Hacker News Reader

It's pretty rare that I root for a company to crash and burn on principle. I'm an entrepreneur myself so it takes a lot for me to go there.

I hope every single one of these cloud-streamed remote-app or remote-OS plays fails and fails hard. They're helping lead the Internet and the computing ecosystem in an even more dystopian direction. I've been happy to see Stadia not really take off.

So lets say this succeeds. Then Google or Facebook buys it. Now all your browser sessions including passwords, keys, authentication codes, private messages, etc. are globally visible to be data mined.

Who's to say they're not doing this already?

What if this is hacked?

This is worse than that Amazon idea of giving Amazon delivery people keys to your house. In the physical world it's pretty easy to see people when they come in your front door. In the digital world you have no idea what these people are doing with your data. There is zero situational awareness.

I am in the same club :)

I don't usually care about companies success or failure, this none of my business, after all, but this kind of "innovation" could have extremely unpleasant side effects.

I hope they crash quick.

I think you kinda hit it on the nose. Who knows where or how or who has access to these machines. IDC if it's encrypted in transit or what, but there is no way a corporation with strict data privacy rules would be able to stream potentially sensitive information across the wire especially when it will be stored in the cloud in web form for a period of time. IDK good luck, but I'm definitely tin-foil hatting with this guy above me.

For me it's not so much a trust issue with this company, though for cloud and mobile stuff I have come to a "guilty until proven innocent" rule as regards privacy. It's (1) the trend this supports, and (2) what happens if worse players get access to it either through hacking or acquisition.

I guess I'm more thinking if the target is enterprise (because it's 30$ a month), what enterprise is going to green light workers using a browser where content doesn't reside on the user's machine? I've worked several tech jobs where it's mandated to use a specific browser because it's locked-down to not leak sensitive information. Not to mention it allows users to access internal resources. IDK I'm not necessarily hating the product, just don't know how it's going to work at scale for the listed CPU/Memory/Price point

While I totally agree with you, if this succeeds my hope is that it will finally push browser vendors to come up with a good authentication/authorization story. Make it totally integrated in the browser, such that I remain in control and Mighty only sees the equivalent of OAuth token it can't use to login in my name. No more custom signup forms, no more botched login flows redirecting you through 13 sites, no more passwords stored on websites... That is an innovation I would gladly welcome both as a web user and a potential web developer.

Every service needs auth. I can't believe nothing is properly integrated. I still have to click and enter a password, which fortunately the browser can create for me. I still have to receive an email and click on a link to validate my account. Web developers still have to create forms, manage the whole process, hash, salt and sauce my password and not leak it.

That already exists as Apple Sign-In: https://support.apple.com/en-us/HT210318

Not really surprising considering how Apple is so focussed on doing proper UX and privacy. Too bad only Apple users can use it.

You have valid concerns but no need to hope for their failure.

Tech people are the minority. The market IS moving towards cloud. It's happened, it's happening, it will keep happening. Stadia may have failed now, but it IS conceptually the future of gaming. It's like you're arguing for blockbuster in a netflix world. We cannot stop this from happening, no matter how many choirs we preach to. All we can do is find ways to make this happen better.

I think it's more constructive (and technically difficult) to accept that the market is heading to full cloud and we as tech people need to find better ways of making this vibe with good privacy practices.

Personally, I would not use anything like this without knowing a lot more about their security. Even then, maybe we're still a few years out from a security perspective before I would feel comfortable storing my passwords and browsing data with a 3rd party server AND pay for it (wild). But, I could see myself doing this if my privacy was ensured.

I hope these guys really focus on innovating in that aspect, and then I hope they succeed big.

Privacy is not the main concern here, although it's huge.

Imagine you buy into this service and they go bust: Suddenly all your history, passwords and cache: poof, gone.

It's exactly what local-first advocates tell us is the current enemy, not closed-source. See Kleppmann's latest blog post about the GPL: https://martin.kleppmann.com/2021/04/14/goodbye-gpl.html

> Tech people are the minority.

That's irrelevant, though. The "tech people" aren't preferring local solutions because they're funny this way - they prefer them because cloud-streamed remote apps objectively sucks. It takes some knowledge about computers to comprehend how and why exactly, but it doesn't change the facts.

(To use an analogy - doctors are a minority too, but you listen to them when they say you should vaccinate.)

> The market IS moving towards cloud. It's happened, it's happening, it will keep happening.

The important question to ask is, why. Why it's happened, why it's happening? The answer has little to do with providing value to customers - it's mostly about creating ability to seek rent. Privacy issues only happen on top of that - they're not the entirety of the problem.

> I think it's more constructive (and technically difficult) to accept that the market is heading to full cloud

Or, we could fight it. Maybe it's a quixotic quest. Maybe not. The market is a dumb greedy optimizer, it flows down the profitability gradients the way water flows downhill. If you want it to flow elsewhere, you have to put obstacles in the way, or cut out a better path.

> It's like you're arguing for blockbuster in a netflix world.

No, it's not like that at all. Nobody is arguing for going back to distributing software in boxes with floppies or CD-ROMs in them.

Reason from first principles, not by analogy. Context and details matter.

Here are some major reasons for the push to cloud. None of these reasons are immutable or universal.

(1) Wimpy mobile devices with constrained power, storage, and bandwidth requirements.

(2) Cloud is the only kind of DRM that works. It's a way to lock things up and make piracy virtually impossible. As a bonus you can still build on "open source" and placate the open source zealots who don't understand the current state of things and are still living in the 90s.

(3) Application delivery and installation/uninstallation are terrible. OSes are broken.

Here are some solutions:

(1) Moore's law, huge improvements in battery capacity, 5G, WiFi 6, etc. are eating away at this problem. This issue will die of natural causes.

(2) The hopelessly naive idea that "information wants to be free" and everything has to be "free" (as in beer) needs to die, be cut into a thousand pieces, burned, encased in concrete, and sunk to the bottom of the ocean. Nothing is free. Software takes a vast amount of labor to produce, and that must be funded. If it's not funded directly and honestly it will be funded indirectly and dishonestly (surveillance capitalism, cloud lock-in, etc.). "Everything has to be free" and piracy actually help push us toward a surveillance capitalist panopticon future.

(3) This might be the toughest problem. Windows is by far the worst offender here with its nightmarish installation subsystem. Closed app stores are another huge problem but eventually I think anti-trust action is going to chip away at that.

That's not by any means a complete analysis. This is just a comment on a HN thread. It does hit the major points I think.

Your point about security is valid. But we are already past that point when we started moving all of our apps and our data to cloud. Before Saas was popular, almost all of our data stayed in our local machine. But now everything is in the cloud. we have already lost the privacy battle.

No, we haven't lost. When all our devices are thin clients and everything runs in the cloud, we will have lost.

> And I fail to see why anyone would use this, you need high speed internet capable of streaming 4k for one and if you have access to that, then chances are you also have access to a sufficiently powerful computer capable of running chrome locally.

Plenty of people who can’t afford a fast computer currently have access to a fast internet connection. The ability to substitute internet bandwidth for CPU and RAM will be very valuable for them.

Most people don't need a fast computer. For many a 5 year old average computer is good enough in terms of hardware.

What makes this hardware not great is the many developers who have fast machines who are ok using a lot of it with the software they develop. This makes the experience on older systems slow. It's unplanned obsolescence.

For chrome stuff and using the web I shouldn't need a killer system. No one should.

Harder to find a fast connection than a PC. PCs are still cheap

> Plenty of people who can’t afford a fast computer

I’m pretty sure this set of people can’t afford 50$ for Mighty either.

On the one hand: yes, $50 * 12 months would go a long way toward a machine upgrade, so it doesn't make a ton of sense purely on your-machine's-too-weak grounds.

On the other hand, I don't really run Chrome or Firefox on anything that operates on battery, because I don't like seeing the little battery icon deplete twice as fast, and it barely even matters how powerful the machine is (M1 helps, but there's still a noticeable difference). Maybe there are people who really, really want to run Chrome all the time, but also work mostly on portables and like them actually lasting as long on battery as they're supposed to. Maybe that's worth $50/m to them.

I'm curious how much battery the video streaming would use here. Wifi tends to use a fair bit of power.

Good point. Decoding's usually pretty efficient, but you're right that use of wifi plus everything else related to this program might erase much of the power-savings.

You can get a more than enough powerful computer to run a web browser and more for $400 if you buy used, and you'll get to keep it forever. A subscription of Mighty would only last you a year for the same $400 price tag.

I'm not sure people who cannot afford a fast computer will be able to spend their money on a service like this.

I'm quite probably overlooking something and I'd be curious to learn what.

I imagine VNC can't do this well because it streams pixels with no optimizations other than antiquated compression (it can't even match WebRTC screen sharing), and crappy color depth.

The idea is interesting for lightweight computers e.g. chromebooks and ultrabooks, but it would irk me a lot to have my browser and personal information running on some other machine that I don't control.

What I would be super-interested in though is a self-hosted version of Mighty, that I could install on a Linux box anywhere of my choosing. For example, the server runs on my powerful desktop at home, and my ultrabook in the bedroom can be a client.

Yeah, but this is a VC backed venture, they won't do that.

The cloud could be the worst thing that can happen to the Internet.

Privacy and Ownership should not be treated as abstract ideas.

> The cloud could be the worst thing that can happen to the Internet

Is this not the same as saying "The internet could be the worst thing that can happen to the Internet"?

This project actually made me think that, since the X-Window protocol is practically a dead-end and everything's gotta be made with web tech now (ugh), it'd be really cool to have a version of FF or Chrome that's smart enough to send some kind of render instructions between a server-instance and a client-instance. Process server-side, render client-side, like X-Window but for web junk.

(the notion that this is completely fucking absurd since those "render instructions" are called "HTML" and I'm just describing server-side rendering isn't lost on me, but it's not my fault things have gotten so bad that having a server-side browser forward draw commands from bloated "web apps" to a resource-light client might actually be kinda nice)

Cloudflare browser isolation does this by forwarding Skia commands: https://blog.cloudflare.com/cloudflare-and-remote-browser-is...

That's interesting! Did they release any products based on this technology?

Yes, but it seems like it's only available to enterprise customers trying to allow selective access to things behind the firewall: https://www.cloudflare.com/teams/browser-isolation/

You can just use NX, RDP, Splashtop or even Chrome Remote Desktop.

You can also use Opera Mini in MicroEmulator for server-side rendering of web pages.

I see this being extremely valuable for companies that hire contract workers, especially for UI-intensive tasks. Say you do labeling for self driving cars where you'll have to render a lot of images to the end user. Rather than giving all these users a powerful computer, you could give them chromebooks and reduce your capital cost massively.

And spy on everything they do in their “browser”.

Agree, good insight!

This is already a thing and has been for years.

Many thin client setups like teradici PCoiP are used by tons of film studios in post production. The last few animation Oscar winners have all been done without computers at people's desks.

There's already services like Nimble collective (bought by Amazon) that streams 3D apps to your browser like Maya, Blender etc...

WebRTC is already seeing tons of companies move to streaming content to thin client end points. Epics MetaHuman creator for example runs in the cloud.

> And I fail to see why anyone would use this, you need high speed internet capable of streaming 4k for one and if you have access to that, then chances are you also have access to a sufficiently powerful computer capable of running chrome locally

That's a weird assumption. Where I'm from gigabit (or at the very least 100mbit) fibre is the norm, which means fast 4K-ready internet cuts across virtually every socioeconomic demographic.

Just because you have a fast internet connection does not mean that all your client devices have a lot of RAM or a GPU. Even if they do, pushing computation to the cloud could mean improved battery life when you are on the go.

Would be interesting to see how far you can take a raspberry pi with mighty.

How much lithium battery degradation is due to some mobile tab going rogue?

Isnt video streaming more intensive?

Seriously though. They must be doing some crazy magic to make this claim...

At their price point buying more ram is a cheaper alternative. Most folks have little use for dedicated gpu.

for the moment, I would consider the concept and not so much the price. What they charge is likely not a lower bound on their internal cost structure. The product came out of beta today, so their pricing seeks to first attract those users with a high need and to test their pricing capacity. Better to try to charge too much and then go lower than to take too little.

Given that their engineering expenses are a fixed cost and the majority of their spending, they'll be able to lower prices as they scale.