Where is the next big (little) think in home automation?

At some point last millenium I had control over the heating and air condition for a reasonably sized building. There was a GUI or you could telnet into the machine it was running on. At last resort you could go up to the roof where there was a room full of bit switches that made a really satisfying clunk when you threw them (do people still throw switches or has that gone out of fashion).

The system wasn't very sophisticated, it basically knew if a room was supposed to be in use at that time or not and what temperature it was supposed to be if it was (or indeed wasn't).

There were sensors so it knew what the actual temperature was in each room and it could control valves to let hot water into radiators or cold water into HVAC units. What more could you want?

Well the thing is that as companies have tried to bring this sort of thing into the home they have given people systems that learn the times they are in the house and allowed control from anywhere in the world. This has often been done by pulling a lot of the control aspect of the products away from a computer that is attached to the the systems directly and into the cloud.

Which would be fine except that there have been a number of situations where this had lead to the same sort of security flaws as with the Internet of Things or the cloud services being turned off so the hardware in people's houses isn't smart any more.

You still see the 7-day all-in-one controller and thermostat unit, the only visual difference being that they now tend to be white instead of beige and just of a pain to program, although some of them are now wireless. But they still only tend to control one service. 

In boutique hotels and karaoke suites you get multi service automation, one touch button at the door turns everything on/off and puts it into moods, but these are just flipping relays and you can't say "I'll be back at 6:30, make it 22° and run a bath"

Who is taking the best bit of all three approaches, smart(ish do we need things to learn our habits, just tell them, or give them an ical feed), all the processing power in the house so it doesn't get bricked by the supplier going bust or being bought out, multi-service "lights, camera, action", and securely controllable from outside the house. Okay two of those may be contradictory, you need some remote reliance to get the message through but if that is all you lose when it breaks, or indeed if you could replace that service because it is documented not proprietary.

Where should I be looking for the friendly packaged control software in a box, with minimal secure external services, and a decent sized set of interfaces into other systems?

That pulse oximeter scandal

Something has been bugging me since I first saw the story that Pulse Oximeter Devices Have Higher Error Rate in Black Patients other than the very obvious racism.

Say you have decided that the way in which white people decide they are the default and don't bother to do any work to see how the technology they sell affects people with different skin colours is a lesser evil than actively joining the clan.

Say you also accept that not a single one of the companies that makes pulse oximeters managed to see a copy of Effects of Skin Pigmentation on Pulse Oximeter Accuracy at Low Saturation (April 2005) or similar.

In order to forgive oversite in this matter you also have to believe that collectively the companies manufacturing these devices  and/or integrating them into more complex products have at no point seen any coverage of the controversy around Apple Watches on dark skin, which to be frank was everywhere five years ago.

I don't know about you, but as someone working in the technology product space, my first reaction whenever there is a story about a product failing in a similar space to mine is to go and ask the specialists "are we vulnerable to the same problem" because (and shamefully so) in terms of reputation damage, worse than being called out for racism, worse than being in the papers/Private Eye/The Register for your product being broken, is being the company whose product is still broken in a way that everyone noticed five years ago and fixed. After all while learning from your mistakes is very important, learning from other people's mistakes is better.

So either there are loads of product types in medical technology that are failing people because they don't engage with the wider technology space, or they spotted this and decided to keep their head down to avoid costs, or worst of all pulled on white hoods and decided that non-pale-skinned people weren't worth R&D time.

As I said at the beginning, there are those that are happy to dismiss accidental racism as acceptable and I'd be lying if I said I was confident I'd never done it myself, but in this case people are actively not doing their job.

P.S. If I have failed to spot someone more appropriate to make this point posting on it, please get in touch and let me know and I'll promote their writing instead.