucial things first (like windows support) is more important, currently arc is available to very few people, so I think extending the availability and getting more feedback is the better thing to do for now
let them work out the concept first because it's not just another bland browser. once they know what they've got, then extend it, otherwise they'll get go off course trying to service to the masses and end up in waysides of blandville
counting down the days until windows support!!
Have been waiting for Windows since last year. but they are just adding new features for a very limited audience.
yeah, windows support should be priority n. 1 right now, tho I'm down for this new feature as well, but only if it comes with some sort of a gallery with community made tweaks for me to use
Indeed this looks amazing! However I also agree that linux and windows support would have a higher priority :)
I absolutely adore this feature, but I'm honestly only going to find here-and-there use cases for this. I use both MacOS and Windows almost every day and I really can't wait to have access to my Arc spaces and tabs on my Windows PC. I'd much prefer to prioritize that over this.
I love that you came to RUclips to ask us how we felt. You guys are doing amazing work and I'm finding myself relying on (and loving) Arc more and more as I continue to use it!
great comment
i'm sure that the community will build some specific use cases that are easy to install and will enhance the UX greatly, very hyped
i've always kind of assumed the problem is more related to the syncing, where they've relied on icloud, rather than windows per se as a platform
That AI thing is cool. But I don't think it is needed as much as Arc for Windows or making it more stable, as you mentioned.
I’m loving the feature! But I do see other stuff like windows support as something which will give more value and attention to the app.
It’s pretty much a nice to have for me 😊
Yah that's our instinct as well!
@@joshmiller973 Yup, I am ditching chrome whenever you guys will release windows version.
@@joshmiller973 yes! i'm also anxiously waiting for the windows version, so i can move away from chrome... so i vote for focusing efforts on the foundations and what's in progress now, which is a lot already, and leave these cool AI features on the backlog for a future release.
a lot of their innovative capabilies cannot be done in windows without inefficient workarounds, it'd need a whole lot of extra work to make it work (regardless of the chromium engine), and they're at a stage of experimenting with the experience so extending to windows at this stage would drain resources unecessarily.
don't encourage that as they'll spread their resources too thin and we'll end up with another bland browser than pleases most people and excites none
@@neanda I am fine with basics on windows and innovative stuff on mac for now.
I need to be able to daily drive this across the devices.
THIS IS AMAZING, dont't get me wrong but this is stuff that should be a feature flag like (arc://flags) or some experimental tab in the arc settings where it can stay untouched (without someone constantly updating and supporting it) and as a nice to have. It definitely should not be prioritized over actions like platform support (Windows, Linux, Android), optimization, code clean-up, basic features (DNS over HTTPS, Search engines on iOS), etc.
It looks cool ! But I don’t even use boost, I don’t think I would use this. I don’t want a browser that packs a lot. Just want the convenience of tabs management and excellent performance (my arc crashes more and more (maybe because of all those widgets)). Keep on the good work ! Design and performance is all I need ! 🎉
I think arc should have a different update streams like "nightly" or "preview" where we can test out these cool stuff which might be in alpha.
I thinks thats a feature you can implement after the windows version and the optimisations. Right now the normal Boosts are good enough :)
As a new ARC user, I hope you’ll hold off introducing the feature. Focus on the essentials (including iOS and iPadOS). It seems every app out there includes some kind of AI, so those who need an AI fix can find it. When you are ready to open the Boost store as a revenue generator that would be the time. Or offer a separate ARC+ with AI subscription.
Just a word of caution about new features, especially the use of AI: I’ve been drawn to apps that offer this or that feature because it’s so cool. But (surprise!) it has often turned out I never actually used that cool feature. Right now you’re reconceptualizing the browser experience and 3 cheers to your team. But the siren call of AI can actually stall your project. Stay the course. Please!
Agree with you on this. Recently it's been a downhill experience for me after they introduced 'cosmetic' features instead of focusing on performance and efficiency.
Completely separate to the feature, I just wanted to say I really love these CEO Diary videos! So cool to get an inside view into what you all are working on, even cool features that might not make it to the user (that's me!). The transparency is really unique, and I don't know of another company that does it like y'all. Keep up the awesome work!!
You are insane. Amazing stuff
Edit: Maybe for now focus on the stuff you already have going on. The feature is amazing but on the top of my head I wouldn’t have a good use case.
right?! it demos so well but... it may just be a neat demo?
It is pretty cool but I don't think it should be something you absorb the cost of at this stage, maybe later on once costs go down or becomes a feature that has more use cases. Love Arc, thanks a lot for all the work and love you put into it!
Cool feature! But as many mentionned, I don't use Boost, and I don't see myself using this kind of feature very often. Plus, it will probably be veeery expensive.
I believe there are a ton of other features than the team could focus on (Arc mobile, Arc Windows, stability, and so on), that are way more impactful improvements and that could take the extra time left from this nice to have. But do not throw this away! It should still be on the roadmap when there is an extra time slot to fill :)
Very cool! I recently started using boosts more heavy to clean up websites I use often and I love it. I would love it even more if I could get a bit more granular like this! There's dashboards / websites I use for work with a bunch of links that I don't necessarily want to zap into oblivion but I rarely need them. When you say "change the way we use the internet." It's really a great set and forget feature. Calling it a "nice to have" feature is literally what it is. But the productivity boost of getting rid of more distractions can really add up to TIME and MONEY for us. I'd vote for it!
I think using AI to create code snippets for Boosts can be very useful. It reminds me of how I got started with coding by copying and pasting custom HTML code for MySpace and then slowly changing a few attributes here and there. It got me to learn from something versus a blank page.
Great to see demos!
I think using the test feature to select specific elements of a website and changing them at ease is really a fresh thing! It can lead to wonderful ideations and in visualising what something could look like. As a UX designer having that feature would allow me to explain things better to developers ( as of now i have to animate effects in after effects or premier pro). Great stuff!
You guys keep raising the bar. You're not just dropping ai into the browser, your implementing it in a way that's user friendly and genuinely useful. so dope
Hey team! I appreciate the transparency as always. I really like how you consult the community with decisions like these.
This honestly *is* a really cool feature - however, I can't think of any ways that I'd actually integrate this into my workflow currently. That's not to say I wouldn't find it useful at some point, but if it were to be introduced right now I'd only spend time playing around with it for fun rather than actually creating a boost I'd use permanently. The average person could spend 5 or 10 minutes spending your tokens just messing around with it, then never use it again. Because LLMs are quite expensive, I don't think it's worth implementing this for something that I think would just be a 'gimmick' for most people.
Now, I do imagine that some people would actually get some use out of this feature - so it's not a bad idea - but I feel that the money could be better spent elsewhere; at least while Arc is still unprofitable.
I know you've heard this a million times haha, but the one thing that I feel I'm missing from Arc right now is performance. The average person probably doesn't use boosts (I'm sure you're tracking metrics for this) but *everyone* cares about their apps running smoothly.
Thanks for reading!
This is very exciting and potentially revolutionary (in the long term) - by allowing non-programmers to customise the web experience much further than what boosts GUI currently allow. This could change the way we think about the Web.
On the other hand you still need the foundation of a stable, reliable browser. So for me personally those would be the priorities:
1. Undoing the "Better Windows" debacle (i.e., dealing with the change that broke the mental model of what an Arc "space" is, broke syncing of tabs, and messed with a lot of people's workflows)
2. Making Arc more memory efficient
3. The exciting AI stuff
I would love to play with that feature. I can imagine using it aside browsing like an assist. Imagine it cannot only reorganize pages but collect all sorts of content from a website to make researching much easier 🙌
I definitely would love using such a feature as a person who has trouble with coding and loves customization but I think I'd prefer wait for it to release after the Windows release as I'd love to already use the Arc browser with how amazing it seems to be.
I would start using arc frequently if this was integrated- very very cool! 👍👍👍
Y'all are amazing at social and community engagement. Thank you for taking the whole world along for the process. Amazing demo, but I agree with other commenters that stability and core basics are more important than anything. Though I wonder, what would it look like for this solution to come from userspace? Maybe external devs can make these awesome exciting features, and you support a rock solid stable core. (A lot of questions and complexity there though, of course)
Thanks for the transparency
This should definitely be added! But more importantly, I hope to see the Windows rollout very soon!
This is certainly a great idea!
In my opinion you can release this feature but for the more technologically advanced people to use.
For example, give us an option to put our own open-ai api so you won't have to pay for the ai use.
And if so, I would actually use it a lot to quickly make summaries of different pages etc. Less like a page editing feature.
It's cool and well executed. I don't have any use cases for this that I can think of.
I'm already so happy with the lowercase text style boosts!
Incredible stuff! This experimental feature shows just how groundbreaking and forward-thinking the team behind Arc is. I think adding LLM from the ground up, is key to shaping the trajectory of where the internet browsing experience will be. Can't wait to see where this leads. Kudos to the team!
I'd leave it for now. Far too niche. Super cool to look at and show the power of Arc, but...
we agree tbh
Yo Josh, I think y'all may be at the same place we have all been. The "need" seems so arduous that we chase the "want," and The Browser Company is probably chasing both, which is tiring. This AI capability should 100% be implemented and I look forward to when it is, but your product is an A+ with or without it. It'll be worth the wait.
I think this is a really cool feature and could be a gamechanger. It could operate like Boosts too and you could create a "market" where you could download prompts made by other people as well.
This is amazing to be honest.
Seems useful to me for learning developers or any dev in reality, also for screen recording, you can interact however you want with the web meaning less editing post-recording.
This is also useful on a daily basis to surf the web results / thumbnails, and other things we come across, maybe by creating an automation on this feature to dynamically update "long titles" or "content with (x) words" etc etc.
BUT, I wish this would be available with some other tests / prototypes under another section, maybe "Arc Labs" or something else, but a section like this with every disclaimer and saying it won't work on most machines just to get feedback, will really be amazing.
////
But please, focus on the basic stuff, optimization, speed, reaaally solid browser essentials, thank you! You are doing an amazing job.
this is incredible!! this would be incredibly useful for not even modifying webpages to make things better for you, but also for web dev!! quick prototyping of ideas, asking arc to "change this box to (x)" etc etc, for developers, this is incredible
I think the concept is incredible! Several ideas spring to mind that could be super useful in my everyday life. For instance, having a summary of news right at the start of the page, integrating emails, etc., seems like it could be a valuable feature. I believe it would be an interesting addition, at least for the beta testers!
This is like on demand, ad-hoc extensions. I love this idea and I want it yesterday!
I think it's pretty cool! I think there's more potential than just in boost though. Like one thing I use a lot is grammarly go ( a feature of grammarly) where it helps me write out letters, group text, etc. It would be cool if there was just a feature you could highlight text or whenever you are in a text field just have ai write something in a prompt for you.
I think this is a sick feature down the road. Make using it currently the best experience ever and then add snazzy features later.
I really like these demo videos.
For the topic, I personally think that If you have stuff going on in the pipeline, you should maybe focus on that. I have some windows friends who cant wait for the browser to come out, also the polishing you are doing lately is just amazing, so maybe I think thats more beneficial to the project. But the demo looked sick, so I wouldn't scrap that.
This is awesome, I know I'd use it to blur spoilers by topic/title 😆
Very hyped for this feature; keep up the great work ❤🔥
I like this thing where you show us demos and ask the user base for feedback on the features that you've come up with. It will definitely avoid user backlash also wasting resources! This is awesome. I wish more companies would do this.
I do have to say adding windows support first would be amazing. Thanks for all the work that you do!
Honestly - this is a GAME-CHANGER and mind-blowing! This can be used by anyone, all day long.. I have an endless list of such examples and implementations! Would be happy to connect and share my ideas with you.
Would be great to have these
a) A way to generate an X word summary of the article, video I am currently on . That tells me whether I should spend time on reading/watching the whole thing. like a blinkist for the internet.
This would be nice for text selections too. Would be cool if these were overlayed on the page as section by section summaries.
b) A way to highlight bias in an article. just highlight areas in the page that have significant bias towards whatever and tell me what the bias is.
c) semantic search over all my archived stuff. I don't mind running a light weight LLM locally for this.
Looks like an amazing feature! I think it's one of those things that you don't really know all the possibilities until you are able to mess around with it. It seems like it would work well with like community prompts for various websites that you could just browse and apply, that way people can be creative with it and share their uses for it with others.
2 good use cases relevant to me: 1) content creators or people reacting to stuff and 2) documentation and screen capture
I think this is the most useful for content creators for sure. Another boost would be to simplify pages for being displayed while doing livestreams. Sometimes people read through blogs on livestreams, and it's good when you have something that can be smart about how it turns that page into something simpler that can be livestreamed.
I would use the blurring to blur things when taking screenshots for documentation. I've always thought it was annoying to have to edit a screenshot when I just want some simple blurring of something or want the screen to look a certain way.
You guys are the innovators. Seriously awesome stuff.
Stability!!! Simplicity!!! Arc is an amazing browser experience already. Everyone I show it to is blown away by how different it feels. Make that core experience better. The AI stuff is really cool (and I am a professional data scientist), but, it just doesn't seem to have a killer use case here yet. And thanks for keeping us in the loop on this stuff ✌
Loving Arc! Genuiniely it's changed how I use the web. I am all for cool use of LLMs, but the biggest reason I love about Arc is that every feature feels well thought out and essential to the experience itself. Like they are working on concert rather than in stutter step. I would say hold off on frills like this for now. I know myself, and probably a lot of other people, would be better served if the time was used on honing the essentials.
This community rules! Peace :)
I agree with most others here - focus on speed, squatting bugs, improving arc mobile, and bringing arc to windows, and maybe work on this later down the line
Hi Josh,
Thank you for sharing the new feature; I genuinely find it intriguing. At the moment, I'm struggling to find frequent use cases where I'd interact with the text filtering & customization capability as described.
If I were to suggest any enhancements, I believe integrating translation capabilities for commonly used languages would be of immense value. Moreover, it might be beneficial to prioritize strengthening Arc's core functionalities and performance. I'm confident that there's a large audience eagerly waiting to adopt Arc as their primary browser.
On a related note, I've faced some challenges accessing Arc on my iPhone. There seems to be an issue when iCloud isn't connected, preventing me from using the browser.
Overall, I'm quite impressed with the direction Arc is taking and eagerly await the ability to use it seamlessly on my phone. I appreciate the update - the new features make the experience feel more personalized and enjoyable!
Cheers,
Sel 🙌
Would love it for the future for being able to creating demos, but not sure if it's prime time. The blur feature is a great showcase... but I'd have some more ad specific solutions for it.
However it can bring the potential of a market place...
This is super cool, and if you could create a custom layer of AI code that runs every time you open a certain website, that could be incredible
Realistically it's more important to nail the simple things. In part because they are what really serves users long term, but also because you don't want lots of headline-grabbing features that encourage new members... only for them to be turned away because the simple things haven't been nailed.
Also, for that specific demo, I think there might be a better way to do it. Instead of restricting to needing to select *1* element, allow us to just write what we want it to do, then provide references for the elements we are talking about. For example "Make all buttons [insert element reference] in the carousel blue, except for the profile button [insert element reference] which should be green.
Also think about email and any replies with AI, text write with AI, code generate with AI write into Arc browser. Lens feature is cool, but you can make even more fundamental approach of AI with providing the tools listed in this paragraph.
Would be nice for us to have some kind of simple web scrapping tools. Basically I was waiting for something like that in Arc, i think it is extra USP to make Arc even more unique.
You have to publish this. This is the tool I always needed, saving endless time as web developer. PLEASE PUBLISH!
I think it would be amazing, it won't be used on a daily basis however, it'll allow users to use Boosts without having to re-write code, making boosts relevant for more people but also increasing the number of boosts templates. Great feature, can't wait to test it.
I love it, for sure, and I may be biased since I'm on Mac and been using Arc for a while now, but for the growth of the company I think it should be secondary to whatever is on the front burner at the moment.
Great initiative you guys! Love what you guys are doing! I would vote to have AI features it's gotta be sooooo amazing!
I'd definitely use it for things like the example shows. Rewriting product names would be awesome. I'd want it to go further to where I can just describe what I want it to do to the page in general without selecting anything and it gives it a shot. Maybe tricky to make work consistently though, haha.
But yeah, as far as priorities, I'm definitely waitin' for windows for my home PC and hopefully Android down the road. But please continue to have some folks at least working on wild ideas like this. It's important.
These are remarkable ideas! I say go for it, but make it easy to turn on and off the features we add, as well as the ability to turn off all of them at once. It would also be a good idea to curate a gallery of these add-ons where we can share and steal ideas :D
I really only use boosts to change a font or remove annoying elements. It's a cool feature but I think it does what I need it to for now, and I don't think the monetary cost of AI or the time cost of improving boost is worth anywhere near as much as getting arc on windows (or fixing sync, for that matter)
It is awesome! I would love to see this feature in Arc
This would be absolutely amazing, even though most people wouldnt use it, power users and enthusiasts would figure out some amazing ways to use this, its like making website usage completely personalised in any way even if you do not know JavaScript
I am just learning about this browser and I am not super duper techy, but man this feature would be such an amazing one for me, and a huge relief because I am very picky and sensitive about the content that I want to encounter/involve myself with online. I like to keep my space positive and ideal, so I would definitely want to use this to block out posts with certain keywords if that is possible, and also, hide news sites and articles because they tend to be overwhelming and unwanted. I literally have been dreaming of a way to make the internet feel more personal to me and less jarring and random. I want to see what I want to see. I am way too excited for this browser regardless though, due to the other features I just learned about. But this one in particular would be my cherry on top! ^^
This is an amazing feature! I am so excited to see this and other AI tools integrated into Arc. That being said, I think it's best to prioritise the basics and not risk spreading yourselves too thin. Focus on brining Arc to Windows and fleshing out the mobile version so it doesn't feel like demo. Once you've done that, then you can start adding all these cool extra features
I think adding with pre-defined prompts, that would appear as a set of tools could be interesting (the prompt itself would be transparent to the user) .
This is super cool, but, as a windows user, i'm biased to say - bringing arc to more platforms is more important than this.
I would still love to see that after windows release though.
It's really a cool feature I really would love to have as soon as possible but again like lot of people said supporting for windows will always comes first, Anyway I really love your work guys! Thanks a lot
I would love this so much! Boosts are something I will use so much once Arc comes to windows, and this would add a lot to them. It does need to be on windows for that to be the case though, so that should be a bigger priority for now if you as me :P
Man, I would love to have this at work as mostly backend coder that sometimes has to do frontend. Since I don't think the average user would use this I guess it's better for the company to develop crucial things first, but again this feature is AMAZING.
Holy moly! As a frontend developer myself I would love that feature!
I would use this to translate pages into languages I’m studying but adjusted to my current level. For example “translate the body of this blog post to intermediate Japanese” or “translate to Korean using beginner level vocabulary and grammar”
This is amazing! This could literally force better UX/UI for most websites, because otherwise it would be filtered out! Personally though, I think that having built in summarisation and powerful text-to-speech for articles would be very more useful. I do wonder if this could be on the client side though, using the GPU.
It's cool to have in the future, I instead choose another cool stuff that is even more surprising from u guys! 😄
Lets wait with it. First basics
I feel it is pretty easy to do this yourself with the inspect tool, I was literally doing that yesterday.
BUT, I would say there are other implications, I think the grabbing text from elements on a page could be used for many other things with AI. I think it would be more useful for it to able to grab the text and put it in a Markdown Note, or generate in a markdown note similar to your Easels. This would allow users to create content from broad sourced information quickly, such as a market report, or blog sourcing a lot of the best material, or even notes for a podcast.
it’s a nice to have and more about the future vision, it seems like more and more people are having performance issues which is probably a bigger concern when it comes to a browser.
This is useful! Please add also global boosts ) To have consistent css etc. And also a keyboard shortcut to open boosted css for a current site.
oh i want this feature!! it's really powerful! Good Job guys
Never prioritize this before finishing the fundamentals in the backlog!
This feature looks awesome! It would allow for a whole other level of website customisation. The possibilities of what people could do with this are endless.
I'm sure this feature will be handy in many different situations
Would love this addition AFTER all the other planned projects are finished (perf and windows are high in my list). In my daily work as a lead I would definitly use some of the blurring and filter features on dashboards that don't have this build into themselves, makes it easier (and less preparation work) when discussing performance with a few employees
I woudnt use it that often being honest, for a few webs maybe and casually imo! Thank you guys as always
I would hold off on this feature to prioritize the other basics. Waiting to release this will also let the LLMs develop more.
This is very cool, I love this idea and ability. I think it really takes things to the next level of making the internet 1) Designed for users, and 2) Making Arc boosts more than just (extremely well done) cosmetics. My concerns... Does the AI feature need users to have their own license? What potential for inaccuracies does this expose users to? Are there dangers to interacting with the web when the content they see isn't what was intended by the website?
Your team is awesome, I'm sure you'll figure this all out either way!
can't wait to use it on Windows! Really want to switch already
Honestly this would make it a lot easier for anyone really to make the internet their own realm and have a good Boost ecosystem. I really also hope Arc comes to Linux some day but it seems I need to manage my expectations as Arc is still working on its windows version.
Amazing feature, really useful while building websites in order to preview directly the test without having to build it for nothing
As a product owner of a scrum team I work with user stories on a daily basis. Stakeholders make very long and detailed comments in those stories. With this feature I would for example have all the comments under a user story be summarised. I can see many more useful areas especially for summarising longer texts and so forth.
This is really cool, but feels more like a "nice to have feature", rather than a needed feature. I wouldn't mind having it on Arc, but if that meant slowing down the speed at which the Windows app is developed, then I'd be happy to wait. AI isn't going anywhere any time soon and will still be there to implement after the windows release
I absolutely love Arc and use it as my daily browser but will echo the sentiments of others: I don't personally use boosts and to me, it's more important to have a rock-solid browsing experience (performance, great tab management, etc) and would love to see most efforts being put forth into those things and making Arc more accessible across lots of different devices
looks like someting you can add later on. I would focus on windows as I feel that will give you guys more of a foot hold in the market place. For me personally I would check it out, then most likely put it on the shelf to collect dust. Stick with adding Arc to windows as I use both systems... .Keep up all the amazing work you and your teams create.
I keep trying to find a way for Boosts to be useful to me, but I havent found it so far. So, in my eyes, resources you spend on Boosts are resources away from features I care about (ie., spaces, cmd+t, organization, profiles, etc.).
But I know y'all want to change the way we use the internet, which requires me to break my old habits - even habits I don't know I have. So I want to be open and excited about this, but idk if it's "better enough" for me to really get excited about.
I really think this feature is awesome and could really useful, You should maybe add a community library where user sorts every code for popular site than every one can use them even if some people doesn't have enough creativity. ❤😍
this is a great idea! i really like the amazon bit.
Amazing! I really want to use that in future.