I love the motivations behind the idea, but not the idea itself. I think there's a better way. Dynamicland undoes half the progress we've made in the last few decades, particularly when it comes to inclusivity. The amazing thing about computers is their ability to convert information between representations at will, and to transmit them anywhere, instantly. This is crucial for accessibility. Not being tied to a single representation means you can access and manipulate information with any method you want. It might be a touchscreen, a keyboard and a program that reads information aloud for the blind, or a voice recognition / eye tracking system for those who can't use their hands. With Dynamicland, you need to manipulate physical objects directly. If, for some reason, you're unable to do so, you're completely locked out. Contrary to what the article says, I think Dynamicland could strengthen the young white male living in California programmer culture even more. The beautiful thing about computers of today is the ability to learn and hack on programs, for anyone, anywhere. Open source even allows anyone to contribute, and being an OSS contributor gives you an advantage when trying to find work. Remote work is also becoming more and more pervasive, enabling people from poorer countries to work in tech without physically moving somewhere else. Dynamicland makes all of that significantly harder.
A better idea which might solve some problems it aims to solve, without sacrificing inclusivity, might be going back to the Unix tools philosophy. Tools should do one thing, do it well and do it only. They should also work well with other tools. The design of smartphone operating systems forces apps to do the exact opposite. They're sandboxed and prevented from accessing most features of the OS, as well as from communicating with other apps effectively. That forces developers to put more and more features into their apps. It also gives many opportunities for corporate lock-in and user-hostile tactics. The Unix philosophy would alow anyone to contribute something, no matter how small or insignificant. There should be some verification, of course, probably similar to what Apple does. Forcing all the tools to be open source wouldn't even be such a bad idea after all. They wouldn't be big apps, just small things scratching small itches, so there would be no big investment needed to develop such a tool. I think such an approach would let us create a much more vibrant and user friendly ecosystem.
I've followed Dynamicland for a bit. I'm completely enamored with the project, and yet I agree with you, the implementation is...not right. The goal is admirable but the approach feels like a step away from it, not towards it. Still, that's likely a side effect of simply trying to break the mold of computing we've all grown familiar with (even as it morphs away from the PC and into mobile and IoT).
The primary indication I see that the project is at least somewhat flawed is this: I don't think you could build Dynamicland's systems in Dynamicland (the same way you might write a new C compiler in C). I don't think that's their target, but I don't believe you can accomplish their mission without that characteristic.
A better idea which might solve some problems it aims to solve, without sacrificing inclusivity, might be going back to the Unix tools philosophy. Tools should do one thing, do it well and do it only. They should also work well with other tools. The design of smartphone operating systems forces apps to do the exact opposite. They're sandboxed and prevented from accessing most features of the OS, as well as from communicating with other apps effectively. That forces developers to put more and more features into their apps. It also gives many opportunities for corporate lock-in and user-hostile tactics. The Unix philosophy would alow anyone to contribute something, no matter how small or insignificant. There should be some verification, of course, probably similar to what Apple does. Forcing all the tools to be open source wouldn't even be such a bad idea after all. They wouldn't be big apps, just small things scratching small itches, so there would be no big investment needed to develop such a tool. I think such an approach would let us create a much more vibrant and user friendly ecosystem.