The Internet of Things has a dirty little secret(medium.com) |
The Internet of Things has a dirty little secret(medium.com) |
There are already 100's of millions, if not billions, of IoT devices in the field, monitoring and controlling electrical, transport, logistical, agricultural, manufacturing and other infrastructure. There are hundreds of companies that are highly active in this space and doing quite well. Some are new, some are old. They operate at every layer from device manufacturing, to data collection, to analysis and more.
The consumer stuff, like Nest, or colored light bulbs are just fads, IMO. The only area where I think IoT may have some viability in the consumer space is healthcare. I do not mean things like jogging bracelets or the other "wellness" silliness that gets sold directly to consumers. I mean devices that monitor actual health metrics (blood sugar, heart rate, etc). There are large trials being done by health insurance companies right now to determine if active and constant monitoring will help reduce costs and improve outcomes among high risk populations.
Anyway, if you want to keep on top of IoT, ignore the SV hooplah and read industrial-focused rags instead.
And its one thing to have a IoT thing, but how you consume it and what you can do with it needs to be pretty good otherwise a lot of value is lost.
At the moment, lots of the IoT opportunities is where the value of the information / control is pretty high.
Consumer household and personal Gadget IoT is interesting, and yes, sort of fadish at the moment, but I think it will develop, manufacturers as standard will put in the cheap chips that allow their things to be connected. But the big thing is, what does it connect into? Ideally some open platforms, but more likely the likes of Google will start hoovering IoT info / providing control and incorporating it into their platform. It does a lot to try and understand your patters of behavior, so seems like a natural extension that it knows what info to tell you about from your IoT devices, and what kind of control you want
Interestingly, most of the scrutiny of these early entrants into consumer IoT come from software developers themselves, who are already well aware that "all software is shit." (In the tongue-in-cheek sense of course)
That's one angle. Another one is that people who have enough understanding of how "smart devices" exactly work can see that cloud-dependence is ridiculous engineering, if you look at the product as something designed to provide value to users. The question becomes, whether or not it bothers you that the companies building this stuff are actually user-hostile. Personally, it does bother me - that's why I steer clear of them and warn people about this.
Not many products can claim to be ROI-positive to the consumer. When the payback period is short enough, that's pretty exciting for the analytical part of consumer thinking. Paired with the "looks good on my wall" and the "burns less fossil fuels" triggers on the emotional side, that's pretty appealing to a bunch of people. (Same basic feelings but much cheaper than buying a Tesla.)
No excuse for being buggy though.
The website says stuff like
> The babysitter calls to say she picked up the kids from soccer and they’re heading home. You adjust the temperature from your phone so they’ll be cozy.
..which is a first-world problem if I've ever heard one. Just wear an extra layer for 10 minutes while the heating warms up. Jeez.
Personally, I do have interest in a thermostat that can learn how to be more energy efficient for the sake of saving me money. But I don't think that requires Internet connectivity at all.
And it is beyond me why anyone would want or need an Internet connected toaster or coffee maker as well.
1) Peak shaving. The utility pays homeowners to be able to change the temperature slightly during peak electrical load.
2) Short term adjustment. One challenge of integrating large amounts of wind and solar into a grid is that sometimes you get short-term fluctuations in power generation. Smoothing these fluctuations out by turning off a bunch of residential AC units can be very cost effective.
Utilities are willing to pay consumers in both of these situations. And both of these work better with 2-way connectivity.
Looking forward, if electric cars become more popular, it's likely that they'll have smart chargers which will optimize when overnight or at-work charging takes place.
I'm working on Vanadium (by Google) [1], which will allow all this stuff to be server-less and secure (by default). Once this or things like it are adopted there's less a need for revenue to pay for the backend, nor the ability/need for companies going out of business to brick their devices.
[1] http://v.io
It's open source. Not everything Google does is directly for money.
Update: didn't see your brillo comment. First I've heard of it actually. Looks like Brillo is an OS and meant for embedded devices -- Vanadium is Go code that works on iOS, Android, your desktop, your server, your raspberry pi, etc.
I just came back from a "smart home" trade show, talked a bit to the regular companies doing regular home automation. It's expensive, yes - you get industrial-quality products and a solution designed (and priced) for your particular needs. Basically, I'm talking about companies that were doing IoT for the past two decades, long before anyone ever thought of that acronym. I have a few observations:
- Everyone does exactly the same thing. HVAC, alarms, lights, window blinds, access control. There isn't really much more to sensibly automate at home for an ordinary person at this moment. So they try to differentiate by look&feel, UI and the type of installation - e.g. whether it can be laid over an existing building easily, or whether it would require to dig up the entire electrical installation and is therefore best done during construction or big renovation.
- Everything runs local first. As it should. Most solutions are wired, some are wireless - via Zigbee, LoRa, Wi-Fi, or some protocol, but still over a local network. Cloud services are often added to enable remote viewing and management via mobile devices, but this is a bonus, not the core thing in the installation. One company I talked with today ensured me that their pretty mobile apps / tablet control panels have configurable network URL, so you can expose your automation server however you like or not at all - just point the app to an IP in your local network.
- A typical setup is expected to work fully locally and is resilient. In many cases, even if your central server goes down, your "smart" wall switches still work and control the stuff.
- The business model is honest. You submit your needs, get a quote for the installation, and when you decide to go ahead then the company comes, set things up, and it's done. No silly recurrent fees, no bullshit subscriptions. You bought the hardware, it's yours.
So basically, it works as it should. It's not hot and sexy (most of your installation will be hidden in the walls anyway), but it is reliable and it is honest. It's how grownups do business. Contrast that with the startup IoT bullshit.
As some of HNers mentioned already over the last few months, what we need is not Internet of Things, but an Intranet of Things. Exposing all that data to cloud by default is user-hostile and bad engineering. Adding a subscription-based business model on top of that is literally screwing people over. Ownership is like privacy nowadays - too easy to give up for convenience, because the people selling people stuff make money off their mistakes.
Your Energy/Water supplier knows a lot about you already, they use this information to give you better experience or offer services. Sure there isn't the risk that this will brick or stop working but you don't have that privacy in some parts of the world, and in the end it's about making money.
If you buy an IoT product that can't function without the internet, or can't have that smart turned off you you bought a bad product and you should feel bad.
For anyone listening, don't allow others to tell you what to think. The statement above, for example, is in dire cognitive dissonance and directly in conflict with itself. "nobody really knows" vs. "they don't want to tell you".
The fact is, if we keep on keeping on with this stupid VC model thing with startups, it's likely to end badly.
Consumers don't really -care- about that data, for the most part. What benefit do they get? The data by itself largely isn't useful to them, and the extra device 'smarts' often doesn't require internet enablement. Nest doesn't need to send your data anywhere; it just needs to be smart enough to figure out when you're home and what temperature you like. Maybe a (secure, obviously) web panel to allow you to configure it remotely. That is, the benefit of a 'smart' device to the consumer is currently largely orthogonal to the benefit a company gets from it, that massive trove of data. I think, as we move forward, and more and more leaks of data occur, more and more egregious abuses of the data occur (even by the company that sold the device), there will be a second generation of consumer devices, that don't send their data to anyone except you; you gain 95% of the benefits of the current tech, but with a guarantee of end to end encryption.
In consumer spaces, there is, however, market opportunity for sufficiently complex devices. For example, Tesla, the data coming from the car can be useful directly to the consumer (notifying them of parts operating outside of expected parameters, etc), and the model could be changed so that data goes back to Tesla only when the consumer decides to as part of getting that part serviced; Tesla still gets relevant data, but not all of it, and the benefit of some people sending it (much like anonymous usage statistics for software) affects everyone. While over the air updates are cool, they could be done without the user information being shared, and they could be set to be disabled, instead requiring user intervention, for the security conscious.
But both of those miss what is perhaps the most intriguing application of IoT - industry. Where the collection of data from numerous disparate parts and places is beneficial to the customer. That is, a Nest (and its functionality) for a consumer is hardly necessary; they know when they're cold, they know when they're hot, they can change it. But for a multi-national conglomerate? They can run analytics against that data and find places that are losing/gaining heat too fast, indicative of poor insulation, or poor placement of heating/cooling devices. They can be notified of failures in cooling of rooms containing mission critical hardware, and respond quicker. Etc. The data gathering across many devices directly benefits the customer, rather than only really benefitting the company that sells the device.
The good news for IOT builders: standalone remote controls have tended to suck too. On-device non-remote physical buttons also often suck.
Three years ago I tried very hard to go wireless+online for music. I diligently uploaded thousands of .flacs to Google Play when I bought my new Nexus. Yet when I was around the house, or taking the dog for a walk, or going to and from work about half of the time there was silence as it was buffering or whatever. This was in Newcastle, not the third world or rural Scotland. I'd have expected much better, like connectivity and enough buffering to play nearly all the time. I'd have expected good enough even somewhere with spotty data coverage.
Around the same time I bought a nice Denon Network Music player for the home system. Paid a good few pounds in the hope of decent midrange quality. Since when, BBC have reduced stream quality, and Last FM radio (the main reason I bought it) was discontinued.
Internet of S* seems fairly accurate then.
Now all my music is back as lossless on FreeNAS, and I have a LAN only app on the phone as remote control. When out with dog, the old ipod nano is good enough and lasts much longer than the Nexus did. The Denon now ranks as pointless waste of money. It plays my freenas archive, but I could do that already.
So for me, my online music experiment is over, and I'm >£500 lighter.
>for 10 minutes
Whenever we'd come back from vacation, it would take my house at least 3-4 hours to move from ~40 to ~70. Our house was actually pretty small for the neighborhood. Being able to kick the heating back on from a few hours away in the car would have been amazing. I think you have a hilariously overpowered furnace, a tiny space relative to Midwestern suburbia, or an exceptionally temperate climate.