by chidiw on 11/13/22, 3:23 PM with 318 comments
by nisten on 11/13/22, 4:27 PM
Before the value proposition was good. You can still use your old laptop and still have a very fast browser experience via remotely running it on a AMD server chip, this way you save on battery life and RAM too.
Then M1 speedometer scores started beating AMD Epyc on single core web browsing performance. Not only that, but even providing a much better experience in terms of battery life because of power savings.
Then yeah there was no point paying 20$-30$ a month for a remote browser instance. If you can afford that as a user you're either already set on CPU/GPU/RAM or better off just buying a used m1 air for <900$.
It's quite sad of a lesson where you do everything right, work consistently hard, be quite innovative, have plenty of financial and talent resources, and still fail.
Tech obsolescence is brutal.
by danso on 11/13/22, 6:36 PM
At $30/month, this sounded like something that could only be worthwhile for businesses (though I'm at a loss to imagine what size/industry of company would want this). Did Might have any corporate clients, and if so, how did they navigate the security/privacy compliance issues?
by obblekk on 11/13/22, 4:48 PM
That said, I work with lawyers as customers and I suspect there was an adjacent market in data privacy that a hosted browser would have had PM fit for.
A lot of industries have strict legal requirements on controlling employee access to data (think healthcare, legal, compliance). In these cases, SaaS becomes risky if it can be accessed by anyone off their work computer.
The standard solution is to limit the webapp to be accessible on VPN, and limit VPN to be accessible on MDM controlled devices, and limit MDM to be accessible on company owned devices.
It would be a lot easier to just control the browser viewport and prevent data harvesting (essentially what Mighty was doing, focused on performance).
Sounds like a small change, but it makes a world of difference when it's not feasible to send a physical machine to someone (3rd party contractors, overseas employees, low wage employees, etc)... something I discovered when talking to lawyers about using contractors overseas for routine data entry tasks.
But that's definitely not the same kind of user, or tech expertise and would have been a big pivot in itself. Hopefully someone buys the IP and builds something cool.
by talhof8 on 11/13/22, 4:35 PM
Probably high server costs and being a vitamin and not a painkiller killed it.
Also, people are not too happy about giving away their browser history data to server-side powered browsers.
That being said, it's still sad to see startups fail. Hopefully they'll have better luck with their new direction. Fingers crossed!
by SergeAx on 11/13/22, 9:02 PM
I beleive, by the way, that the idea of a fast browser for cheap devices came to founder after reading story about Browserling, QA multibrowser suite, going viral in India, where people with $20 phones were able to use WhatsApp webapp via Browserling free plan (HN post: https://news.ycombinator.com/item?id=16103235). Coincidentially, I heard that voice room app (I am not being sarcastic here, I really forgot it's name and don't think it's worth googling) also became succesfull in India.
by doorman2 on 11/13/22, 5:58 PM
by stephc_int13 on 11/13/22, 6:51 PM
I think it was already dead at the time and the criticism got under his skin.
I wish there was a way to make money by shorting bad ideas. I think I could easily fund a few companies by doing that on weekends.
by modeless on 11/13/22, 4:39 PM
The death of this product is a good sign for computing overall. The more of computing that happens on the client the better. Clients are (at least sometimes) under user control. When stuff moves to the cloud users give up control and that's a bad thing. Long live fat clients!
[1] https://twitter.com/Suhail/status/1588906086459150337?s=20&t...
by ivoryloom on 11/13/22, 3:40 PM
by thih9 on 11/13/22, 7:41 PM
> We're excited to finally unveil Mighty, a faster browser that is entirely streamed from a powerful computer in the cloud.
Source: https://blog.mightyapp.com/mightys-secret-plan-to-invent-the...
by datalopers on 11/13/22, 6:50 PM
by andybak on 11/13/22, 6:16 PM
by mbStavola on 11/13/22, 3:53 PM
by pavlov on 11/13/22, 4:25 PM
It was also quite expensive to run, and the stack was fickle because those Nvidia GPU servers aren’t designed to run client apps — they’re AI/ML solutions primarily. So I’m not surprised that Mighty is shutting down: the infra bills must have been running pretty hot and they seem to believe they can better spend their cash runway on AI, presumably on largely the same infra.
(Btw I’m giving a talk at the Kranky Geek virtual event on Thursday Nov 17 about cloud compositing which will discuss this Chromium adventure and what I believe is a better solution for most apps.)
by anhncommenter25 on 11/13/22, 5:44 PM
I can't think of a bigger slice of consumer compute than browser workload; so if the scheme doesn't work there, it won't work for anything. My conclusion is that compute arbitrage isn't viable for B2C. You will have to actually provide a service on top of the resale. For example Github Codespaces is reselling cloud compute while simultaneously solving infra-as-code pain points in the CI/CD pipeline.
by _sohan on 11/13/22, 4:20 PM
by rvz on 11/13/22, 3:57 PM
Doomed for failure from the start.
by ec109685 on 11/13/22, 5:48 PM
The combination of Intel Mac Laptops that had horrible thermal characteristics, browsers that didn’t do anything to throttle background tabs, bloated websites and limited memory created a perfect storm of awful performance that Mighty could address.
Almost all of that isn’t true anymore (besides the bloated websites which browsers manage better), which decimated the potential Mighty market (people willing to spend hundreds a year on a fast browser).
by kilroy123 on 11/13/22, 9:23 PM
On my work computer, I keep everything separate. Chrome is work stuff. Mighty was my personal stuff, with little to no files stored locally. When I run all the docker containers to stand up the local server. My computer is very taxed for resources. Mighty was amazing and not bogging down my computer with a second browser open.
However, on my big beefy personal M1 Max MacBook Pro with 64 GB of ram. It actually felt slow. I found myself reaching for regular Chrome more and more.
Very niche use case. IMO they needed to pivot to B2B and not some niche B2C play, like what they were pursuing.
Overall, it worked impressively so. Kudos to the team for shipping something so stable that worked.
by dhirenb on 11/13/22, 4:32 PM
But, Apple stepped in with the M1/M2 and totally obviated the need for any product like this.
by robenkleene on 11/13/22, 4:27 PM
But all forms of coding and media editing (the main tasks that are truly CPU/GPU-bound, the area where something like Mighty would help) have remained as local desktop app affairs.
I think what really happened is folks misinterpreted the significance of Figma. The web has taken over in collaboration-focused software, e.g., things like Google Docs, Slack, etc... What really happened is that design has moved from specialized professional software (hard to use, powerful), to being more like collaboration software (easy to use, light).
I.e., what Figma really signals is that design is now more like new task that became more like using Google Docs app, not that high-powered tasks like video, photo, and audio editing are becoming web apps. This means what really powered Figma's takeover was that flat design took over, which is less technically demanding.
by landonxjames on 11/13/22, 4:43 PM
by amadeuspagel on 11/13/22, 4:00 PM
[1]: http://paulgraham.com/newideas.html
[2]: https://news.ycombinator.com/item?id=27062713
[3]: https://www.echevarria.io/blog/the-mighty-pushback-isnt-all-...
by funstuff007 on 11/13/22, 9:26 PM
https://twitter.com/paulg/status/1357097710734749700?lang=en
I dig when billionaires are wrong. They're just like us. :)
by fareesh on 11/13/22, 4:23 PM
Can the underlying technology be used to create a first class remote desktop experience? From the ones I've tried, I've found that NoMachine is among the best performing remote desktop tools. I never used Mighty but my guess is they must have achieved better performance than this? Was Mighty's technology tailored to Chrome only?
by stavros on 11/13/22, 3:59 PM
After significant searching, I found it:
Basically Stadia for webpages.
by skc on 11/13/22, 4:55 PM
Always seemed like a very, very niche product .
by crubier on 11/14/22, 3:36 AM
by QuadrupleA on 11/14/22, 5:22 AM
Running a browser to connect to the cloud to run a browser to connect to the cloud to retrieve the contents of a single 2D page to recompress and send back to the original browser is now "the future of computing".
by ykl on 11/13/22, 10:01 PM
[1] https://www.allencheng.com/starting-a-business-around-gpt-3-...
by lmeyerov on 11/13/22, 7:09 PM
IMO Mighty is too early bc they didn't tackle the SW side, 'just' the remote nature, so couldn't get their 10X. The 10-100X shift today comes from apps running differently on cloud resources, not just lift-and-shift. That already exists as the VDI market.
Ex: We do client GPU <> cloud multi-GPU for cyber/fraud/etc analysts wanting to visually investigate how their many events stitch together, and Otoy does same for movie effects. In the consumer market, seeing same rewrites in say gaming. In a sense, Mighty's pivot to generative AI is the same -- faster to use remote multi-GPU services.. but custom built accelerated visual app for 100Xs, not lift-and-shift.
Without tackling the SW rewrite problem, hard for Mighty to get these wins via a lift-and-shift :( I do think federating cloud resources so users can bring their own and devs can reliably tap them would be amazing, but we seem still in the dev-controlled-server era. As someone trying to build predictable 100X experiences, that feels likely for awhile. Such a shift could have been webgl2 etc standards enabling modern multigpu apps in the browser... but Google and Apple web GPU browser/standards leadership have long strangled that path.
In a related note, we are starting to launch our global GPU edge network, and looking for a backend/infra eng on that + related AI services build out -- Nodejs/python/k8s. We are profitably growing: I agreed with Mighty's pattern before it even existed (my PhD at Berkeley explored it!), just easier ROI right now by sticking at the app layer.
by arriu on 11/13/22, 4:24 PM
by ianbutler on 11/13/22, 4:18 PM
Instead of a faster browser they did super easy edge deployments of your JS apps which has a similar end result for users but also solves a giant headache for businesses which is where the money is.
by searchableguy on 11/13/22, 4:13 PM
So much interest and explosion.
by langitbiru on 11/13/22, 4:47 PM
by dandongus on 11/13/22, 8:14 PM
by wilde on 11/13/22, 4:11 PM
by miketery on 11/13/22, 5:00 PM
I’m tired of all these apps constantly doing stuff in the background and tracking my location anytime I use them.
If I had a “mighty” phone app, and could stream an emulated device with the apps on that server I’d happily pay $5~$10/ month.
by abetusk on 11/13/22, 5:04 PM
If Moore's law (or it's equivalent in compute/dollar) really ever does level out, then this will become more attractive. With the cost of compute dropping exponentially, it doesn't seem like this can ever really get a foothold.
by guynamedloren on 11/13/22, 6:33 PM
by abol3z on 11/13/22, 7:12 PM
by dietsprite on 11/13/22, 5:17 PM
by bagels on 11/13/22, 6:33 PM
by iworshipfaangs2 on 11/13/22, 8:20 PM
by raverbashing on 11/13/22, 4:48 PM
As other comment said, it's a vitamin, not a painkiller. And you can always buy a new computer or upgrade it. Would it really make me much difference to run a browser in a much more powerful computer than the one I'm at? Only in extremely limited situations
by hmate9 on 11/13/22, 5:39 PM
by smaudet on 11/13/22, 7:19 PM
The problem with too many tabs wasn't primarily performance, although that was a part of it (disabling background tabs is a better optimization), it was that there were too many tabs.
So, cool tech but I don't think most people needed it.
by fulafel on 11/14/22, 9:13 AM
I feel my web experience is mostly bound by other browser things than JS. I wonder what would be the best way to profile this.
by yubiox on 11/13/22, 8:56 PM
by rognjen on 11/13/22, 5:27 PM
by theGnuMe on 11/13/22, 6:11 PM
by xwowsersx on 11/13/22, 5:59 PM
by memish on 11/13/22, 4:19 PM
https://twitter.com/Jonathan_Blow/status/1387094702139142145
> Wait, Paul Graham supported this?
Jonathan: Very vocally, to the point of insulting world-class programmers who think it’s dumb (he did not know said programmers were world-class).
> Huh, I see. I thought he's all pro elegant and lean solutions, judging by the books and articles at least.
Jonathan: Yeah it doesn't make sense to me either.
by IceWreck on 11/13/22, 6:04 PM
And the few people who really had that usecase could just use a lightweight container running chromium + guacamole on a nearby server. Okay I'm sure Mighty used cool tech to make it faster than just that (idk how it worked but maybe they were transmitting changed HTML) but at its core the idea was not something normal people and 99% engineers would ever consider. Anyone rich enough to pay for it would just pay for a more powerful machine.
Anyways, they said they have 50% of VC money left and will be using it to create an online stable diffusion thingy which atleast fifteen startups are doing already including stability AI themselves. How can they differentiate ? Can they just use leftover VC money to make a completely different thing ?
by tpush on 11/13/22, 4:08 PM
by OmarIsmail on 11/13/22, 4:30 PM
I think this is a really interesting decision from Suhail and one not taken lightly. I think this decision is another data point that AI-infused applications are a potential new “tech platform” and we’re at the beginning of a new “mega cycle”. I.e. web2 2002-2010, mobile 2010-2020.
Folks shouldn’t be spending much energy on thinking about how Mighty may have gone wrong (and seriously the negative analysis is so boring and lame), and instead think about the new AI opportunity.