I'd love to see more of the bleeding edge desktop software like the Adobe stuff, Blender, all the DAWs for Audio, etc. But it feels to me that not very many programmers have the skill to pull something like that off. Also, there doesn't seem to be much of an incentive left to invest into such things.
I would very much want that the mind set changes a little bit. Great desktop software makes a profit from selling to customers who value quality and awesome features. I wonder why that is not enough incentive for ambitious programmers. I only ever hear about developers for mobile and web apps getting a lot of money for things I seriously don't care about.
I have the feeling that we use desktop computers for the same things we used them ten years ago. I can't even remember when I had the last "I need that!"-moment for desktop software. So why can't we reinvent the desktop with software, like many try to reinvent what phones and web apps can be used for?
It's like the desktop computer in this day and age is slowly reduced to a platform (OS) for a platform (browser) for web apps. That's probably why many can get by with a tiny laptop or even a chromebook, which is probably the culmination of that thought.
However, I've not lost all hope. My bet is on virtual reality. If someone can build a really awesome VR application where everybody's like "I need that!!!", I can see how the desktop could return to its former glory, considering that a decent VR needs a ton of compute. But there's still the problem that VR is hard, GPU programming is hard, 3D is hard, parallelization is hard and optimizing is hard.
So are there any pioneers left with an enormous amount of skill and time who would risk to start such a glorious and uncertain journey without some VC guy dropping a few million dollars on the table? Or are we all going to end up as web developers because that's where the capital seems to be?
By contrast I think C++ and realtime audio programming is simple, but I've slowly picked it up over almost 20 years. I retreated back to desktop apps and firmware. I'll take another stab at learning web technologies, but not at a startup where I need to deliver asap and learn at the same time.
As much as developers hate sandboxes, the lack of one (except Mac OS until recently) is why I think desktop is failing. Users just don't feel safe installing random software like they do going to websites or installing mobile apps. Users have been conditioned to only install software from sources they trust, and they trust no one.
If I post a link to my new web project, most people will click on it. If I post a link to an .exe almost nobody will download it. I think that's the main issue killing the desktop. Big downloads, compatibility issues, slow installs are also an issue, but I think they are secondary.
The goal for the desktop that I was talking about had nothing to do with these webby things. These should stay on the web and do their thing. You can have those small simple programs as a web app and on your phone. I think the downfall of the desktop is that people use them as if they were feeble things with a mail program and a web browser and some office programs on it, when they're actually capable of much more, which isn't leveraged except for some professional programs and games.
The desktop is - in my opinion - not primarily meant for little things. The desktop can and should handle the bleeding edge of high performance, parallel and accelerated computing that can't possibly run on anything else but a desktop. If we would stop thinking about whether a program will run well on a phone and explicitly target real computers, there would be a lot more purpose to desktop computing.
I'm not sure what that might be that doesn't exist, yet (maybe VR). But I'm sure that you can't do this all by yourself in a windowless room over a weekend, it'll probably be a large project involving quite a few programmers.
Then, the question of whether users will download your software is not the same as whether they will click on a website. Then, it's the same question as will they download Photoshop, will they download Reaper or will they download Maya? Will people download the next big thing in desktop computing?
A lot of medical applications (signal processing, visualization, robot control) tend to be traditional desktop applications.
Also military applications for analysis, and planning are often desktop based because they can not use consumer mobile devices nor can they rely on the "cloud".
Mac desktop app (native Cocoa + OpenGL, no web views).
Desktop is pretty much necessary for this tool because it needs to integrate with Xcode and Android Studio.
Because latency is an absolute premium for playing music in real-time across the internet, we've focused first on the desktop, where we have the most control ... but still find a ton of challenges.
If you saw this recent article about Android and audio latency, you can see why there are challenges in the mobile space: https://news.ycombinator.com/item?id=9386994
If you were to try and build a web-only version of JamKazam, WebRTC is your best bet but it does not have low enough latency, either.
Have you tried to tell about your product on reddit? There are many subreddits interested in tech like this and in indie music production.
Just built with web technologies vs natively compiled binaries, which is an implementation detail that shouldn't matter to the end user, in my opinion.
But to answer your question anyway, this is the responsible line of code:
http://microisv.com/ has some info.
Andy Brice's blog, Successful Software, sometimes has articles about other mISVs than himself. He has a product, Perfect Table Plan, that has done quite well. And he recently created another product that has also started doing well.
There's always a lot of new game studios, but I don't think they count. There's no end-game in games development, no lock-in, no network-effects, no industry-standard-status. A single bad game can ruin the studio.
For example, a single bad product/service can kill any startup, not just a game studio.
Madden NFL is not the only example of a "safe", long-running game series (franchise). Wikipedia has a list:
We are bootstrapped. An app like ours probably wouldn't make sense for a VC to back but it feels really great knowing that people from all over the world use it to backup their external drives to Dropbox or collaborate on the files on their NAS using Dropbox.
At one point you might feel out of the trend for not building a mobile or web app but I guess what matters is that your app helps someone, regardless of it being a web, mobile or desktop app.
Granted our desktop app is a highly customized version of our hosted web app. We use NWJS to package up our Meteor app (along with elastic search and mongo) to create a secure, local-only version of our personal web search tool.
We've been growing our customer base steadily. So the answer is yes, lots of startups doing desktop apps.
We're ultimately trying to build a decentralized object store Platform, but our initial approach is a desktop application. We're hoping that the majority of embedded and streamed content on the web is stored on and fetched from our Platform.
We have a native Mac and Windows app which uses Chromium embedded to wrap our core web app then adds our call stack (based on PJSIP) so you can make sales calls right within the app.
not counts. sorry evernote, I loved you though.
It's exactly the sort of service that I want a desktop app for too. I mean, I stopped paying after it deleted a bunch of my content, but still…
It's Windows only / .NET
Probably funded by YC IIRC.
Think of a traveling sales rep who is on the road and needs access to a product catalog and pricing information to take sales orders. HTML5 doesn't quite cut it in such a scenario.
This is in fact one of the most common reasons for companies to look for native apps.
Ward Bell talks about it briefly here: http://www.viddler.com/v/94a02c0f
There are a few google talks about it on youtube too if you look. It's great for enterprise.
i.e. 
Also, the reasons people make native mobile apps rather than hybrid apps (which run inside a mobile web-player component) are speed and reliability. Famously, Facebook switched from HTML5 to native for their iOS app: