S05E10 of Connection Problem: Thursday Edition

At reader Peter R's suggestion I'll be trying out a Thursday schedule rather than sending Fridays: Weekends are a busy time, after all, and so we can see if this makes perusing the newsletter easier. Also, thanks for pointing out that my bio had been outdated; it's now up to date once more 🙏

Following up to my notion last week of the imbalance of bot capacities — brands have all the bots, consumer pretty much none — Patrick Tanguay pointed out that exceptions exist in niches, kind of: Remember how Supreme street wear "drops" are scooped up less by humans and more by bots that go in and gobble up their shopping lists' worth of stuff faster than any human can hit refresh and navigate through the site? Again, niches, and run by yet another middle man business of sorts, but more promising. At least in the context of some funky niches like street wear online shops built like a static website from 1998.

×

Know someone who might enjoy this newsletter or benefit from it? I really appreciate any forward or shout-out to tinyletter.com/pbihr. Want to support my independent writing? Join the Brain Trust membership.

×

Personal(ish) updates

OpenDoTT: As I'm hearing from the University of Dundee and Mozilla team, final interviews for the OpenDott PhD program are ongoing pretty much as I'm typing this. Expect official announcements soon, once everything is confirmed. I'm super excited about this; as one of the supervisors I've had a chance to pitch in during the initial round of application screenings, and it's a fantastic group. The only bummer is that there's only five spots available.

ThingsCon: The annual ThingsCon e.V. membership meeting just happened, which is always a fun affair. It's a tad more formal than feels right (there are pretty strict rules to follow to comply with German laws on associations like this) but it's the best group of people. And while we'll want to keep it small, we'll be looking to bring a handful or so more folks into the fold. There's no sales pitch here. It's a thing that serves its own purpose, which is advocating for better IoT. The only perk is that it's a fantastic group of peers with a strong moral compass and a trick or two up their sleeves. If this sounds like it might be for you, hit me up.

×

AI art, amiright?

Over on The Atlantic there's a discussion about what makes AI art special (if anything). AI art is currently going through a lot of firsts: First pieces generated in such-and-such style, first pieces sold at such-and-such auction, etc. It comes with the terrain that this invites debate about what's art, really? The article draws heavy parallels to found art (fwiw, I think rightfully so). I'm not an art expert, but also not a total n00b. The pieces discussed here (generated by an algorithm called "AICAN", more on that below), while novel in their own ways, seem to be pretty much within the frame of reference of 20th century art, and a lot of the "novelty" isn't so much in the origin than in the sales pitch. (For instance, the computer scientist behind the art discussed in this piece, a Dr. Elgammal, coins the term CAN or Creative Adversarial Network, as opposed to the Generative Adversarial Network or GAN from established neural network research. We're working at the level of puns here, people.)
 

Faceless Portrait of a Queen, 2019, Digital Print on Canvas, 48 x 48 in (source)

After all, as the author sums up, "the whole of 20th-century art was predicated on the idea that putting something in a gallery or museum makes it art, rather than the opposite."

One bit I found interesting is that the whole notion of using machine learning (especially neural networks) to generate art is so dependent on large volumes of training data that all too often that training data has a lot of range, a lot of variety. As a result, the newly generated pieces tend to be built like an amalgam of a wide range of styles, putting the whole thing at risk of being… a bit bland, if anything? (For the protocol, I think AICAN produced some visually interesting pieces.)
 

Unity Rising, 2018, Aluminum Dibond, 100 x 100 cm

So I think the jury is still out: What will the truly interesting creative applications of AI be? Will it be iterations on existing styles, which run long enough might be interesting: start an evolutionary style algorithm and let it go for a couple million iterations to see what happens? Will it be introducing new elements, a kind of random noise generator turned up to 11? Maybe algorithms will just produce a giant pool of pieces and the artists become curators, picking and contextualizing. Or will it be something fundamentally new? The stuff discussed here is art, sure, but we're still moving through the motions. Get it out of our systems, because you got to check some things off the list before moving on to the real deal. I think we'll see some much more interesting stuff down the line. This, so far, is not that. That said, if you are aware of stuff in this field that you find interesting, I'd love to hear about it - hit reply, will ya?

Also see:

1. MADE IN MACHINA/E, "an ongoing research into the relationship of Nordic design and Chinese manufacturing culture, machine learning and the authorship roles in design". Read: algorithmically generated "Nordic designs" made in China, based on generating design briefs via Tensorflow. A fascinating project by Simone Rebaudengo and Sami Niemelä.

One example of the results of MADE IN MACHINA/E

2. creative.ai, a toolkit for generative design (currently in alpha).

3. Google AI research blog post about real time AR for self-expression, like a pretty impressive demo of realistic glasses put on a moving person. (Already a feature in YouTube Stories, apparently?)

×

Who's protecting your data?

I was pretty hard on Foursquare last week. On The Verge, Foursquare founder Dens makes the point that they’re going against industry practices by specifically not selling data: That they're consciously taking a business hit for protecting user data. If that’s really the case (and not just a weasel worded excuse along the lines of not-selling-just-renting or something) then an apology is in order. To be perfectly honest, though, I haven’t read the privacy policy and I’m unlikely to. I’ll have to rely on some expert analysis. (If you can offer or point me to some: 🙏) The fact that they’re testing the acceptability of a new live (anonymized) location tracking feature at SXSW however doesn’t inspire confidence in their judgement recently. After all, the SXSW audience — which hugely, disproportionately leans towards online marketers — is decidedly not the crowd to judge data privacy standards by, if anything. So... I genuinely hope I’ll have to/get to apologize eventually?

In very much related news, Elizabeth Warren says she wants to break up Amazon, Google, and Facebook. Which to me seems unlikely and quite heavy handed. But if it went hand in hand with campaign finance reform then things would get truly interesting. Campaign finance reform and heavy handed regulation of US businesses are, of course, two proverbial third rails in American politics… And while we are at Warren's proposal/positioning as the breaker-upper of big platforms, why just go for the big ones, except for obvious populism? It seems like they are, bizarrely enough, the "easy" (or at least obvious) targets. Why not first (or simultaneously) get rid of the data brokers? They're smaller, but just as ubiquitous and damaging, and in my mind far more insidious. Show you mean it by taking them out; we'll all be better off for it. And then we can talk Big Brand Name Targets.

×

Bots, bots, bots

Two items about robots, automation, climate caught my eye this week:

Exhibit A: GACHA, Helsinki's fully autonomous shuttle busses go live.

I've been name-checking these in presentations as the most exciting use case for autonomous driving for years, and am so happy to see them go live. (The article and video are a little awkward; I'm guessing that might be related to a non-native speaker author more than anything.)

Image: Umar Zakir Abdul Hamid, Sensible 4

I'm so excited about this because you can frame autonomous driving as either boon for cargo trucks (utility/industry); luxury individual transport (the Tesla promise, a 1% solution); or as the solution to the last mile of public transport for everyone. The latter builds on existing infrastructure and extends it, maybe even amplifies it. This is an approach for much closer to 100% of the population, and it ties right into existing mobility concepts. It's a humble (if high-tech) approach to a non-sexy, important issue: Infrastructure at its best. (Fun fact: These were developed in partnership with Muji, apparently, in that very Finland way of having a strong minimal design affinity and equally strong bonds to Asia.)

I reached out to the team and they confirmed, public roll-out starts in April: "We start driving in Otaniemi, Espoo, FI in April with open road and open access to all passengers. Later this year in other Finnish cities. Welcome to hop on!" 🙌

Exhibit B: Self-replicating carbon capture robots?

Here’s an interesting story about a scientist who has been working on carbon captures for something like over two decades. At some point he wrote a paper on the feasibility of self replicating robots that would capture carbon out of the air and turn it into rock, to the scale of an area the size of Egypt. But, a snag (and not the obvious one of turning a giant piece of land into rock!):

"They eventually published a paper working out the math and exploring several applications, including self-replicating robots that could capture massive amounts of carbon dioxide and convert it into carbonate rock.

The robot armada, solar arrays, carbon-­converting machines, and piles of rock would all grow exponentially, reaching “continental size in less than a decade,” the paper concluded. Converting 20% of the carbon dioxide in the atmosphere would generate a layer of rock 50 centimeters (20 inches) thick covering a million square kilometers (390,000 square miles)—an area the size of Egypt.

The hitch, of course, is that self-­replicating machines don’t exist. Lackner moved on from that part of the plan, and briefly focused on solar power as a replacement for fossil fuels. But the more he studied the problem, the more he came to believe that renewable sources would struggle to compete with the price, abundance, and energy density of coal, oil, and gasoline."

That’s right: This guy who thought it possible to cover the world in self-replicating robots gave up on solar energy because renewable energy would never be cheap enough to replace fossil. That from a guy who wanted to build robots that created an Egypt sized rock! And here we are, renewables now increasingly the cheapest type of energy in more and more places and contexts. Ah, the times we live in.

×

Miscellanea

Best Practices for Public Good Data Things

Via Laura James' weeknotes I stumbled upon The New Practice of Public Problem Solving (Stanford Social Innovation Review): "A new class of innovators is advancing the public good by figuring out what people actually need and then testing, improving, and scaling solutions that may already be out there." It's a breakdown of approaches and research/UX/dev techniques, with lots of examples. A bit of a long read that I only skimmed, but bookmarked and surely will come back to for reference many times over.

Video game violence

Speaking of best practices, there's a new study looking into the impact of violent video games, and it appears to be a scientifically well done study. (More on why that's noteworthy below.)

Before, I'd like to preface this one with a trigger warning of sorts. There's nothing explicit here, but I know that the impact of violent video games are a bit of a hot topic issue, and one where personal opinions are often strong. So proceed with caution, and an eye on your pulse.

So - as research goes, this area is also a bit of an old chestnut in that it's been research for a long, long time (if not quite as long as, say, the impact of consuming violent content on TV). Now, the issue with this is that the quality of the research over the last 20 years was… sub-prime to say the least. I remember this even from when I was doing my masters degree in communications science (our local flavor of media studies): Even then there were scientists battling it out over these two issues (impact of violent media and games, respectively). And even then it was blatantly obvious that many of these scientists desperately wanted to prove their point, confirm their own opinion.

Now, finally, this may have changed. Here's a study that appears to have a solid methodological approach and large case numbers and most importantly was designed to counter the researchers biases specifically, to not let them cherry pick the results they favored. So here it comes, the result in a nutshell: Violent video games found not to be associated with adolescent aggression.

Book recommendation

As per Robin Sloan's high praise, I'm reading So Many Books (by Gabriel Zaid). This paragraph about the changing meaning of commerce over the centuries from social exchange to today's business-transactional meaning had me do a double take:

If engineers could disassemble a vending machine, go back in time to its remote anthropological origin, and get the basic blueprint, they would find a miracle saying, "Good Morning."

The book is full of little gems like this. And although it projects a somewhat bougie attitude and has a lengthy passage about the how "the new media" aren't skimmable (!) that has aged spectacularly badly, that's easily forgiven. It's a short book, and reading it is more like listening to a conversation over dinner. If you like meta books about books, it's really quite good.

×

If you'd like to work with me in the upcoming months, I have very limited availability, so let's have a chat!

×

I wish you an excellent rest of the week. Enjoy a cup of tea, lean back, reflect!

Yours truly,
Peter

×

Who writes here? Peter Bihr explores the impact of emerging technologies — like Internet of Things (IoT) and artificial intelligence. He is the founder of The Waving Cat, a boutique research, strategy & foresight company. He co-founded ThingsCon, a non-profit that fosters the creation of a responsible Internet of Things. In 2018-19, Peter was a Mozilla Fellow. He tweets at @peterbihr. Interested in working together? Let’s have a chat.

Know someone who might enjoy this newsletter or benefit from it? A shout out to tinyletter.com/pbihr or a forward is appreciated!

×

Pictures: Edward Hopper, Ground Swell, NGA Corcoran Collection (Museum Purchase, William A. Clark Fund). Footer, NYPL Sketchbook Ariyoshi Kondo 1826-1840.