If you're in that bubble, so most of the innovation ecosystem events, I can, I know 98% of the people there, so I actually don't get any new conversation. Everyone's just going like, this is how we've always been. We'll see. October 28 when we meet with the Banff Center will all digest whether, what, what this actually led to. Yeah, I'm super excited. Mostly, partially because you're organizing it, and you're good people, but also because, um… Yep. Exactly, eh? Yeah. Just like not the end, the beautiful unexpected place like Calgary, eh? Hey Julian, is everything okay, Calgary?

Like everything is beautiful, man. Yeah. Beautiful. We'll hope that is like that nice fall. We are very close to Halloween, so we as kids and those who have kids are conditioned to plan the Halloween costume potentially to go around a snowsuit. So we'll see what weather you get. By the end of the month, but hopefully. I thought you were going to say that there would be plenty of candy corns, but I'll dig out my snooker. Oh, wait. suit. We are going to like go in that direction of like what kind of candy is actually decent because I will fight you on a little candy corn.

You need to stay like five more days so that everyone who has overdone it with the candy then goes to look to like unload it on the adults that shouldn't be eating it. But yeah, the hope is that, so a couple things even for this group is. end of this forum, we're going to be announcing a fellowship program for four to six Calgarians here. And as part of that, we're going to have a six-month program. It's not a full-time fellowship, but we're going to be looking for. Mentors, those sort of things and connections so that they can connect outside of the outside of the city on a big bold future or a topic or an idea so that over the course of that 6 month they're gonna create an artifact or an experience, whatever they shape what that is.

To bring back to the city, so there might be something if people want to be part of that network of mentors, but hopefully. this is a flywheel for so much more, so Julian at the end were asking for his help on the big question is like, what if Calgary was actually a global leader in futures thinking and what would that look like? What evidence if we time traveled, what would that be? And so that will also point us in the direction of things and programs and whatever that we can start to mobilize around, so there could be more programming in addition to what we hope is a recurring. forum, but we didn't want this to just be an event.

It's literally us trying to start building infrastructure in our city around futures thinking. So yeah. Julian's being put to work, so, like, if we I missed you this year, just know that Julian will tell you how much, how many sessions and stuff we've roped him into doing. Much appreciated. Good stuff. I was at MetaConnect, so my apologies, I didn't manage to get, uh, to get the Ban Ray stickers for you. Oh, I, as I was walking around, I was like, I really could, I mean, I might, I'm, maybe I'll, uh, get lunch there sometime. And try to throw them up.

But, uh, yeah, no. Well, the offer is still up. Yeah. Pretty awesome. Yeah. Um…

And it, uh. It was more or less as people kind of saw it, uh, as an event in terms of the The, uh… Yeah, does anyone else have a, have a quick, uh… The different things they had there, they, they were sort of responding to the privacy concerns without responding to the privacy concerns insofar as they sort of are making, uh, glasses without cameras. That was an interesting concession for them to be able to make to say that they were. Conscious of the… the… the… the sort of reputation that they've had. But there wasn't. Two minutes share out before we move to the share out, share outs. there wasn't a lot of description of how it, sort of, constitutes part of somebody's whole life.

They tried to say that, you know, the future is for everyone. I've seen that as an ad for for people, uh, for, for Meta. And that the future that they imagine is a sort of a plushie Yeti. That, um, sort of stares off your shoulder and looks at all of your email, the… The charm. So Meta Connect was this week, they released. a bunch of new glasses, some without cameras. Um, there are these things called MetaVR glasses, which we got to see but didn't get to use. Um, which are the sort of the next generation of Oculus Quest kind of things, and it's going to be $1,300, they said.

They're announcing a kind of a super Tamagotchi with spy cameras. Called the Muse Charm. Um, and they want to have a thing that stays resident on your. On your computer at all times to be able to kind of monitor Okay. And help you, uh, pay off all of your parking tickets. They made the claim up on stage that Muse AI will make you money. And they anticipate that they will be able to just clip the ticket on that, rather than charging you money for it. So that's an exciting and bold claim. We'll see where they get from it too.

I also, uh… Wasn't here last week, but they… that, uh, last week, the, The snap spectacles. I think you might be able to get to their pop-up soon, Julian. They're down in LA. They have a physical store where they'll be showing them off. So definitely worth taking a look at that. But they have… Sort of similarly sort of positioned 3D glasses, but they are glasses rather than there are things, uh, and those will be $2,000 and. Hopefully useful, I don't know. I don't know that anybody's really kind of done a better job of Explaining what it's all for.

But I look forward to some more people trying. But yeah, that was an entertaining thing. The event itself was about as obnoxious as Meta tends to be. It took me about 15 minutes to figure out how to be able to actually exit the campus. And. So that, that kind of left, left me with a slightly bitter taste in my mouth. They left us in the midday sun. I got to see people sort of throwing around Casey Neistat. As we were sort of trying to… Portion out our shaved ice in order to make it through the, um… 90-minute wait in line in 86 degrees heat in the middle of the day. to get into the keynote.

So that was, um, also. Pretty far from the courses I've. I've seen the event in the past years, but yeah. Yeah. That's… that's… Thanks, Brandel. Uh, anyone else? Just briefly? Brandon, you're looking very, uh, John Lennon there with those glasses. Oh, yeah. If John Lennon survived, you might be. We might be twinsies. Yeah, good stuff. Well, I'm super excited. My, uh, my, my friend Alex.

We, we, we caught up a couple few weeks ago, amazing conversation, it's still kind of rolling around my head. There's some things I need to talk to you about. it just kind of pops in my head yesterday, I was like, oh, I wonder if Alex is around he wants to share some of the stuff he's been working on. And so I pinged him, feeling like I might not hear back for a couple days, but he wrote me back straight away. He's like, yeah, totally. Let's talk about this. And so he very generously agreed to do a share. of what he's working on, and I think I told you, Alex, like, generally the way we do it, it's, like, short, sharp, 10-minute, and then the bulk is this wonderful kind of conversation that kind of rolls around, um, sort of inverting the Dynamic, so we get to talk about what was, what was shared rather than

just have it shared, which I know you, I know you're the kind of guy who appreciates that sort of stuff, so… I'm just gonna hand, step up on the dais and hand things over to you. All right. Well, that's good, because I don't… I still haven't figured out how to talk about this stuff all that well. Yeah, yeah. This is a good group. Hello all. My name is Alex Kauffmann. I have I'm future… I think I'm future compatible And so it's like I worked for a decade in China at a innovation consultancy, sort of inventing new shampoos, felt a little bit useless.

My work product was entirely PowerPoint that got ignored So I went to ITP at NYU and I learned enough engineering to be able to actually make stuff. And I made a scratch and sniff TV, which got me recruited to the then nascent Google X Where I worked on face computers very reluctantly for about a year and a half, till Sergey Brin fired me Which I consider… I consider an honor, though financially disastrous. And then I worked at Google Research and a variety of other, sort of, R&D parts of Google for about 10 years. And in 2023, I left And since then have been running a little consultancy called Squeeze Labs, where I and a very, very smart ML engineer who has no interest in making money make things that we think should be made rather than things that will make us money.

So we are nonprofit, not on purpose by circumstance. And so recently we've been working on we've been making sort of very extreme local AI stuff trying to figure out how we can run LLMs and image generation on microcontrollers. We're not quite there yet, the silicon's not quite yet there yet, but it's getting there. And as sort of part of this, we No, it's not… it's not, here, sorry, I'll tell you it's Squeeze lab with no E, because we squeezed it out We made a thing called Crank GPT Which is entirely… the idea was basically, can we run a voice agent, which everybody loves voice interfaces.

Can we run an entire, like, a voice agent, all parts of it, entirely locally, on a Raspberry Pi I built a really sloppy prototype that had about 5 seconds of latency, and I showed it to Kat. And Kat looked at it and said, oh, this is really cool. And half an hour later had built it so that it has less than a second of latency, and you can actually have a conversation So I'm just gonna show it to you. The conversations it has, like, the models we can run on this particular board are not super smart, but you can… it's the kind of thing now where you won't be impressed, but if I'd shown it to you 10 years ago, you would have you would have lost control of your bowels.

So I, I'll just show it to you. I don't know how this is going to work over Zoom. I don't know if it's gonna… if it's gonna kind of I don't know if you'll be able to hear it, so we'll have to see if Zoom sort of recognizes it as voice or just sort of mutes it out. It works a little bit like It works a little bit like a bagpipe. You have to keep the power reservoir full And you don't have to be cranking constantly, but you do pretty much have to… you have to keep that reservoir full, otherwise it just dies.

And so, I'll just… I'll show it. Is there a topic you… anybody would like to discuss with it Right, because my default is post AGI cockroach recipes, but that's not It's not that great at that, but is there anything we want to talk to it about?

Any Andy? I'm pretty sure we should celebrate Sabrina's sweater, so like, let's make it bone mushrooms. Much… about mushrooms. Okay, I think that's okay. We'll probe its mycological knowledge. Here we go All right So, the first thing I have to tell you is that I have to boot Linux and that takes about 20 seconds, so you're going to have to bear with me while I actually Boot Linux Mm-hmm. So, it's like an old car, you have to prime it first, but I primed it before we started talking. It's about a minute of cranking to prime it.

You get… I think it's about I think it's four tokens per turn of a crank. I think that's what we calculated using this model. So anyway And the best part of this whole thing is that because of the way hand generators work based on sort of how much power the board is drawing, the crank actually gets stiffer or less stiff, so you can feel the inference in your… in your cranking arm as you go And so right now, I'm just loading sort of loading Linux, which doesn't use a particular… it doesn't use a lot of power, so it's not getting what do you know about mushrooms?

I'm knowledgeable about mushrooms, including their types, uses, and ecological roles. What's your favorite mushroom I'm more of a mushroom enthusiast. I just ate a mushroom with red dots on it, and I don't feel great. All doubts on mushrooms could indicate fungal contamination. So like I said, not particularly smart. She can't hear me while I'm cranking, so And if I change the subject, it's most likely she will go back to mushrooms Let's see, tell me about ostriches. Ostriches are large birds of prey found in Africa. They mainly eat insects Do they like mushrooms? Both that's just not mushrooms Anyway.

That's pretty much it. And probably she'll talk for about another 30 seconds before before, before she dies. Oh, that's… we're gonna figure it out for you. We'll help you out. Julian, should I show the hardware? Is anybody interested in that? Like, sort of how it works? Is that Yeah, totally. Yeah, definitely. All right, so this is my — there's a website which I'll put in the chat, which is just crankgpt.com that has Some pretty banging copy and a couple of cool generated videos. But this is the this is the demonstration version. And so I will open it up and you can see it has a handy dandy Amazon $30 hand crank And then a little power board, which actually I just I got turned into a PCB, so I don't have to hand solder.

This is just, it's basically a taser. It's 150 farads of capacitance that just holds charge Because the board, which is a Raspberry Pi 5 which is under here is very… it draws a lot of current, and so you can't just power it directly from the hand crank. And that's it. That's all it is. And then a lot of software and a lot of optimization. And I can show you, this is something that we're just working on right now, which is related So this is for my son hates creepy clowns, so we built a creepy clown generator.

This is entirely… this is running on a little $40 board. And we can generate images in 3 seconds. And so this just generates creepy clowns But this is happening in real time. There's no internet connection or anything, so Can you, like, use this to punish him? The creepy clown generator. So, like, my interest isn't Yes. Well, he likes it, he likes it. It's a kind of like, oh, don't show me another one. So I also made a Yeah, smell the spoiled milk. It smells so bad. Smell it again. Well, the thing is, it's a very small screen, but it's high resolution.

So I wanted to do something that was life-size. So I made a belly button generator, which is actually strangely compelling and also really disturbing. But it takes too long to reload, so I'm not going to show it to you. But just take my word for it. I can generate very realistic belly buttons with all kinds of hair or piercings, or yeah, it's, yeah, it's good stuff. This is why we don't make any money. But yeah, that's what we're doing. That's awesome, man. That's so cool to see. As opposed to just kind of read about or infer what it might be doing.

Yeah. So can you can you clarify? So is it… so the model's running locally? Yes. How'd you shove, how'd you squeeze it in there? Is it like a garlic press or what are we talking about here? Yeah. It's actually the one that I'm running right now, the talking one is an off the shelf, I think it's a liquid. There's a company called Liquid that makes relatively small models. I think it's a 1.2 billion parameter model that just fits on the board Yeah. We've done fine-tunes of all of the small ones to be good at talking about specific things, but this is just sort of the general purpose one.

One of the Google Gemma models, out of the box is incredible at translation, and so with just proper prompt, you can get it to translate whatever you say into the language of your choice, as long as it's a sort of a well-represented language. So, French, German, Spanish, Italian work really well. You know, if you ask it to translate into Basque, it just says nonsense But without any sort of fine-tuning or anything. So the models themselves, the point of the project is, one, feel how much energy you're using in your arm every time you

ask an LLM something stupid. And the second thing is also, like, hey, if you know what you want, these things can be… you can run the thing that you need locally It's only if you're doing this, like, general, oh, it knows everything about everything that you need these huge data centers. Yeah. So Yeah, that's weird. I was gonna say, I think there's even like a level there of intent, like, Knowing what you want, there's even a deeper level of… Like forced intentionality when using it that I think is, is a cool angle with this if you're willing to put in the actual, like, physical work.

To get the crank going. You have to think about, like, what are you going to talk to it about, or what are you going to ask it before you start doing that? Because you might be a little distracted while you're actually Cranking it. Looks like Brandon's hand cranking his connection right now. Did I miss the question part or was that a comment? Sorry, I didn't Mm-hmm. All right. I think it all just faded for all of us, so we're not sure. There you go. AT&T is the worst. Was, was good times. See what he comes back with.

There was your… so, so hearing this and seeing this and kind of, you know, um, allowing, you know, all this is kind of, like, rolling around in my head at the moment, and it took me back to a moment, um… When… I don't know if I just felt the moment, or I was working on a project around this time. It might have been like 2006, 2007 when When small electronics kind of wearables were, that was the thing, wearables are going to change everything and there was a raft of projects kind of like both in the art technology world, but also, I think, you know, definitely being kind of contemplated within the commercial world that were activity tracker kinds of things and trying to make activity legible in some other kind of expression was… seemed to be interesting and exciting to people.

So it was all like, Data visualization was more in the commercial sector and then other kinds of representations. And I remember this was also a time when Tamagotchis were kind of catching on. you know, around, and there was something, I don't know, at the time, I was spending some time in Japan, so it's kind of like they were there, and it always seemed like, wouldn't it be cool if your physical activity could be expressed somehow in your Tamagotchi? So if your I don't know. Shitamagotchi got tired or wasn't hanging out with you as much, or seemed a little bit peckish, you would have to go and do a walk or a run or something to kind of reinvigorate them.

And this feels like they're really fascinating kind of opportunities. I don't want… opportunity is the wrong word, but it makes you think about, actually, if you want help with your agentic accounting system, you're going to have to go for a run first. Yeah. Sort of thing. Well, you know, the… the larger sort of context that this was in, and there's a couple of things. There's one which is, I've had so many conversations recently with people who are like, oh, you know, we're throwing all this money at automating screenwriting, and it just turns out that the robots are better at writing screenplays, and they're even better at prompting the screenplay writing robots to write screenplays than any of the screenwriters we've hired, and gosh, it's too bad that all these people in creative industries are going to lose their jobs.

Which, I mean, I haven't seen any of these screenplays, so I can't judge. But the thing that's interesting is it's this… it's the same thing that the Anthropic people are doing. They're like, oh god, hold our hands back, because we're going to invent the end of the world, and we can't stop And I think this idea that you can't stop working on the thing or that you should work, this woman who was talking about these screenwriting robots was like, oh, I feel really bad. I was like, well, then don't work on it. Like, work on… work on the thing that you want rather than the thing that you think you need is inevitable to build.

And so part of this project was this idea, the sort of… if you've seen kids play soccer, how they all crowd around the ball and the rest of the field is empty, it's sort of this, like, well, everybody's crowding around this field of gigantic cloud-based, all-your-data-consuming AI What if you didn't do that? And Is it practical? Is it because part of this feels like a little bit sort of old man get off my lawn kind of like contrarianism. But part of it is also I don't know, like, I watched… I watched the… I watched this sort of whole IoT thing implode Because the premise of, hey, if you add the internet to a thing, you make a better thing.

Like the math was wrong. When you added the internet and compute to a thing, you just got a shitty thing shaped computer But what's exciting about these little models, I think, is that it makes that math right again. You can add contextual awareness and intelligence to a thing and have it still be a thing, and have it not require any sort of external connections or any of the sort of elements that made smart stuff actually much worse than just regular stuff I just don't know it kind of forces you to then interrogate, like, well, what is a smart X and do those smarts actually make it better, or is it interesting?

Because the technical limitations are kind of gone now. And unfortunately, because of RAM prices, this is going to it's going to push this out. The board that I made this crank with initially was 60 bucks and now it's 35 Nothing has changed except brand prices. And so when you had like these $14 These little $14 things that are almost things that you could have put into consumer products, they're now $140. And so, that's going to push it off, but, like, I think that's the more exciting stuff is, you know, we've been talking about stuff that can think for a long time, and finally you have I think you've got the raw ingredients and it's just that work isn't going to happen because of RAM prices and because of all this attention on the sort of bigger cloud stuff.

Sorry. No, that… that's totally valid. Um, I, uh… Oh, did you? Yeah. I really loved seeing this here, and I also saw it at demos and chill a couple months ago, when I think, uh, maybe Kat demonstrated it. Um, one of the things that I really… I'm excited about it is the sort of the lack of. dependence on the sort of the massive infrastructure, not connected to things, but also not even connected to an electrical grid, and short of the, sort of, the prepper-adjacent kind of inclinations that one might have with it, Like. Where do you want to, where do you want to put Uh, these things, what have you thought more about, like The difference is that it turns into, like, the difference that it makes for it to be situated in different environments.

Not necessarily because of apocalypse, but just because, like, Right. It's on a wheel, it's spinning, or if it's in a rock on the middle of a field or something, like… Where do you want to go? Yeah, so one of the things that we haven't built it yet because we're still working on the image models. But one of the things that it seems really — I mean, we're working Our serious work is with toy companies. They're obviously really interested in this, and it's really fun to work for toy companies. When I was at Google, I worked on a Harry Potter magic wand for four years, which was the best project I've ever worked on.

So I was kind of anxious to get back into toys. And this has been a very good way to do so. But so you've seen those, like, now that they've got the nice big color e-ink displays, people have been making frames and doing the sort of the artwork thing. They need to be charged and they're getting stuff over the network So, here's an example of a thing. Picture frame with you give it an aesthetic guideline. You say, hey, I want minimalist geometric art with this color scheme And you can sort of line the frame with solar film, and there's this new solar film that works with indoor light.

It's not as efficient as with outdoor light, but it's much more efficient than current solar film. And so, over the course of a normal day, you can You can get enough charge to run a diffusion process to get a new image and refresh the display. That's an object. That's an object that exists on its own. You don't have to ever plug it in, do anything. It goes anywhere. It has digital elements But it is not… it just doesn't… it never requires plugging in, it never requires a network connection. We wouldn't even put a radio on it.

And so all of a sudden, you have this thing that's, like, completely dynamic. And you could, cool, abstract art. What about if it's like, here's my grandma Okay, I'm going to train this little model of my… on my grandma's face. Show family pictures of grandma doing she's never done before, or what… and so he's like, look, grandma on a ride… a ride-along mower, grandma on a Ferris wheel So you just… there's just sort of a dynamic element to it that I think is really compelling And so does the defense industry. They're really into the sort of targeting without having to ever phone home.

But, yeah, I think… I think, like, if you Especially with larger format things, I mean, we're really focused on really small, really inexpensive, but if you're willing to spend, you know, $1,500 on a frame, you can put a big juicy board in there, and as long as you're running this kind of stuff on the NPU, it doesn't actually use a tremendous amount of power, and so you can keep it as long as you've got some sort of storage in there You can be harvesting power all the time, and then boom, be generating stuff whenever you have power.

And because speed and latency aren't a problem You know, you just… there's just, like, something really exciting about that particular… that sort of, I don't know what I'm gonna get infinite possibility, that when you detach it from needing to be plugged in and needing to have a network connection, suddenly becomes, like, just so much more compelling. Okay, I'm going to sort of follow up on what Randall said because I looked on the homepage, right, and the demos are… there is… I love the project, by the way, so this is not me shitting on that. It's… but it looks… it feels like a manufactured necessity, right?

So it's more of an art project, because I might as well just get up to the wall. Yes. Yes. get the same thing, right? So, or buy batteries or… whatever. So I'm thinking, and because the examples you've given are LLMs, so it's a human facing interface. So it's not like we use it for… Mapping out bird populations or, you know, whatever, because those already exist. We have libraries to map bird songs and whatnot, and see populations, so I'm thinking. What would be, sorry, I'm just repeating basically Bradl question. What would be that is in the… What's a useful use for it?

No, that's the wrong way to think of it because then it then becomes, but what's the niche that this would… field that gives people a new sense of the technology because I. Sure. Like, there are always these kind of projects that show up where it says, this re-evaluates our connection to technology, right? This requires you to sit on an exercise bike, or this requires you to jumping jacks to activate it, or, you know, Right. What Julia said, right? You need to go for a run to counteract whatever other habits you have. I really like this to be able to do this locally and this becomes so physical with the crank.

Yeah. I'm just and I'm annoyed with myself that I can't come up with a good. Use case. Well, so I will say like the crank part of it is very much, it was crank GPT with the pun very much in mind. That's not intended to be the thing you take away. The experience of it shows you how much energy exactly all the stuff requires, and you get to feel How like a long conversation is much harder than a short one. But obviously, the point is, like, very little power can actually produce what 10 years ago would have been considered an incredible amount of intelligence.

And so The idea that you can do some of the things you're talking about, whatever it is that you want to do, as long as you have a scoped project You can actually probably do that with something that never sends data to anywhere else. And that was… that was more the point. I mean, it's… every email you write from now on, every conversation you have is going into somebody's training set. Okay. You can't help it, and that is, I mean, the meta glasses are sort of tip of the iceberg, because granola is taking notes, and, you know, somebody's recording, and it's like, every single thing you do and say is going somewhere.

And it's nice to just say, conceptually, that doesn't have to happen Yeah. And it's kind of a political statement, but I think it's one that more people should be making. you know, I also think Americans are in for a reckoning. They've never lived through somebody going through subscription lists and knocking on their doors at night. Like, that hasn't happened in the US, but that's happened in a lot of places in living memory. My father still doesn't subscribe to magazines because he's worried that the fascists are going to come knocking on the door. And so I think also maybe that something will… I don't know, maybe something will push us over that.

Hopefully, European regulation will make it just more convenient to do it this way. But it's definitely possible. That's the thing. I don't think anybody realizes just it's not that hard We didn't have to do that much work to get this on a Raspberry Pi. To get it on an ESP32, that's a lot of work. But to get it onto a Raspberry Pi, you just have to figure out what is it you want it to talk about, and do a little bit of fine-tuning, and then it just works forever. Yeah, I think, like, the toy angle is really good.

I mean, I could totally see, like, how can I get my transformers to talk and behave in a particular way and add to that some robotics as a product, that's fantastic. I was thinking about the one thing that I've seen more. Okay. I live in Sweden, so one thing that I see in restaurants, and sometimes when I shop is you can see a CO2 equivalent cost of the food that you're buying, right? So that's the attempt being that by being more aware of the environmental costs and energy requirements. Yeah. Yeah. You will make better choices, or at least you will become aware of the consequences of your actions.

So, this project is in the same vein, right? So this much work for this much output, and it's a pedagogical issue. Yeah, no, I love it. I'm just trying to come up with the… Yeah, yeah, yeah. Good use there, yeah. I think it's interesting. Go ahead. Oh, sorry, Sabrina, sorry. Oh, I was… I was saying, I was like, I've been thinking about… context of this forum, there's like this conversation around how we power ourselves and that it's not specific to the generative side of things, but it's dealing with this sort of these tensions around how we want to live, like everything from density to sprawl, autonomy, interdependence, this visible infrastructure hidden systems, things like that.

And so, for me, I was like, toys, great. These sort of things, but, like, this is also a tangible way that I bring to life, like, this sort of, like, how you want to live, or how you might want to live when there are these archetypes. Calgary is a very energy-centric city, and so when people go, can we have a conversation about energy, I'm not going to talk about oil and gas, but talking about sort of how we choose to power ourselves and how we choose to power, how we work, live. I think this is a really interesting piece where it's talking, like, people might want an extreme of, like, these sort of ambient, invisible infrastructure, those sort of things, but we also want to decentralized that we have control over this, and so it illuminates some of those tensions of, like, can you really have ambient, invisible infrastructure?

in a way, or maybe it's just in your household to get to some of that decentralized side of things, but it's just in context of some of these archetypes that we're looking to challenge people on to think from a lens of energy. how you want to power your life or what you think about when you're powering yourself, to your point around you have to crank, you've, you feel and see the energy that you're using in the choices that you're making. I want one for my kids because they just literally asked me whether they can create new class images for their axolotl slash whatever theme that they're doing in class and I go. have to go run outside to earn the energy to do that.

Yeah. Yes. Screen time changes when you have to power the screen, right? Well, I will also say that the conversations with the toy folks have been very illuminating. Because right now this whole moment that we're in is, as somebody who's worked in bleeding-edge R&D for a long time. The conversations that are happening are very much the ones that happen when the thing is still in the lab And unfortunately, this thing, like, got escaped to the lab and got all of the… the money of a thing that's kind of ready for prime time, when really it's still… it's still at the stage where it should be a very plastic, and people should be playing with it and figuring out how it should be directed And, you know, all of the toys so far that have come out, and all of the conversations we've had with the toy manufacturers are basically like, hey, let's put ears on

ChatGPT. That's a toy. And The thing that's more interesting is thinking about, well, what are the things you can do that don't change the interaction with the toy, but make the toy better? So, things like, can you make a toy aware of what other toys are around it? And is that… does that unlock new clay patterns? Can you give the toy… can you have a commentator that's watching you play, and just doing a live running commentary on the game, rather than being the game itself So much of this stuff is about, like, bringing attention to itself, and I think part of what we're trying to do is kind of what you said, Sabrina, is like making the infrastructure invisible, making it sort of something that happens in the background that isn't the point of the interaction, it just sort of brings more richness to the interaction.

And that's the kind of thinking that happens normally when the thing is ready for prime time

And it hasn't been happening, mostly because I think it's still so raw and still so much in development, and the people who are most involved in it are the people generally least qualified to understand how humans relate to other humans or to things. Very interesting concept. The one interesting thing that comes up for me is actually the cranking Because to me, if you flip it The implications to me is that it's introducing a movement ecology If we think about how this might scale locally, right? So I saw a weak signal a couple weeks ago that some school in Australia is teaching Children like toddlers and so forth, interoception Okay. how to be more in touch with their inner sensations and feelings to communicate more So, as you were talking, I was like, oh, what if The feedback from the AI was about how I was actually cranking And inviting me to be more in a

more reflexive space about what's happening in my body, sensation-wise and everything else, as I'm doing it. So what happens if you get a bunch of kids in the room doing that? right? I think it's fascinating what you're doing That's a super interesting idea. So one of the things, one of the sort of first ideas that everybody who's working on computer vision has is like a workout thing that tells you that watches your workout and corrects your form based on some sort of pose analysis But I'd never thought of actually a weight machine That says, hey, that jerk is too… you're doing this too hard or too soft, like, having it actually built into the machine itself is a really interesting idea Oh, yeah.

Sure. Yeah, all kinds of things, right? Like, a card… a card that sort of tells you about your driving style, but not… not… it's like… maybe it knows your driving style over a long period of time and can see that something's wrong and suggests you take a break, or changes the radio Yeah, and to play with that, if I could If I could hit a toggle or something to have it intentionally disrupt my habitual movement patterns so that if I'm cooking in the kitchen, it will call me to come crank in a very particular way, but invite me to walk towards it in a very unusual way, and so forth All for the sake of neuroplasticity?

Yeah, man. Yeah, that's cool. There's ? Well, the powerful thing about this is that anybody can build this stuff. There's nothing magical or difficult about it. You just kind of have to put in a little bit of effort. But yeah, I mean, something like that is totally buildable It's not hard. Hey, Alex, uh, so my name's Peter. I… Yeah I was working with IDP way back in the day with Tom Igoe and Clay Shirky, and the time when open source hardware first came out. Um, Arduino… Yeah. I guess it came out, like, 2005, 2006. I founded a company called Bug Labs in 2006. run the same error.

Oh, we're sorry? Okay, so… Oh, I visited when I was at ITP. I visited Bug Labs. We went on a field trip down Broadway, yeah, to see it. Yeah, yeah, yeah. Okay, so there you go. And uh… one of the… hard lessons, and I'm not trying to throw cold water on this, because you… you had me on LLMs on microprocessors. I think that by itself is worth the, uh, the cost of entry. I think that's a very cool concept. But… so we launched in 2006, We won Best of Show at CES in 2008. But then what happened was, uh, this thing came out.

Yeah. The iPhone came out, and it literally sucked all the oxygen out of the room. Because, you know, IDP at the time, IoT was like, oh, we're gonna do all these cool things, with sensors and Bluetooth and all that sort of thing, and wonderful, wonderful ideas. But because this behemoth came in, with a beautiful design piece of hardware, that everything could work either on or attached to Um, you know, there was… it was game over for us from a hardware standpoint. Arduino managed to get along, because first it was… it was a brilliant technology, but also they were, like, 30 bucks.

Like, you could… You could do things with Arduinos. Um, but I still feel like we're in that shadow where All the investment. So, you've got great ideas as a hardware person, You've got… you want to go attract investment, the investors are looking and say, okay, I can invest my capital here, but then always in my head is Tom Igoe's, you know, hardware is hard. you know, building hardware, scaling hardware, building it, and so on, just takes an enormous amount of capital. Um, and the VCs go, uh, I'm just gonna build this thing that attaches to that other thing, and… And so, that's always been a problem, I feel, for these small hardware companies.

And uh… and so I'm listening to what you're saying, and I'm like, that's great. You can imagine all this stuff going down that path, but… I've… we did this for many years. Not so much with toy companies, but… other folks who are trying to do hardware. And uh… and you get to a certain point, but then getting past that point from a scale standpoint, or a business standpoint, becomes a real challenge. And if you look at even just Meta recently, they've announced seeing what did they invest in? They invested in hardware, But it's hardware that connects to the internet. that does a whole bunch of stuff, right?

So it's like an iPhone on your face. And so, anyway, it's… it's, uh… Okay. I think I'm discussing my PTSD from that whole thing versus anything else, but I do… I would love for something which is not from one of the big guys to start getting traction, you know? Yeah, I mean, I think I think our ambitions end at telling a story, a viable story that's an alternative to the current story. After we put out the Crank website, we got all of these people saying like, hey, can we order one Yeah. Right. And we looked at each other like, no.

Like, we're not getting into that game. And so we put all the plans online, and a couple of people have actually built them, and I just actually just got some PCBs back, so… to make it a little bit easier, so that everything is basically off the shelf But the ambition, especially with the toy companies, like, we're not doing any hardware Yeah. Yeah, yeah, I think that's smart, yep. All we're doing is saying, given that you're using an ST 32 whatever and that you want this kind of play pattern, here's how you get something close to that using these big models.

And here's how you run them on this small hardware rather than saying like, hey, we're reinvent the stuff And I think also, like, the phone thing It's true, but also, I think phone exhaustion might set in at some point. I do think we might see some sort of swing away from like trying to do everything on the phone, especially if you can do little things Yeah. Well, there are limited capacity, too, yeah. situated in objects. The reason that didn't work before is because those objects were more expensive than much, much more expensive than their non-sort of smart versions, and also because they didn't work.

So if we can make them work And I don't think like, I don't think you want to compete with the phone. The phone is a great Swiss Army knife But you don't, you know, sometimes you want to cut a steak, and the Swiss Army knife's not the best for that. I… again That premise sounds great in theory. In practice, it has yet to be borne out, but I do think like showing at least the people who are working on this, and that's really the goal, is like If there's another story, then your thing isn't inevitable.

Yeah, I think that's great. then it's a choice you're making I think that's… I think you're… that's why I was saying that the tagline you have, which is LLMs running on a microprocessor, People who want to get it, get that, and that's the cool part about this, is like, holy crap, that's true, you know? That's… Yeah. to me, I think is really neat. So, cool. I'm still processing. But also, I'm sorry for your loss. Hey, Alex, uh, I love the project, this is super cool and I think about things that think. I worked at an agency doing digital stuff in the mid-'90s and got to visit the MIT Media Lab back then, so I remember the things that Think Project from that tour, which was super cool.

There's a couple patterns here that I love. One is. I love the crank pattern in the sense of, like, An artifact from the past that's almost necessarily obsolete, suddenly coming to life again in this future. You know, by being connected to technology from the future. Just as a general pattern, I think that's really kind of interesting. And then I thought about your conversation with Rodney and the. Like the notion that this crank, you know, kind of makes you stronger, right, and you started talking about things that think, and the gym, and. So one is another inversion pattern of like these digital technologies have this tendency to make us weak.

They make us fat, they make us, you know, it's AI is kind of like making us dumb and not think. So I like this idea of, like, technology that can make you strong. And then you were talking about weights, and the idea popped into my head that like. What if, rather than making weights. Smart, you made something await. So what if you could go to a gym? that was called the Romper Room. And all you did was play games. You didn't lift weights, but you lifted weights because you were playing a game that had you lift weights.

Like, edutainment, physical edutainment. And I don't know, these are just, like, things that that. your project inspired, and this conversation inspired. Yeah, that's the dual purpose object is that's a really interesting idea. Again, I think it's hard to plan those out But I think that's a nice pattern to have in mind, the sort of… I worked with this guy named Thad Starner, who's a professor at Georgia Tech, who's been wearing computers on his face since the 90s, and he always talked about dual-purpose speech So he would say something that was for the benefit of anyone who was listening, but also for his base computing system, and this sort of dual-purpose action feels like that… that feels like a nice kind of paradigm to explore.

I'm gonna… I'll noodle on it. I don't know, that's the first time I've heard somebody Sort of frame it in that way. I'm sure there's some interesting ideas in there. When we get into these conversations, I'm always inspired to think of the lyrics from the John Prine song, Living in the Future. Which, if you don't know it, I'll post them there. They're pretty entertaining. You took me back to, I think it was Paoloca, I can't pronounce his last name. Do you know what I'm talking about? wind up… what the wind-up girl, yeah, yeah, yeah. Wind Up Girl Chronicles, yeah, yeah, the King Springs.

And I remember when I read that, I… I started kind of like… You know, just, you know me well enough to know, it's like, aside from the story of the world, like what are the billboards? And I just imagine these kind of advertisements for different quality kink springs, which was the one conceit, like kink springs, whatever that is. But these devices and so when I, when I, when you took me back to that as I was listening to you. And then Matthias's question about like, okay, so what is this kind of, what's it good for sort of thing, or what's, what's the utility?

I just went there very quickly because there was something, the main kind of like value. you know, artifact thing in the, in the world that that world was calories, you need calories. And so there seems to be these wonderful correlations between all that and the thing that I think that book did really well was… effectively was I really felt like I was in that world. You know, in the reading of it, it was very kind of immersive in that way. And so, why am I saying that? I'm saying that because I think, you know, part of what I know that, you know, you'll be able to do, and I'd love to help, is, like, to kind of construct the set dressing around the world, You know, it's almost like, it's like I wanna see the the the this is in creative direction, I'm just saying it's like, man, if I went to crank GPT, but

all of a sudden I tumbled into that world, and it was available in some kind of way, but maybe it's sort of sold out, And so there are different editions, and one edition is like the, you know, were you seeing someone talk about it that is, you know, this is great under these conditions. And you create the conditions of the world that you sort of imagine this. existing in and maybe some of it is like a little bit of prepper, so you have the guy in the kind of like in the in the duck blind who's like I'm hunting for my family's dinner, but this is a great way to kind of like, um… You know, have asked questions about what kind of buckshot I should be loaded for, for hunting mallards or something.

And they're doing it in just such a matter-of-fact way. Like, this isn't a weird art project, not at all. Like, why would you not have this any more than why would you be the… The buffoon who's walking, who's not doesn't have, you know, luggage with wheels on it. Of course we have this. No question, and maybe the circumstances are implied in the way in which you kind of Oh, you know, have the Crank GPT website that's like. It gives you all the reasons why this should exist in the world. Right. You get what I'm saying, it just… yeah.

Yeah. from there. We have one more presentation from Rodney. So, Matthias, do you have a quick one? I have a really quick one. I was going back to what Rodney was talking about about how we move and so on. If this was a communal thing, right? So you bring three of these together or four of these together, and they have to crank in sync. They have to do a particular tempo and that generates a more complicated pattern or that, you know, Orchestration, you gotta… orchestration. So you have orchestrakish, right? So it becomes a way to enforce communality between Yeah. user somehow like not necessarily the musical instrument, but… Yeah.

But something where you create the connection or a relationship between people that didn't exist there before because they have to match templates. or intensity, whatever, right? For the performative element, I feel like it has to be musicians, it has to be 2, 3, 4, everyone crank. On the website, we have the three versions of Crank GPT. We have the one that I've showed you, then there's the coder one, which uses a foot pedal, and then there's the one for running your startup, which is actually a spinning studio next door to the development studio. So you have this sort of jock nerd, a reconciliation.

But yeah, there's definitely lots of You can take, like, if you actually lean into the kind of self-powering thing, there's all kinds of… there's all kinds of sort of communal things you can imagine But that's Yeah, but the spinning studio, that's just like one Black Mirror episode away from being Lamp and Proletariat, right? But I'm thinking more like if you want to do a communal thing, but yeah, okay. But I mean, that is the thing. You can spin it as a church, you can spin it as whatever you want, but ultimately you are a battery.

Like that is what this comes down to. It's like in this future that we're being sold, we provide energy and maybe a seed and then the machine does everything else. Uh, Seb, really quickly. Hey, Alex, I really, really like your project. I like the way you think. I think it's really valuable to have these sort of objects that help us imagine, you know, the possibilities I really appreciate that it's It's not something that is connected to the internet, and it's not something that's asking me to train it because the technology is not sufficient. That's usually how I feel about using AI.

I'm kind of like frustrated because I have to teach it things And it feels very extractive. It made me wonder if we could have, like, some type of application for kind of like being a tour guide. And you would encounter it in the city square, and you would be like, Hey, there's information about this place here. But if you want it, you have to crank it Right, and it could be fairly simple, I guess, to just train the model on that location's information. It would only give you information about that place. I like the idea of like distributed systems and like not every piece of AI technology needs to be connected to everything, right?

Like, if it's just like a smaller Okay. application right and sort of just thinking of function instead of like bolting the AI And kind of overpowering it almost right. It's like You don't need to bring a nuclear weapon to this sort of crossing the street kind of scenario Right. And and then and then the crank thing is really interesting, because We could experience energy crisis, right? Some places are experiencing blackouts. Right for a couple of days. So it does become kind of interesting to have this companion during a blackout where you're like Well, there's nothing else to do, but I'm going to like crack this thing and have a conversation about ostriches.

X or Y, right? Ostriches, yeah. Anyways, I really appreciate it I'm not sure it's a great conversational companion. But yeah, I mean, yeah, there's and there's some people that are really leaning into the sort of end of days blackout kind of scenario. This Dutch professor took our schematics and has built a thing called survival GPT or something that sort of leans into that idea a little more Let me think about All right, thank you for sharing. Yeah. Hi, Alex, thanks for sharing. It's really, really interesting. Um, like, one of the ideas that came up to my mind, like, immediately, but I was just, you know, listening to heart to everybody else, sharing great ideas.

You should definitely try to put it on this arcade boxing machine, where basically you punch it very hard, and the harder you punch it, like, the longer the joke is, or the harder the more interesting the… A stoic quote you're given, like, by the machine. Okay. Okay. And also, it could be interesting to consider a thread mill that basically roasts you while you're running, and you're the only person who synced headphones with that machine, so you're the only recipient of that roast. So, uh, yeah, those could be really, really interesting things. And, like, the punching one, that could be really, um… fun one, fun way to prompt, basically.

And of course, exhibition design, what Seb mentioned, that would be also really interesting if you could, like, pre-train.

I like the sort of the workout art exhibit Thank you. Nice, uh, appreciate it, Alex. Thank you so much for, for showing up and sharing that. Super fun. Rodney, I don't know what you want to do. Do you want to. Because I don't have much time past 10:30. Do you wanna, do you wanna punt to another week? Or do you You want to go ahead. Okay. Okay, I'll try to be quick. It's a good challenge. I'll, share my screen So first of all, from Rodney, on many things, but I share this not as a plug for myself, but I'm also a self-taught abstract artist It's relevant to what I'm going to talk about.

So just FYI. So back in the early 90s Showing my age, I took one of my paintings and I buried it in the backyard And I dug it up the next day Just to see what would happen. Was it profound and interesting? No, but it was different. It was changed. So back in this May I was thinking about that, because a memory had come up, and I thought Could a model understand that process? Not as a style, not as some sort of filter in Photoshop not as like a predefined look, but actually understand Material impact process over time So given my own play and mostly boredom with using AI for imagery and so forth I began to look at it less as what should something look like In terms of subject style, lighting and mood and so forth.

I was like, what if I started prompting what should happen to it? Process change, emergence, things happening over time Would the models actually get this? So I started experimenting in this way. Started playing with the new ChatGPT that had come out this spring because the imagery was great and so forth And I started to run these experiments And I started to run into noun traps So if I say You know, I want this to be a Polaroid photo of something that's lived in a coastline warehouse for over 20 years. I'd see an ocean I'd see a warehouse and so forth And over time, this continual experimenting actually led me to create use claw code to actually build like little instruments and so forth So again, running into the noun and metaphor trap of actually naming particular objects, it would actually show up in the images And I realized that the shift for me Was to

stop just looking at the AI as the output machine And realizing it's a reader It's reading what I'm giving it in a certain particular way And as opposed to nowadays you hear about, well, we're moving from prompting to directing and so forth Could I direct the interpretation of what I'm actually giving it So, for example, if I gave one of my paintings to Claude or ChatGPT or anything else and say, describe what you see If it's a painting, typically it'll describe it in a painterly type way. brushstrokes, movement, texture and so forth. But if I say, describe this as if this were a graphic design piece Now let's introducing other attributes to a different type of frame, hierarchy, negative spacing, and so forth.

Maybe compositional flow and so forth. So the image isn't changing. But the reading is So I began to realize that what I'm working with imagery in ChatGPT, whatever I provided or AI rather It's encountering a relational field of associations. Texture, color, style, like, all these things in which the models have been trained on And when I switch the frame or add a frame or change the analogy, I'm inviting other associations to it. So I began to experiment with this more so And it dawned on me What if I treated the image as a compositional genome What does the genome do It provides instructions for how something is to be.

So a genome of an oak tree tells the oak tree to be an oak tree. So what does that mean over time? It means as it encounters conditions out in the world, it may get gnarled, it may break, it may splinter and all these things. But it's still going to retain certain invaluable or invariable traits about it. Could I begin to look at images in this way? Right? So here's an example, a quick walkthrough These are two bilateral drawings I did on paper Just with markers and ink and so forth. So I'll show you in a second, but I ended up building a suite of tools That allows me to structure What's the so-called intent, I put that in quotes because that's a trap phrase as well Because then as an abstract artist, I don't know what I'm gonna create.

I just go with it And it helps generate a particular type of prompt that I give to the AI and so forth. So here I take these two drugs And I blend them in a particular way That's a whole other conversation. I have my gripe with how AI is blend. It's still a black box. I don't know what's happening to the compositional structure, I don't understand the material color was involved in that So what you're looking at here This is not peanut butter plus jelly, and bread equals peanut butter and jelly. This is Rodney's dad named Kenneth.

Rodney's mom, named Elizabeth had a son named Rodney That had brothers and sisters as well So this is a descendant Of the merger Happening over generations of time that I can influence the AI to infer So from here, if I can go forward and explore possible descendants I could take this image and go backwards And then for hypothetical ancestors of how this… of the material history of this image may have actually happened So I can take one So for this one And I can also say there's some interesting patterns and motifs here Show me all the patterns and motifs of this image, because I'm going to actually play with one Alright, so I'm going to play with this one.

So from here, I can look at this one and what I call analytical visualization methods, but ways of deconstructing the image in different Patterns, whatever, I'm going to do this type Analytic visualization of this image And then in my creative flow, I do a Shanga yoga, martial arts, whatever. In that moment, I felt the need I need to go do some forward fold, Uttanasata for a five breath count, because I feel like intuitively it's going to open me up more about what I want to do next So what I do is I take that visualization And I blended with the original. the sender, let's say So then I start to say, okay, let me see this out in the world.

Maybe it's a mural. Or maybe I want to see an isometric 3D typology. Maybe it's a structure, maybe it's a museum. And then it hit me. Oh, a playground movement. And let's use that. And notice between the image on the left In the image on the right. Right? Look at what's still in play. The DNA is still persisting So then I take that playground And I use ChatGPT, of course, to create this concept For this generative playground that uses movement that actually replenishes the habitat and the ecosystem around the playground Right? So I was like, oh, this is cool It's like self-subsisting and so forth Right So then, if I were to do some marketing or editorial or what have you I take the analytic visualization I take the image Because I have another tool that allows me to use that analytic visualization As a compositional structure for any editorial piece that I actually create.

And if you look on the image on the right, you can see some of the shapes and movements there. So what this does is that it opens up this idea of what an image actually is. When I introduced this analogy Treat this image as a compositional genome. All the things that come with DNA genetics, whatever, the model starts to infer I can mutate. There's morphology, there's surface, there's lineage, there's all these things happening over time. So what I've done I'd create a suite of tools This is one of the tools called ORCID And what this does, it allows me to blend in particular ways.

I don't write any prompts. The tool writes for me But I'm not telling the tool what to write because I'm not asking the model to show me what I want. I'm working in transforming existing images So if I wanted to do an abstracts blends, you'll see My prompts could be 234 words long. They're procedural. They're not actually descriptive. Right? I may add a little description You know, down here, I can change the medium. Trace, we go backwards. I want to see lineage over time of a particular image. Hypothetical ancestors backwards way of being generative reveal I can all my generative Ways of You know, whether it's using Silhouette, edge abstraction, typographic mapping, I can apply that to images as well.

I can do You know, what I call Bonsai my tool is all the ways that I can shape compositional structures. Right? I can pick one. I could pick, and the images you see here doesn't tell me what the images I'm going to create These are results of the images of my recipes So this way, I'm an old interaction designer, so I'm going to rely on Recognition and recall So Oh, yeah, yeah, yeah. Yeah. Hey, Rodney, I wanna, I wanna provide some space for discussion because I see people, you know, be able to have, uh, you know, we do the Titan and then discussion, and we're getting close to time, so, um… Appreciate that, appreciate you sharing.

Yeah, give me one second just to quickly show that Okay.

I know it feels like you've got one more thing to share, but I don't think you do. I think, I think it's, it's time for feedback, yeah. Okay, I'll stop there. I know, I know, I'm sorry, I'm sorry. I know, I know, I go It always feels that way. If there's one more thing I have to say, if I could only share this, then it would all make sense. I really love the, uh, the sort of reminder of the frames of analysis over the sort of the material having such an important. implication for what what you get out of it.

I think a lot of people have kind of discovered by turns that. Depending on whether you say something is your essay versus something that you've received, you get very different feedback from these agents. And a lot of people get really, kind of, weirded out by the subjectivity of this. The fact that, you know, what you're doing is explicitly subjective, Uh, and also more about plausibility than, you know, actuality. is really great that it's, uh, that it… Um, that it's more about, like, ways of reading and allowing the sort of the. The generation of these plausible candidates as being, sort of, valid for their own sake, rather than, like, oh, ChatGPT told me this fact.

I was like, no, it didn't. Yes, yes, yes. And so, living there, I think, is really, really wonderful to be able to see, kind of, real successes. Thank you. Thank you for that. And it's one of the things that, you know, when I was playing with, I was like, okay, I want, I want to play with this image, but I want you to first interpret my English into Zulu. Then I want you to translate that into Japanese. And then I want you to do a transliteration of them both, tell me what's changed and run and generate off of that Right?

So again, this idea of a reader and this idea where These tools are browser-based. Using the browser as my workspace. And wanted to maintain this feeling of I don't know what I'm looking for. Okay. I want to experiment, I want serendipity, I want to seamlessly where it feels like these tools disappear You know, my creative flow Yeah, I'd give you 10 out of 10 for that. For sure. 10 out of 10. So part of me is like. I get it, and especially when I saw the playground, I was like, 100. It's a perfect way to extend. the value of the work, and also the value of the ideas that came from this and.

I don't know you that well, but, you know, you have an embodied practice. And so there's a certain amount of energy that you're manipulating to get this. In the spirit of abstraction, You're not 100% aware of it until it Gets created, but then in the spirit of leveraging what you're what your creation is, you've taken it, and you've made it. super powerful. The thing that it does, though, for me, It also changes… It changes the way that you interact with perhaps. Yes, yes, yes. Abstraction. Like, it stops, it stops being deliberately abstract, and when I say that, meaning it stops being nothing until it's something.

And it becomes something. Again and again and again. And that starts to codify Yes. This vocabulary… That was without. Uh, a codified language, which just brings you back to words and I just wanted to remind you that, like Yes. Yes Somewhere in there, brother, I think that there's words being used to describe the things that didn't have words. And you're knee-deep in vocabulary. I really But it's not obvious to me yet. Yeah, I really appreciate that, Rick, because I realize Couple things emerged in this four or five month of stealthily and quietly working on this, on these things.

An ontology emerged from here And I realized, rather than using AI to Oh, yeah. describe an outcome of a particular look and feel, you would appreciate as a printmaker, I'm describing the conditions, the event of what needs to happen without knowing What the outcome is, and using analogy as my interface for And what happens over time? That's a whole different ball game, because then this idea of intent and prompting and so forth, it becomes less of a feeling like Okay, here's a list of keywords that I need to throw in it. I was like, well, how do I articulate how I perceive Destroying, and what it could possibly be And it becomes less about the tool These tools aren't meant to be sticky.

It's a car wash. You get in, you get out But As a maker, how is it sharpening my eye? That's the test. Can I persist the aliveness I feel with analog ways of making when I start to step into this domain of working with digital and AI That was the impetus. Yeah, powerful stuff. Thank you. Thank you. What it made me, what it makes me think that you're actually doing is you are examining influence Yes. It's… you're speeding up the generations of influence. So you're like, well, what if I made this drawing, and what happens to it when it goes out into the world?

Or where did I… where did I get this drawing from? And then sort of going back to look for the influences that produced it. And there's something really interesting about about that. Like I said in, like, in the chat, like, there's a Saul LeWitt, but it's like Saul LeWitt, like, two levels of abstraction up. You're not giving instructions, you're giving in… you're, like, putting seeds, and, like, what happens to those seeds? Yes, yes, yes, yes. And that's, like, the playground thing, and the marketing material for the playground. It's like, you've just basically sped up 40 years of people digesting, sort of, visual influence, and you've just, like, compressed it into the thing, which is kind of what the models do, too.

So it's this very nice kind of leaning into this mechanic that already exists Sure. But, like, embracing it and adopt… adopting it, which is… is very compelling Thank you, man. I appreciate it. Okay, so one… One question then is… What do you feel that this offers you that it didn't have before you made these webpages? Great question. What it offers me, because that was the test when I turned the computer off. and I'm out in the world I can now envision potentials or potentialities that I hadn't before So, if I'm walking on the street, waiting for my daughter's bus, and I'm looking at telephone pole And I noticed a particular pattern with the metal and a bent nail or whatever, and I take a photo of it.

Just like when you step into futures, you can imagine certain scenarios already I can start to see different compositional things that I can do. What if this was made out of clay? What if I used a Foot photography, you know, dark room technique to adjust the lighting. It just becomes like this All right, I get to experiment with And not feeling like these disciplines are so siloed based off of the material consequences. I have a photo here. I have a painting here. I have a printmaking here, or ceramics here. Each one has its own conditions conditions and procedures and ways of doing things, but they don't necessarily, in the real world, come into play so seamlessly And thus the mental model is still in this bifurcated type way, but now it's all liquid for me, man.

It's everything is material now But okay, I'm. I can see that it's… It has… real value. for the way you're presenting it. I think about a year ago, What you are describing now, it does sound like you weren't able to do this a year ago, right? You could have stepped into a design library and browsed the clay section and the woodworking section and the painting section. No, no, no I'm just pushing because I'm trying to find out, okay, what's the new thing here, right? The new thing is Being more intentional about experimenting with conditions And material history over time that understanding Like, I knew from 30-some years ago, burying a painting in the backyard, and you're like, oh, that was kind of interesting.

I know if I use acrylic on certain woods and the light hits it a certain way, it's gonna change over time, maybe even the colors fade and so forth. But I see that as like a side effect Of the craft. Not to look at that more intentionally And to think about other associations with images that aren't part of the frame So genome could be one. Another one could be financial instruments could be one. What type of verbs and associations do they bring, and what can I do and play with that aren't part of the space So it's not just the end result, but it's also the interaction grammars and models and patterns that I'm kind of exploding and experimenting with here Sort of, sort of… And it's altering And it's altering my relationship and how I work with AI in a way, because the difference between mid journey can talk to me about what it

just created, and what it thinks I'm doing. So there's this feedback loop In this process. That is new as well. And it helps me infer what the models may be doing. So I've done a lot of research in that regard. And I'm still learning. And all these models are not the same, and I truly believe Claude Hicks, ChatGPT, I think it's built in. But that's just me. That sounds, so it sounds like, oh, sorry. Just a quick one to say that Dave Gray and I were like, hey, maybe we should get our groups together. We could move over into his backyard and continue the conversation and meet some new folks, but it's, uh… Sure.

How would we do that? Okay. send a note saying like, hey, we're over here and the burgers were on the grill if you guys want to come over. There's a link, there's a link in the chat. If I click that link, then this one's going to shut down. Sure, sure, I'll go Yeah. Thanks. Yeah. Yeah, I'll head over. Thanks. Thanks, Rodney. Thanks, Julian. Did you guys want… I don't know, I mean, I love Dave, I've never met his crew, but certainly a vibrant, interesting, creative group, probably. Worth at least saying hello. This is beautiful. You guys going over there?

Yeah, yeah, of course. Thanks, Randall, as always. A way to click. Did you find imitates? There you go. Jeff, I'm not sure who you are, but if you want to go over there, you're more than welcome. Yeah, Sandra, where are we going? Yeah. I know, I know, I know, I probably should. Yeah. Uh, yeah, it's almost Halloween. Oh, it's Sandra. Dude, just put your name in! Freak me out! Um, yeah, are we going to the… what is it, the coffee room or whatever? It's so cool. No, Dave Gray has a community kind of parallel to ours, and he sent a note.

He's a really interesting guy. up in Portland, and he's got a pretty vibrant community as well. Okay, so how do we get there? There's a link in the in the chat I put in. Okay. Let me know if you can see if you don't see it before I leave. Um… Why don't I see? Oh, here it is. You got it? Okay, see over there. Yep, got it