View profile

🛰 Spammy Home, Microgrids, Disengagement numbers – Internet of People Observatory #18

March 21 · Issue #18 · View online
The VUCA Observatory
glad you could make it.
I’ve been travelling last week, and this week tends to be a bit heavy on the conferences, so this issue is a bit briefer than we’re used to. Nonetheless, here’s the Observatory.
In this week’s issue: Spammy Voice Assistants, Microgrids, the the latest Uber stuff.
❤️ Love this? Please share it via Twitter or Facebook.
📬 Got this forwarded from a friend? Sign up here!

Connected Devices
I guess it was only a matter of time. Back when Google first announced their Pixel-exclusive Assistant, which also powers Google’s Echo competitor Home, Marcel and I wondered on our podcast what Google’s business model might be with regards to an AI-first world. Now we’re hearing reports that Google included ads in its “Daily Summary”, ostensibly trial balloons testing the acceptability of this kind of advertising. Google, of course, walked back, claiming it’s just a test of a new feature. Walks like a duck, quacks like a duck…
Between this and Google’s clear problems with their approach to the “one true answer”-problem that powers much of Assistant, it increasingly looks like just porting the experience Google gained with web search on desktop and mobile might not easily be transferable to voice. Experimentation and A/B testing might work on screens, where the difference is not so big, and where you have a lot of visual context to orient around. On audio, devoid of context, it can be disorienting. Additionally, you’ll have to do much more of the heavy lifting of reducing complexity for the user, which of course means taking on more complexity yourself. Designing systems that know when they’ll fail remains hard.
Remember the vibrator maker that was caught tracking how customers use the product to a rather intimate detail, and, crucially, without appropriate disclosure or consent? Well, they’ve been ordered to pay punitive damages to the tune of CAD4m, which should be about 7% of their annual turnover. It’s an encouraging signal, and in line with the latest Vizio settlement south of the border in the US, but it’s important to remember that both cases hinged on disclosure and consent of data collection, not on the collection itself. It’s questionable whether many customers would have read the Terms of Service and Privacy Policy of these products to begin with, had they appropriately existed, and even requirements for prominent placement can be subverted.
Lastly, another big announcement for a smart city rollout, this time in San Diego. I usually put them on follow-up to see what came out of it. This seems to be one for that folder.
As was pointed out, I was amiss in talking about the Brooklyn Blockchain Microgrid initiative last week. I am, of course, inconsolably sorry. So to cut to the chase, here’s the gist of the story: Brooklyn folks set up residential PV (but refuse to say how much capacity they have installed) and instead of selling to the utility, they trade amongst themselves (over the utility’s wires, presumably), put the transactions on the blockchain, and call it a revolution in the making. In simpler times, you’d have set up a cooperative, which let’s members either partake in the profits of selling the generated energy, or use parts of it, which would have reduced the computational overhead of proof-of-work for logging the transactions to the blockchain.
Speaking of actual revolutions, Fortune has an excellent account of the German energy transition, with a good focus on how expensive it can be to kickstart technological development with policy. The complaint here has often been that the subsidies for renewable energy sources tend to be excessive, tail off too slowly, and distort the market heavily. All of which is true, but then, this is how the sausage is made, and if you want cheaper RES, you’ll have to kickstart the initial demand until superlinear scaling effects kick in and drastically lower costs.
The question then remains, however, how to structure a market place that slowly reduces the excessive burden again, and serves as a model for other countries to follow suit. I’m still involved with a local Think Tank to try to figure out how a market design might look like for higher shares of RES. (Of course, there’s discussion of blockchain, too…)
One thing is certain: More RES in the grid will need more storage. So it’s encouraging to see how large battery storage facilities have gotten rather quickly. Tesla just opened a 52 MWh-installation (and attendant solar farm) on the Hawaiian island of Kauai, with a contract to supply energy at a LCOE of $0.13/kWh. That’s quite remarkable.
Let’s keep this brief this week, because we’re all getting a bit tired from the Uber drama, aren’t we? Bloomberg has a nice long read on the Waymo ./. Uber suit. It’s all gotten a bit messy.
What’s more interesting is that Uber’s disengagement numbers leaked. And they don’t look particularly good, especially if compared with what the competition filed with the DMV. Google has about 0.2 disengagements per 1000 miles (~0.12/1000km), BMW ran 1.57/1000mi (0.93/1000km). Uber? According to BuzzFeed, this would come in at roughly 1500 disengagements per 1000mi. That’s in line with Bosch, who are not in it to build full cars, but to test components. It’s a long road ahead, indeed.
The Strange, Weird, and Interesting
End note
As always, if you like the Observatory, don’t forget to share it.
Feedback is always welcome. Hit me at
That’s it for this week. See you next Tuesday!
Did you enjoy this issue?
In order to unsubscribe, click here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Powered by Revue