S05E19 of Connection Problem: The best of times, the worst of times
Today’s issue comes to you from Brussels, where I’m writing this in transit. I’m here to discuss digital futures, and so it just seems appropriate that this issue covers a lot of AI-related things. From music to military, from surveillance to digital rights. And while the rest of the world is moving at breakneck speed, it appears that in Berlin’s never-finished airport we now have — wait for it — human smoke detectors on staff. Because the automatic ones won’t work. (Not a joke.) To paraphrase Gibson: The past is already here, but even that is unevenly distributed.
×
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 directly? The best way is to join the Brain Trust membership.
×
Personal-ish stuff
During a major clean out I recently re-discovered an old notebook. It was black (as in, completely black) and the notes were taken in metallic Sharpie. I posted a few pics online (the one in the header, and the ones included below) and of course some folks recognized it right away because they've had some like these themselves: My notes and sketches from a Museum of the Future workshop; it was one of two I’ve participated in over the years, this one from Berlin. I include the notes here for entertainment more than anything else as they only make so much sense as a stand-alone but this group of people is interesting AF, and we need to see more groups convene and tackle the big issues. Black notebooks and metallic Sharpies are not a requirement, but I have to admit: pretty fun. As Scott jokingly points out, “Bilderberg doesn’t skimp on workshop supplies.”
×
AI writing things
Talk to Transformer, a text-generating implementation of OpenAI’s machine learning model crossed my radar multiple times. You give it a prompt, it auto-completes. It’s clearly in an early research state (that bit is indicated in the part I left out in the quote, see the three dots below), but quite interesting:
“Built by Adam King (@AdamDanielKing) as an easier way to play with OpenAI's new machine learning model. In February, OpenAI unveiled a language model called GPT-2 that generates coherent paragraphs of text one word at a time. (…) While GPT-2 was only trained to predict the next word in a text, it surprisingly learned basic competence in some tasks like translating between languages and answering questions. That's without ever being told that it would be evaluated on those tasks.”
My prompts are the bold opening lines; it’s quite interesting to see where the neural networks take it:
Example 1: "It was the best of times, it was the worst of times. You could go see The Walking Dead, or Game of Thrones, or something, and they'd all come into the room with, you know, these amazing actors, but what they were doing was basically, we'd like you to be able to play whatever part that we wanted you to play and it would just happen.
In the second episode, there's a really fun moment about the car dealership. You're trying to work up enthusiasm and you'll be like, I'm looking at these little wheels, and I'm like, they look like they're making a mess. So the car dealers are doing that very similar to the first episode, right, and they're like, it's terrible! Can you do whatever I want with it? Is my character going to be too good for it? You know, it's great to see people come together and try different things, and it doesn't mean that you have to get your ass nailed. And that was one of the things we wanted to do in the first season of season 2, and it really worked."
Example 2: "We should really build IoT things more responsibly. If you want to build more things, why don't you start with more sensors, more sensors in the cloud, and more sensors in your cars, and more sensors around your building, and more things to make sure that when you move, you're not moving too fast."
Example 3: Of course these things do go very, very wrong very quickly. I didn’t get to see the original output Sjef refers to in his tweet but you get the idea:
×
AI killing things (and people)
DARPA is training AI to win dogfights. Fully automated weapons systems should clearly be a no-go. That said, there’s more - this is part of the DARPA section that tries to invent the next generation types of warfare, which of course are hybrid models. Centaur chess, but with guns/drones/AI, you name it.
DARPA’s “mosaic warfare” vision (…) shifts warfighting concepts away from a primary emphasis on highly capable manned systems — with their high costs and lengthy development timelines — to a mix of manned and less-expensive unmanned systems that can be rapidly developed, fielded, and upgraded with the latest technology to address changing threats.
So hybrid models, but also lots of cheaper materials to throw at the enemy. We’re banning plastic straws, they’re inventing single use drones.
And who are they trying to recruit in their research efforts? Exactly: game developers.
“Through the AFWERX trials, we intend to tap the top algorithm developers in the air combat simulation and gaming communities,” Javorsek said. “We want them to help lay the foundational AI elements for dogfights, on which we can build as the program progresses.”
×
AI not scanning your face
In what might be the most-shared piece in my timeline, San Francisco bans facial recognition systems, or at least police use of these systems.
Why is it significant that SF of all places bans it?
Aaron Peskin, the city supervisor who sponsored the bill, said that it sent a particularly strong message to the nation, coming from a city transformed by tech. “I think part of San Francisco being the real and perceived headquarters for all things tech also comes with a responsibility for its local legislators,” Mr. Peskin said. “We have an outsize responsibility to regulate the excesses of technology precisely because they are headquartered here.”
Also, great to see the big picture framing of possible negative impact here:
Matt Cagle, a lawyer with the A.C.L.U. of Northern California, on Tuesday summed up the broad concerns of facial recognition: The technology, he said, “provides government with unprecedented power to track people going about their daily lives. That’s incompatible with a healthy democracy.”
And again picked up here:
Luke Stark, a postdoctoral researcher at Microsoft Research Montreal, described facial surveillance as “the plutonium of artificial intelligence,” arguing that it should be “recognized as anathema to the health of human society, and heavily restricted as a result.”
×
AI tracking polluters
A coalition of groups is turning to satellites and AI technology to help monitor emissions from all the world’s power plants. Live satellite surveillance usually makes me nervous but this might just be an exception.
×
AI being more inclusive
A great list of resources on designing/making/teaching/etc inclusive AI.
×
AI making tech workers leave their job
Everyone’s favorite British tech non-profit Doteveryone’s new report People, Power and Technology: The Tech Workers’ View is an excellent and quick read about how ethical leadership is now required more than ever. They tie it to brain drain and the cost associated with it. Personally I always find the cost argument a bit lame in this context (ethics shouldn’t be a line item) but hey, whatever works.
×
AI making music
Holly Herndon's New, AI-Spawned Album Is Full of Humanity, titles Vice, and it’s really pretty neat and interesting. The interview maybe even more so than the results, but that’s worth some listening, too. Herndon sure knows what she’s talking about, and not just the musical bits but the overall field. Some quotes from the article that got my attention (highlights mine):
(1)
Herndon wrote about PROTO, she noted we needn’t raise technology to be a “monster.” PROTO is a love song to the capacity for machine-made music to be full of life.
(2)
“I have a different relationship to the internet than I do to social media. Those are two different things. The internet is still something with a lot of potential. It depends on the way the architecture of the platform is designed. Right now, most platforms we deal with on a daily basis on the ubiquitous internet are platform capitalist mega-companies like Facebook and Twitter and Google. They function according to advertising logic. They’re not necessarily designed with the best kind of human interaction in mind—there's a different incentive structure.”
(3)
“if you just train a bot on Twitter, you're going to see Twitter reflected back at you. And of course, Twitter contains so much racist garbage. One thing we were trying to do with PROTO was have a very small data set. A small group of people who are aware that they're training, and are acknowledged as people who trained. You can hear their voices, instead of this invisible ocean of people and data where everything gets swept up in this scraping. We were able to control our data sets by having it be more limited.”
There’s a lot going on here, around ethics, best practices, overall awareness - and keep in mind that this is not a thing to roll out as a platform that implicitly defines or shapes the behavior of hundreds of millions of people like many of the others but “just” a music album. If a musician can get these things right, then for crying out loud, the platforms should be able to.
×
Miscellanea
ThingsCon presentations
The presentations from last week’s ThingsCon Salon Berlin are up. A bit thank you to Ester Fritsch, Dr. Isabel Ordoñez and Chris Adams!
Virtual space colonies
No Man’s Sky’s best and brightest are trying to build a perfect colony. Fascinating piece about a subculture of virtual homesteaders inside No Man’s Sky. If you’ve ever tried playing the game, you’ll appreciate that these are people with a lot of patience and grit. Same kind of rhetoric as settlers, too, and they haven’t picked a comfortable place. Even within the virtual they’re going for the hard environments, and even within this weird context someone’s taking to art. Of course!
- “carving out a space for yourself in such an unwelcoming environment makes your homestead feel that much more earned, necessary, and secure.”
- As of right now, there are around 70 players who have claimed land on Mount Loper, spread out across all platforms. Each base is mapped and spread out to give players plenty of real estate to build in, while also giving them an opportunity to link up with other players, if they wish. Much attention has been paid to building extensive road networks between bases, partially for convenience — Mount Loper can be hard to navigate — but also for protection. As an extreme weather planet, Mount Loper sometimes becomes enveloped in whiteout conditions where players can’t see very far. Roads, which are occasionally outfitted with rails and barriers, help players see where they’re headed.
- a recent arrival on Mount Loper, built a base with a gallery exhibit that other players can enjoy
×
Currently reading: How to Do Nothing in the Attention Economy (Jenny Odell), Exhalation (Ted Chiang)
×
If you'd like to work with me in the upcoming months, let's have a chat.
×
I wish you a wonderful end of the week.
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 firm. He co-founded ThingsCon, a non-profit that explores fair, responsible, and human-centric technologies for IoT and beyond. 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. my own, and one Twitter screenshot.