The Future of Microcontrollers, as seen from 2015

How can you choose a microcontroller for your next project when there’s too many options?

We’ve seen this before

In the 1970s and 1980s the personal computer market was taking off. There were hundreds of brands, all mostly the same. Sound familiar? Today a handful of them now live in museums, while the rest were recycled for parts. The market seems to have decided that there’s room for three choices: Big company A, big company B, and DIY models (everybody else).

The state of affairs

The microcontroller market is flooded. Raspberry Pi has at least four versions. Arduino has nine. Then there’s every other board struggling to get a foothold in the market. It’s so bad even Intel can’t get anyone to care about their devices.

The latest Arduino? Stinks. The latest Pi? Don’t care. The features they’re adding either don’t matter or are rushed and kinda broken. This should be a giant red flag. I’m in the target market, and it’s not worth the effort to switch from a board that I know works to a board that might or might not.

The bad future

The saving grace for these new boards is to get into schools and hook the kids when they’re young.  Stay with them all the way through school and they’ll be too invested to switch when they’re older.  They’ll make great new inventions and do awesome things!  … but the technology behind those inventions will only update once a generation.  I believe this is a local maximum – it might look like the best we can do from here, but there are other ways that improve better, faster.

The good future

I believe the next big thing for microcontrollers is not more ram, or more pin outs, or even more sample code and tutorials.  Those are nice features, but they don’t address the central problem: ease of adoption and use. Apple is #1 because they’re the easiest to use. Arduino has carved out a market by being simpler than others. Take that to it’s logical conclusion, and you have a microcontroller that doesn’t need to be programmed. I believe the next thing is a microcontroller with a microphone and a web connection. A service like Siri will understand you and write the code for you. I wouldn’t be surprised if most of the code lives in the cloud and pokes your microcontroller to do things at the appropriate time, like the dumb network terminals you can find now in large organizations. I would be equally unsurprised if the on-but-not-busy microcontrollers were helping the service to convert speech to code.

Final thought, paranoia edition

Thought the first: If you still believe you privacy in a phone call, remember that Siri can convert any phone call to text and then store it in a database, forever, just as easily as an email.  “Online Privacy” is an oxymoron.

Thought the second: There’s nothing stopping Siri from having a voice with emotion instead of that flat tone it normally has. So what if Siri has decided to keep the flat voice? Maybe showing emotion would sound too human and creep people out.