[th0ma5]

The hardware is very cool. The start of the integration with the dashboard is neat, but I'm not entirely sure how useful it would be other than making a video like this. Perhaps something like OSC or something could help keep it modular from a software standpoint and not so toy seeming. I also think a lot of software has whole knows what layers expecting hardware to not just suddenly disappear. Appear is maybe less a thing.

[ajsnigrutin]

Yep, looks great as a toy, but seems not-that-ok for any serious use, because for static setups, all the alternatives are a lot cheaper, because they don't have to be modular (eg. touchscreen display for controlling home automation).

But I like the concept for areas such as schools etc., because it makes it possible to try a bunch of different things very fast.

[vidarh]

The alternatives may be cheaper if you can find one that does what you want. But in terms of the ability to quickly throw something together where existing stuff doesn't quite do what you want or is locked down, it looks amazing.

In terms of "serious use" I think that depends on what you mean by "serious". E.g. my hone automation setup is not static, because I keep finding new little things I want to tweak, and often the ready-made solution lack options I'd like to have unless I'm willing to spend time tracking down very specific modules. My time is valuable to me - I'd happily pay extra for a modular system where I know if something is lacking a button for something I want I can literally just plug one in instead of having to search all over for a different model of something.

Of course this will not replace mass market "close enough" solutions.

But I think there are still more than enough people who want to tinker but don't want to have to whip out a soldering iron.

[ajsnigrutin]

You can always get the same components that this kit uses, solder them together and do it cheaper. Want a dimmer for your smart lights? esp8266 + one of many potentiometers available + maybe an adc if you need more than one. You'll never use this whole kit just for one potentiometer, with the rest of the components in a drawer somewhere, because this would make it too expensive. Cramming multiple modules just not to "waste" them, makes you build a tool around what you have, instead of using just what you really need.

[vidarh]

Yes, you can, but I have no interest in building my own. I don't need these things badly enough to be willing to invest much time in it.

Things like a single dimmer etc. is not the problem, because for simple stuff like that there are plenty of off the shelf options.

The issue is anything that deviates from what manufacturers expect. And yes, I'd be happy to use a lot like that for that even at a quite high cost because the existing options to do it cleanly without a self build are also expensive.

[unfocussed_mike]

Yeah. Fairly obvious applications in teaching/education devices. Though I think it should be noted that a lot of what schools want to teach with devices like this is precisely the pin connections and circuits this thing elides.

But I am not sure about the rest.

And what you don't see in demos like this are the practical limitations that the demo avoids. I didn't watch all the way to the end but I had questions about how the device communicates its limitations and handles situations where, for example, it cannot supply enough power.

[XorNot]

I feel like the power of this could be in something like building industrial control panels quickly? Being able to rapidly setup control boards, and equally rapidly configure them into something else would have a lot of utility in that application.

EDIT: Though that does make a key weakness the fact that there doesn't appear to be any provisioning for locking down the blocks more aggressively to the board.

[nsb1]

I don't know that I'd go so far as to call it a toy, but I do think that this device is more likely to be used in education or prototyping than anything else - which is still great! Just having a platform that will autoconfigure the array of devices that he's demonstrated is amazing. Being able to snap on two cameras and a TPU and have everything 'just work' so you can focus on your vision processing software is incredible.

I'm betting that it will be cost-prohibitive to have this device as a permanent fixture for any one application outside outside of the sorts of things an rPi4 already does (Home Assistant server, etc). I don't think anyone is going to use it to control an LED with a slider, but for prototyping that sort of thing it has endless possibilities.

Hats off to Anil for the huge amount of work he put into this project.

[Tade0]

As a third generation hoarder I have to say that the number one thing preventing me from doing projects such as home automation is that since I struggle with throwing stuff out, failed/deprecated devices would just pile up and I don't want that.

[goodpoint]

The magnetic connectors is what makes it an expensive toy. The pluggable grid layout is nice but it could be achieved with a daughter board with pins sitting on top of any cheap SBC.

[rpmisms]

It would be trivial to modify the design to allow some form of locking blocks in place, whether through a sliding tile system, a lever lock, or a pin design. You could even do it yourself with 3d printing. Really depends if there's a professional environment that could use something like this. Maybe a lab or machine shop?

[jsmcgd]

I think a lego compatible version could be popular.