TL;DR: it makes Chrome faster by running it in the cloud and streaming it to your machine.
One thing I don't see anywhere on the page: pricing. No-one is going to run a giant fleet of cloud servers out of the goodness of their own heart, so either I end up paying for this service or they extract some icky level of personal information to pay for it. The site says "Your data is your data. You’re not the product", so I assume it's the former. But without any pricing details I can't really evaluate whether this is worth trying or not.
My personal method of making Chrome faster is to use Safari. It consumes way less battery and sites run more than smoothly enough for me. Everyone's situation differs, obviously, but I'm more comfortable running that locally than depending on a remote service (and a very stable internet connection!) to do my essential everyday tasks. At a bare minimum I'd want this to have an option to "downgrade" to local browsing for when I'm tethering, etc.
If you sign up to hear from them they ask how much you'd be willing to pay for it. The options are <$10/mo, 10-20, 20-30, 30-40, 40-50, and 50+. To me this suggests they are looking at charging similarly to tools like Superhuman which is $30 a month.
I know a couple of people that like to open lots of tabs and also swear by Tree Tab. I haven't used it personally because I rarely have more than half a dozen tabs open at a time.
Tree tab is a must for me when doing research. This morning I closed a few Firefox windows. The largest had 421 tabs, the others around 150-200 each. I might be a compulsive middle-mouse-button and new-tab user.
I saw the founders tweeting yesterday about launching today, but the page still shows a button to request access. Makes me wonder if I could also launch like this.
This is a great technical solution to a problem that is really about users not understanding how their use of a computer really affects performance, and companies under-spec'ing the machines they give people.
4GB of RAM in a MacBook Air is not enough for your average knowledge-worker living in their web browser. 8GB is probably fine for most, but if you're a designer using Figma? Maybe not.
Also I suspect that while most users know that lots of tabs makes their computer slow, I think most users also have a fairly fuzzy idea of what's a slow computer, what's slow internet, what else might be slowing their computer down, etc. If you're here on HN you're probably not one of these users, but they're not uncommon.
While I applaud the technical solution here, I think a lot of companies should be seeing their logos on this page as a sign that they have failed to create accessible software. If your target market is considering renting cloud compute to run your webapp, maybe that's something you need to fix.
> This is a great technical solution to a problem that is really about users not understanding how their use of a computer really affects performance, and companies under-spec'ing the machines they give people.
Let's not forget developers and site owners stuffing webpages with tons of fluff, especially megabytes of JS that is mostly tracking code.
I'm not sure the business case makes sense? A web browser is something most employees would need to be running throughout their working day. So, something like 8 hr/day. Even assuming that their claim of "16-core Xeon per browser" is slightly embellished, these machines aren't exactly cheap.
I can see €64.26/mo (~$77) from Hetzner for a Ryzen 3700X octa-core box, so assuming a box can be shared by three users (maybe from different time zones), that's still ~$25/mo per user just to pay the cloud provider.
25/mo = 300/year... if you figure that box is equivalent to a 1200$ laptop = 4 years... except you don't have to pay up front and you get to upgrade to more modern hardware as it comes out.
The flip side is you need a fast internet connection and a relatively cheap endpoint laptop.
I don't know, doesn't seem like a good deal to me, but it's not a ridiculous idea. Being able to timeshare the expensive hardware would be a good thing.
When you're financing a computer, you're still stuck with it if you stop wanting it. Not paying up front here means you can cancel your service at any time and not have to continue paying, you can upgrade to the new faster computer the second it comes available for the same price. Etc.
> I think a lot of companies should be seeing their logos on this page as a sign that they have failed to create accessible software. If your target market is considering renting cloud compute to run your webapp, maybe that's something you need to fix.
The goal of a company isn't to create software that's runnable on as many computers as possible. It is to create a product that is valuable to their customers. If those customers are willing to spend loads of extra money running your product that is a strong signal that you are doing something right.
Alternatively, it can mean that customers don't have any other better options. Take work chat for instance, in which all of the available options are heavyweight web/electron apps. In fact for that category specifically it's a common gripe that all the options are bad and one has to select based on which is the least-bad for their particular situation.
> 4GB of RAM in a MacBook Air is not enough for your average knowledge-worker living in their web browser.
You can't even spec a MacBook Air with 4GB of RAM any more.
Base configuration MBA has 8GB of RAM. You can even finance it for $83/month for 12 months, which isn't much more than a $50/month Mighty service. But you get to keep the laptop at the end of the 12 months if you finance it.
> 8GB is probably fine for most, but if you're a designer using Figma? Maybe not
Doing anything interactive, I'd be more concerned about latency. If I had an extra 100-200ms round-trip latency on anything I do in the browser, design work would become a lot more frustrating.
Looks really cool! This was my biggest pain before buying a M1 MBA. In the "running Slack/Figma/SaaS web apps" space, are they competing directly against low-energy, more-powerful chips like the M1? Whereas I can imagine lots of use-cases where it's impractical to buy a machine like that.
On one hand, you can pay ~$30/mo to have somebody else do your web browsing for you. On the other hand, you can use an adblocker (free), Brave Browser (free), and just not have 8 quintillion tabs open. I'm trying to be open minded here but this seems really over-engineered to me
To me the interesting difference with mainframes is that you were generally taking advantage of the extra processing power available on the mainframe. Whereas this is more like taking advantage of the extra RAM available on the cloud machines. But yeah the spirit of it is very similar.
There is a recurrent pattern where systems are rebalanced where resources allow for 'more'. Centralized comes first because concentration obviously helps having more resources then market distributes capabilities (desktops/laptops). And now people are going the other way, maybe because local resources are not growing fast enough.
WebTV did this in the 90s. Web pages were reformatted and rendered on the server, then sent down to the WebTV box (originally over modem). Microsoft bought the company, called it MSN TV.
I don't think they stream video back. It could make more sense to take over memory consumption and CPU/GPU heavy threads in a compressed binary format, especially from background tabs and restore them on demand.
What's the problem though? The market clearly exists (right now), so I don't why blame the company trying to get a share and help with this issue.
It's clear that Google isn't going to optimize Chrome and people aren't going to switch to Firefox with Tree Tabs / Sidebery with background suspension.
Lots of browsers are implementing background suspension. Even if Google decided that they'll never implement this, it will be hard to convice potential customers to pay x$/month to solve a problem already solved by others browsers.
And my issue was not about wether the market exists or not. It's about an unreasonable solution to an unreasonable problem.
Market clearly exists if you are able to mass product diesel-powered personal jetpacks for $49.99. But it says nothing positive about our future. (but i'd be glad to try it at least once anyway :D )
Well, it sounds like a parody product because browsers are supposed to be the lightweight, fast clients for the things we've offloaded to cloud servers.
By tradition, web browsing is the quintessential lightweight task, letting laptop vendors report "10 hours of web browsing" and the budget-conscious to say "8GB of RAM is more than enough for everyday tasks like browsing facebook"
Hearing that someone runs their web browser on a cloud server is like hearing someone has hired a personal assistant for their personal assistant.
It isn't insane, it's just an engineering tradeoff. Streaming a browser takes more of some resources, probably bandwidth, processing power for video decoding, and the cost of the remote hardware. But it takes less RAM on your local machine. If you are running out of RAM and not running out of the other resources, this tradeoff makes complete sense.
Streaming your web browser and thereby exposing all information accessed and sent, including passwords, to a third-party company because you had too many unused tabs open is insane nonsense.
Yeah, chrome sucks and logs your every move, but it's not live streaming passwords and giving a third party the ability to just... Take over the session while you were logged in to online banking.
Let's say this ends up being 30$/month (I've no idea how much it will cost, but I saw this number floated above).
Let's say you replace employee machines every 2 years.
Doesn't it make more sense to just spend 720$ extra per employee on hardware? It'll be a much better experience, with much less risk (what if mighty is unsustainable and closes down?), and that machine will still have value in 2 years, unlike throwing money at a cloud subscription.
Fwiw, I still feel the pain on a top of the line 16" MBP. For many tech workers, a laptop is what they get, and what they rely on in terms of portability vs performance trade off. Using this is probably more convenient than getting beefier hardware, which I suspect is gonna have to be a bulky souped up workstation.
After the pain of many, many years of dealing with IE, I will never again use a Microsoft browser, no matter how nice it looks. Fool me once, and once only.
This clearly isn't marketed towards me (I'm happy running firefox locally), but from what I can tell I'd like to use every part of this product except the core offering.
- Mirror my tabs in the cloud? Great!
- Opt+Tab to navigate my overflowing tab bar? Sure!
- Cmd+J to instantly join meetings? This might be the killer feature for me honestly
- Search through all my google docs from anywhere? Sure, why not?
The problem is, I can get most of this through google calendar alerts and firefox extensions. I wonder how their value prop will evolve over time, because right now I don't see it being worthwhile for anything except crash recovery. With M1 Macs being as quiet and power-efficient as I've heard, it sounds like the main market these folks are targeting (execs/higher ups that aren't as tech-savvy) would rather just use newer machines?
M1s are heavy on processing power but not too heavy on RAM, and it seems like this offering makes the most sense for people running low on RAM, so perhaps M1s are a logical market.
Ah, classic HN. There can't possibly be a post on here about something new without someone going off about how $more_complex_technology already exists.
Your parent probably meant that there is nothing specific to Chrome in this technology, which / because it looks like VNC, not that the technology is useless because there is already VNC.
It's because we're not impressed with people reinventing the wheel and putting a new sticker on it. History repeats itself. We've seen it before. Maybe this is drastically different, innovative or helpful, but so far its just meh.
I've been following Mighty's development for a while through the founder's twitter. It seems to me like they are doing so much more than putting Chrome behind VNC. They've been doing a lot of deep technical work [0] to make sure that the experience of using Chrome through Mighty is as frictionless as possible, like delving into Chrome's scrolling algorithm to get it just right [1]. I don't think it's fair to dismiss what they've achieved on the basis that VNC can do it too, but much worse.
So Figma is written in JS and C++, compiled to WebAssembly so it runs in a browser, which runs in a datacenter, with video streamed to Mighty, an Electron app where the front-end is written in JS and some C++, running inside Chromium.
Then good news, between modern software development practices, NFTs, and just plain fucking laziness and incompetence, there won't anything left of it soon!
I feel like we engineers are putting too many abstractions on things. It's like we are all peddling "get rich quick" schemes to people trying to weasel our way into some super popular process. This screams like an anti-direct-to-consumer model.
STOP CREATING MIDDLEMEN! It's going to cost me 30 bucks to just browse the web where I spend another dollar amount to where someone collects a "handling fee". Jesus I feel like the world is going nuts.
Aside from your HTML interpretation there's another interesting thing that's been tried by a remote browser start up bought by cloudflare, called, I think it was, s2. They hooked into the chromium rendering engine skia drawing instructions and instead of sending screenshots or video from the remote browser to the client they sent the skia drawing instructions and then rebuilt the entire rendering of the HTML client side.
Point taken. But there's a time and a place for everything - nowadays we have the tech to stream the contents of the desktop, and quite probably to it real time. So you could say the problem HTML was solving is gone.
Do we really need this though? Web browsers are slow because of all the javascript running on them, not because we all don't have Intel Xeons clocked in at 4 GHz.
I even have Firefox sitting at 5 GB ram usage right now for 150 tabs. I don't think I've ever had an issue with performance on browsing the Internet.
Most probably, the bottleneck is bandwidth/CPU for most users.
Typical pcs have 8gb ram. If Firefox is using 5gb then the system is under memory pressure.
Assuming they have slack or discord and Spotify running for example
Firefox is really only using 5 gigs because the system has so much RAM in the first place. It's just their default tuning options. If the RAM is needed for slack/discord/spotify, then just restart Firefox. It'll make use of a fraction of what is left.
You can also reduce Firefox RAM usage in about:config and by reducing the number of content processes in settings.
1. I don't have the problem of feel like Chrome being slow, and I don't hear this complaint much. The complexity of the web is not increasing as quickly as computers are increasing in power. This seems like a temporary and niche problem to be working on.
2. Reliable low latency streaming on wired connections is pretty straightforward, and should work fine. This is an easy problem.
3. Reliable low latency streaming on wireless connections is an unsolvable problem due to the nature of physics (basically), and will be an endless source of frustration. There's a reason no FPS gamer would ever play on wifi by choice. It will work fine at times and then randomly start sucking right as you're trying to do something important.
4. If it turns out this is useful in some cases, Google can easily do a better job than Mighty. And there's no reason this couldn't be done by AWS and Microsoft as well. It's trivial for a major tech company to do this better than Mighty does. They already built Stadia and the rest. Unlike when Dropbox launched, these companies aren't sleeping on stuff like this anymore.
I would love to see this on iPad and iPhone. I haven’t been too bothered by Apple’s rules, but in this case you can clearly see how they stifle innovation.
There are potential privacy upsides as well. Even compared to tunneling your traffiv to a "vpn" service, this might be more resistant against traffic analysis.
If you are tunneling between home (A) and "vpn" provider (B) that terminates your traffic and sends out it to the public internet, and you are browsing websites C and D, both sharing the ad / tracking site E. 1. It's easy for anyone observing traffic between A and B to deduce from the traffic patterns what kinds of requests you are exchanging between C and D, 2. It's easy for someone obsrving traffic between A and B, and colluding with tracking site E, to correlate your address at A with your browsing behaviour on sites C and D by correlating A-B flows with information gathered by E.
In the Mighty type architecture, the leaked signal from the flow between end user and Mighty servers will be harder to match to the HTTP(S) traffic based on packet timing and sizes. But not necessarily hard enough unless there are purpouseful andi-TA measures employed. Hence I said "potential" privacy advantages.
Both of these scenarios place equal amount of trust in service operator - the difference is in security is against other adversaries.
I don't think that there is much of a TA benefit -- you make a request (and let's assume it's just a packet count/timing analyzer) and 10 packets burst out of your device to Mighty, then 10 packets burst out from Mighty to the destination. How is that not equally analyzable?
> Both of these scenarios place equal amount of trust in service operator - the difference is in security is against other adversaries.
This is where I would strongly disagree. With a VPN they can analyze your endpoints, but assuming E2E encryption, they cannot see your actual packets. Whereas Mighty is definitionally a MITM browser, they can see everything. Therefore you are trusting Mighty incredibly more than a VPN provider.
> you make a request (and let's assume it's just a packet count/timing analyzer) and 10 packets burst out of your device to Mighty, then 10 packets burst out from Mighty to the destination
By my understanding, the web browser runs at the Mighty server farm and it's streaming it to you using a RDP or Stadia style system. So, it's more like you send 10-ish packets containing input events (like mouse motion and click) using the Mighty client protocol, and as a result the Mighty service does 150 web requests that happen due to loading the new web page (probably around 10k packets). While the page is loading, the Mighty service sends you screen updates that are again much different than the packets the Mighty service is receiving in response to the web requests.
Yes, these can still be temporally correlated, even though the correspondence is much more distant than in the VPN case (where you can just observe the similar sized packets). But there's potential to fix it rather easily, by eg using chaff traffic to the client in idle periods, and/or by pulsing updates out to all users to the service in sync, etc.
It's like VDI, but for apps that we originally designed as thin clients. Does that make Chrome a thin thin client? I think if a modern PC can't run your web app it's a sign that web developers have gone off the deep end, and either need to re-architect their app or ship it as Electron with a lighter web-only substitute.
Also, Cloudflare recently launched a similar thing, but it's designed for situations where you want an employee to access a service but don't trust their browser: https://www.cloudflare.com/teams/browser-isolation/. That's the only situation in which this makes any sense, and even then, if you don't trust the employee's device you are probably hosting them a virtual desktop anyway.
If you don't trust the employee's device at all, you shouldn't let them use a keyboard.
Even if your authentication is password-free (say, TOTP plus pick-the-right-icon from a set of 64), you don't want a keylogger picking up anything else, do you?
I'm curious how they deal with local printers, file downloads, file uploads, links that launch things like a native local Zoom app, etc. That wasn't fun last time we did this with thin clients and Citrix :)
I found this in their hiring page regarding file uploads.
> We’ve implemented cross-platform Drag and Drop file uploading. When you drag a file into a Mighty window on macOS, we simulate that same sequence of Drag-and-Drop events on Linux. We trick Chromium into thinking it's uploading a file from the Linux filesystem while, behind the scenes, we stream the file from the user's Mac; we accomplish this using Filesystem In Userspace (FUSE).
I don't see the actual business case for this at all.
Most businesses amortise a laptop/PC over a number of years. Would you rather pay $10/month/user for this cloud service, or spend the additional $360 (laptop/PC lifespan for a business that can amortise the asset over 3 years) in the first place to get more powerful hardware locally, and benefit all apps rather than just the web browser?
In case you're seriously wondering, it's vastly preferably for a business to spend $10/month in opex instead of $360 amortized over 3 year in capex. That's why everything is going rental and outsourced, even the plants in your average fancy office are rented by the month.
Honestly, I can see the value proposition for saving battery life on a laptop if you're doing something resource intensive in browser. I mean, if you still have to stream video I wonder how much battery most people would save.
But why would anyone want to outsource their web browsing to a third party? If this is something you need, you should setup a homeserver and RDP into it...
This is a very very weird product. From the site which looks like a nice pitch btw, this basically looks like a VM that only serves as a browser. What problem does it solve?
Btw. My teeth gnash at the thought that my assumption above about the product is correct.
774 comments
[ 4.1 ms ] story [ 399 ms ] threadOne thing I don't see anywhere on the page: pricing. No-one is going to run a giant fleet of cloud servers out of the goodness of their own heart, so either I end up paying for this service or they extract some icky level of personal information to pay for it. The site says "Your data is your data. You’re not the product", so I assume it's the former. But without any pricing details I can't really evaluate whether this is worth trying or not.
My personal method of making Chrome faster is to use Safari. It consumes way less battery and sites run more than smoothly enough for me. Everyone's situation differs, obviously, but I'm more comfortable running that locally than depending on a remote service (and a very stable internet connection!) to do my essential everyday tasks. At a bare minimum I'd want this to have an option to "downgrade" to local browsing for when I'm tethering, etc.
4GB of RAM in a MacBook Air is not enough for your average knowledge-worker living in their web browser. 8GB is probably fine for most, but if you're a designer using Figma? Maybe not.
Also I suspect that while most users know that lots of tabs makes their computer slow, I think most users also have a fairly fuzzy idea of what's a slow computer, what's slow internet, what else might be slowing their computer down, etc. If you're here on HN you're probably not one of these users, but they're not uncommon.
While I applaud the technical solution here, I think a lot of companies should be seeing their logos on this page as a sign that they have failed to create accessible software. If your target market is considering renting cloud compute to run your webapp, maybe that's something you need to fix.
Let's not forget developers and site owners stuffing webpages with tons of fluff, especially megabytes of JS that is mostly tracking code.
I can see €64.26/mo (~$77) from Hetzner for a Ryzen 3700X octa-core box, so assuming a box can be shared by three users (maybe from different time zones), that's still ~$25/mo per user just to pay the cloud provider.
The flip side is you need a fast internet connection and a relatively cheap endpoint laptop.
I don't know, doesn't seem like a good deal to me, but it's not a ridiculous idea. Being able to timeshare the expensive hardware would be a good thing.
I mean I get 12-24 mo, 0% financing on all Apple products with my Apple Card.
The goal of a company isn't to create software that's runnable on as many computers as possible. It is to create a product that is valuable to their customers. If those customers are willing to spend loads of extra money running your product that is a strong signal that you are doing something right.
You can't even spec a MacBook Air with 4GB of RAM any more.
Base configuration MBA has 8GB of RAM. You can even finance it for $83/month for 12 months, which isn't much more than a $50/month Mighty service. But you get to keep the laptop at the end of the 12 months if you finance it.
> 8GB is probably fine for most, but if you're a designer using Figma? Maybe not
Doing anything interactive, I'd be more concerned about latency. If I had an extra 100-200ms round-trip latency on anything I do in the browser, design work would become a lot more frustrating.
Solving problems caused by overengineering with overengineering ?
Isn't this just insane ?
If you have performance issues because you use a lots of tabs, just use a browser which is able to pause background tabs ?
Compressing and streaming pages was kind of the idea behind Opera Mini as well https://en.wikipedia.org/wiki/Opera_Mini
So nothing really new…
https://en.wikipedia.org/wiki/Amazon_Silk
Are we also mentioning the fact that the whole browser navigation goes through a third party service ?
GDPR will be fun on that one.
It's clear that Google isn't going to optimize Chrome and people aren't going to switch to Firefox with Tree Tabs / Sidebery with background suspension.
Lots of browsers are implementing background suspension. Even if Google decided that they'll never implement this, it will be hard to convice potential customers to pay x$/month to solve a problem already solved by others browsers.
And my issue was not about wether the market exists or not. It's about an unreasonable solution to an unreasonable problem.
Market clearly exists if you are able to mass product diesel-powered personal jetpacks for $49.99. But it says nothing positive about our future. (but i'd be glad to try it at least once anyway :D )
But yes, I agree on the sustainability future. Afraid there isn't a solution other than Chrome losing market share or Google implementing the feature.
By tradition, web browsing is the quintessential lightweight task, letting laptop vendors report "10 hours of web browsing" and the budget-conscious to say "8GB of RAM is more than enough for everyday tasks like browsing facebook"
Hearing that someone runs their web browser on a cloud server is like hearing someone has hired a personal assistant for their personal assistant.
Streaming your web browser and thereby exposing all information accessed and sent, including passwords, to a third-party company because you had too many unused tabs open is insane nonsense.
They're marketing it to people whose _Google Chrome_ is running too slow. That's already being exposed to Google.
Imagine all the enterprise customers who'd be willing to pay for this so their designers and engineers can be more productive.
Let's say you replace employee machines every 2 years.
Doesn't it make more sense to just spend 720$ extra per employee on hardware? It'll be a much better experience, with much less risk (what if mighty is unsustainable and closes down?), and that machine will still have value in 2 years, unlike throwing money at a cloud subscription.
- Mirror my tabs in the cloud? Great!
- Opt+Tab to navigate my overflowing tab bar? Sure!
- Cmd+J to instantly join meetings? This might be the killer feature for me honestly
- Search through all my google docs from anywhere? Sure, why not?
The problem is, I can get most of this through google calendar alerts and firefox extensions. I wonder how their value prop will evolve over time, because right now I don't see it being worthwhile for anything except crash recovery. With M1 Macs being as quiet and power-efficient as I've heard, it sounds like the main market these folks are targeting (execs/higher ups that aren't as tech-savvy) would rather just use newer machines?
Its VNC. And Cloudflare has pitched same [0] with different value dimension (security).
[0] https://www.cloudflare.com/teams/browser-isolation/
It's easier to use.
[0] https://twitter.com/Suhail/status/1337251861175230469
[1] https://twitter.com/Suhail/status/1259516709074956291
Low-latency (in terms of <10ms) high-fidelity (4k@60fps) video streaming is not a solved problem. It is hard engineering.
STOP CREATING MIDDLEMEN! It's going to cost me 30 bucks to just browse the web where I spend another dollar amount to where someone collects a "handling fee". Jesus I feel like the world is going nuts.
[0]: https://github.com/crosswalk-project/chromium-crosswalk/tree...
This doesn’t solve the Figma case tho.
It is an electron app, as unlikely as it sounds.
That sounds... just ridiculous.
edit: omg really
https://www.workatastartup.com/companies/mighty
I even have Firefox sitting at 5 GB ram usage right now for 150 tabs. I don't think I've ever had an issue with performance on browsing the Internet.
Most probably, the bottleneck is bandwidth/CPU for most users.
But Firefox is really very good with crazy numbers of tabs these days. https://metafluff.com/2017/07/21/i-am-a-tab-hoarder/
"Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something."
https://news.ycombinator.com/newsguidelines.html
2. Reliable low latency streaming on wired connections is pretty straightforward, and should work fine. This is an easy problem.
3. Reliable low latency streaming on wireless connections is an unsolvable problem due to the nature of physics (basically), and will be an endless source of frustration. There's a reason no FPS gamer would ever play on wifi by choice. It will work fine at times and then randomly start sucking right as you're trying to do something important.
4. If it turns out this is useful in some cases, Google can easily do a better job than Mighty. And there's no reason this couldn't be done by AWS and Microsoft as well. It's trivial for a major tech company to do this better than Mighty does. They already built Stadia and the rest. Unlike when Dropbox launched, these companies aren't sleeping on stuff like this anymore.
https://en.wikipedia.org/wiki/Opera_Mini
I used it as a kid to play Farmville (Adobe Flash based) on my iPad.
This is strictly worse for privacy than a VPN run by the same company, no?
If you are tunneling between home (A) and "vpn" provider (B) that terminates your traffic and sends out it to the public internet, and you are browsing websites C and D, both sharing the ad / tracking site E. 1. It's easy for anyone observing traffic between A and B to deduce from the traffic patterns what kinds of requests you are exchanging between C and D, 2. It's easy for someone obsrving traffic between A and B, and colluding with tracking site E, to correlate your address at A with your browsing behaviour on sites C and D by correlating A-B flows with information gathered by E.
In the Mighty type architecture, the leaked signal from the flow between end user and Mighty servers will be harder to match to the HTTP(S) traffic based on packet timing and sizes. But not necessarily hard enough unless there are purpouseful andi-TA measures employed. Hence I said "potential" privacy advantages.
Both of these scenarios place equal amount of trust in service operator - the difference is in security is against other adversaries.
(Traffic analysis is a term of art in security engineering & SIGINT: https://en.wikipedia.org/wiki/Traffic_analysis)
> Both of these scenarios place equal amount of trust in service operator - the difference is in security is against other adversaries.
This is where I would strongly disagree. With a VPN they can analyze your endpoints, but assuming E2E encryption, they cannot see your actual packets. Whereas Mighty is definitionally a MITM browser, they can see everything. Therefore you are trusting Mighty incredibly more than a VPN provider.
By my understanding, the web browser runs at the Mighty server farm and it's streaming it to you using a RDP or Stadia style system. So, it's more like you send 10-ish packets containing input events (like mouse motion and click) using the Mighty client protocol, and as a result the Mighty service does 150 web requests that happen due to loading the new web page (probably around 10k packets). While the page is loading, the Mighty service sends you screen updates that are again much different than the packets the Mighty service is receiving in response to the web requests.
Yes, these can still be temporally correlated, even though the correspondence is much more distant than in the VPN case (where you can just observe the similar sized packets). But there's potential to fix it rather easily, by eg using chaff traffic to the client in idle periods, and/or by pulsing updates out to all users to the service in sync, etc.
Also, Cloudflare recently launched a similar thing, but it's designed for situations where you want an employee to access a service but don't trust their browser: https://www.cloudflare.com/teams/browser-isolation/. That's the only situation in which this makes any sense, and even then, if you don't trust the employee's device you are probably hosting them a virtual desktop anyway.
Even if your authentication is password-free (say, TOTP plus pick-the-right-icon from a set of 64), you don't want a keylogger picking up anything else, do you?
> We’ve implemented cross-platform Drag and Drop file uploading. When you drag a file into a Mighty window on macOS, we simulate that same sequence of Drag-and-Drop events on Linux. We trick Chromium into thinking it's uploading a file from the Linux filesystem while, behind the scenes, we stream the file from the user's Mac; we accomplish this using Filesystem In Userspace (FUSE).
Most businesses amortise a laptop/PC over a number of years. Would you rather pay $10/month/user for this cloud service, or spend the additional $360 (laptop/PC lifespan for a business that can amortise the asset over 3 years) in the first place to get more powerful hardware locally, and benefit all apps rather than just the web browser?
I'd like to be proved wrong.
But why would anyone want to outsource their web browsing to a third party? If this is something you need, you should setup a homeserver and RDP into it...
Btw. My teeth gnash at the thought that my assumption above about the product is correct.