In a couple of days, it’ll be 3 months since I stopped working at Nordic Semiconductor. That feels like it’s been a nice amount of time to reflect into why microelectronics hasn’t been for me.
A particular phrase has stuck in my mind. One of my colleagues was showing around the cabinet full of consumer electronics using Nordic Semi’s chips. At one point they pointed to the medical-grade heart rate monitors and said something along the lines of:
It’s wonderful that we can help people.
That resounded in my mind. Does that mean that eveyrthing else that they do does not help people?
It’s been weird. There are indeed useful applications for all the bluetooth-enabled microcontroller stuff; there are elegant solutions and nicely thought and implemented protocols. But from my point of view, the number of weird crappy niche applications is way higher.
And it makes sense. Ever since I read The Goal by Eliyahu M. Goldratt, I’m sure about one thing: privately-owned companies only care about money. Nordic Semi is no exception, and the way it makes money is by selling chips to hardware manufacturers, and hardware manufacturers make money by selling as many consumer electronics as possible (doesn’t really matter if the end result is a novelty or actually useful at all). And somewhere in that pipeline, the agency of the final user is lost (agency: capacity, condition, or state of acting or of exerting power).
A fine example of this is the windowblinds of Nordic Semi’s HQ. The building is touted as a top-notch smart building:
An example is seen at Otto Nielsens veg 12 in Trondheim – an office complex which is well on the way to implementing ‘next generation’ energy solutions. Complex solar power installations cover the roof surfaces, and the surplus heat from the Bluetooth producer in one building heats the offices in the neighbouring block. All energy consumption and production are precisely measured. The electric grid, sensors and the Internet are all interconnected.
And yet, there is one week during the year where the sun angles just right so it was hitting my screen for six hours a day and I couldn’t get the friggin’ window blinds down because they are fully «automated». And this is because the architect et al didn’t design the building for comfort. Same article, emphasis mine:
Measuring energy consumption is important for machine learning, so that at a later stage we can allow the data to manage the optimum operation of the building. The information flow provides all the knowledge we need on where the energy production and needs are, so that we reap the maximum benefit. The mix of electric grid and Internet – which the smart grid represents – is on its way.
The system is optimized so the landowner saves more money, screw the tenants if they get sunlight on their faces for six hours a day because they’re not the main use case in mind. Geez, I wish some people would eat their own dog food for once.
At the same time, I’d be up for having a look into the building automation systems, and actually implement a windowblinds algortihm based on the position of the sun, digital elevation models, and cadastrial data turned skyline data. Wouldn’t it be cool to apply geodesy and astronomy to architecture? Alas, allowing tenants to do so is hardly seen as a priority.
Don’t get me wrong: it’s not Nordic Semi - it’s the entire industry. The whole internet-of-things fad is optimized for the middleman, not for the final user. It’s half comical that I now think of Nordic Semi as a mere delivery vehicle for the Internet of Shit. It’s all good as long as it sells more units, right?
I might be a weirdo becuase I want agency over the stuff I own, which is also why I’ve been on the free/libre/open source bandwagon for a long, long time. I’ve come to appreciate the agency that FLOSS can give me, and how it’s replaceable.
That’s why I’m very skeptical about Zephyr RTOS. I knowNordic Semi and some other electronics companies are putting effort into it, but it feels off somehow. The feeling of reading through the source codes and the modules is like different people have been just dumping code in there and hoping that other people will fix it. Which prompts the thought of «Wait, there’s stuff that’s been running this code? WTF?».
Maybe it just needs a couple of years to mature, but deep down I’m afraid that key people touting for open source haven’t read the Free Software Definition or the DFSG or the Open Source Definition. Publicly available doesn’t mean open source, folks.
Also, is something FLOSS when the compiled result is written to EPROM or non-writable, non-readable flash memory? (Because that’s how microelectronics firmware is done in the end.) I feel compelled to at least doubt it.
I can’t help but think it should be an ethical imperative to give end users of consumer electronics more agency over their stuff. To make things debuggeable, fixable, hackable. Cue anarcho-communist rant about how the owners and decision-drivers are the middlemen and not the end users.
Maybe I’m burning a bridge here. Maybe it’s just not the time for radical FLOSS in microelectronics (yet). Time till tell.
Im the meantime, I guess I’ll go back to making some maps. I’ve been ogling at that job opening at Geographica for long enough.