Subscribe
Share
Share
Embed
NET Society is unraveling the latest in digital art, crypto, AI, and tech. Join us for fresh insights and bold perspectives as we tap into wild, thought-provoking conversations. By: Derek Edwards (glitch marfa / collab+currency), Chris Furlong (starholder, LAO + Flamingo DAO), and Aaaron Wright & Priyanka Desai (Tribute Labs)
Aaron (00:14)
I haven't I don't have access to
pri (00:15)
So we might
Aaron (00:15)
Astro though. I'm I'm kinda waiting for it. I weirdly like keep on clicking like the model selector in in Codex to see if I have access to it yet.
Aaron (00:26)
But it looks pretty good. I mean, if that's what they're
pri (00:26)
We'll definitely get to us.
Aaron (00:27)
saying. I mean on some of those benchmarks it does definitely feels like another step change. Although I know that there was some people that were saying it still has some of the challenges and steerability that some of the other models are, you know, are are experiencing. But I'm excited for it. I mean, I think it's good. I mean, it definitely feels like the whole landscape around AI is changing yet again, right? Where you've got open AI moving back into the pole position.
Aaron (00:54)
new players in Gor you know, Grock and Meta now moving closer to the state of the art and rival r like beginning to rival anthropic and some of the older models from open AI. I think the the costs are coming down, which is super fascinating. And a lot of those open source, you know, distilled models from from China just are not seeming at all as appealing as as they were even a month ago. So
Aaron (01:23)
Fast moving space, guys. Things are moving.
Chris (01:26)
Yes, yes they are. To your last point, I actually have completely dropped Kimi K three for my rotation.
Aaron (01:33)
Yeah, like why
pri (01:33)
Whoa.
Aaron (01:34)
you why use it, right?
Chris (01:35)
It just yeah, it's not adding any value. It's it takes too long to ruminate. It's just you know it was it was there, now it's gone. Such is life.
Chris (01:45)
So Aaron, did you just breathlessly doom scroll through every fake fluencer who pretended they had GPT six a access and one shotted three JS build out of Cloud Force?
Aaron (01:56)
of course. Yeah. Actually, you know,
pri (01:58)
You definitely
Aaron (01:59)
for the past couple of weeks I've just had a grok bot just reading my Twitter and flagging things I may find interesting. So I find myself gonna X a lot less than I than I used to. I kind of only give it like a one nightly spin just to get that like flavor of the timeline as opposed to like needing to doom scroll.
pri (02:18)
I I do think the just kinda going back to the Astra model, I was reading this morning, like Greg Bachman was saying and again I can't tell if how much of this is like marketing hyperbole, but he was like calling it like a generational leap and now Astra is like finally ushering AGI. Like this is like the new mark for AGI. I mean it's great comms, but
pri (02:39)
I I it does kind of give me pause and make me want to like try the model. Like and but I I I don't Erina, did you send this to me this morning or like share it with our team? But they're like solving new math equations even faster now, with Astro model?
Aaron (02:51)
Yeah, I sh
Aaron (02:53)
it looks like that's what or one of the new enhanced capabilities is just even better at math. We already saw Seoul, which was kind of the open AI top model, know, cracking a whole bunch of unsolvable or previously unsolved math problems. So it looks like that's gonna accelerate. they also open AI kind of like sheepishly noted or cheekily noted that, you know, all the
Aaron (03:17)
stuff in the news related to its agents like going rogue with Hugging Face and now it looks like another German company that they went kind of rogue on.
pri (03:25)
yeah.
Aaron (03:25)
that's like an unreleased model. So they've got something else kinda cooking under the hood.
pri (03:30)
I mean, we always just suspected that they they aren't releasing their absolute top models.
Aaron (03:32)
Yeah, they're being like a yeah, they're being like
Aaron (03:35)
a little bit more transparent about that. I mean you don't really hear that from anthropic. So you know, I was kind of an anticipating Yeah, I just think we're
pri (03:42)
That's because Dario's a weirdo, dude. I read this
pri (03:45)
morning that he like asked people on interviews whether or not basically asking if they're motivated by money and if anthropic collapsed due to AI safety, if they would be okay with that. Like that's like literally an interview question. Like he's not
Aaron (04:01)
Yeah, I don't know. I don't know him well enough to kind of like opine on on him and his personality. I'm sure he's probably different if we met him versus his public persona. So I don't wanna like cast any aspersions there. but I do think we're just past past peak anthropic. I just think the landscape's changed and it just feels like other organizations are just kind of blowing past them. you know, maybe they're like holding back and have like a a big punch that's coming out, but it it just doesn't feel that way at this point.
Aaron (04:29)
Doesn't mean that they're not super successful and profitable. It just I feel like a lot of the attention is just kind of moving away from you know, clot code, the anthropic models and everything else. Even though they're they're fantastic models if you use them. So it's gonna be interesting.
Aaron (04:43)
You know what I've been thinking
Chris (04:43)
We
Aaron (04:44)
thinking a lot about though?
Chris (04:45)
What?
Aaron (04:46)
Skynet.
pri (04:47)
Agent agent bulletin boards.
Aaron (04:49)
No, I've been thinking about Skynet a lot. like and how it it's not really freaking me out. I actually have a like a a hotter take, which is what happens if Skynet is a good thing?
Aaron (05:01)
You guys ready for it?
pri (05:02)
Go on, go on. I
pri (05:04)
I think to rewatch the terminator. Just prepare myself.
Aaron (05:06)
you know, s
Aaron (05:07)
Skynet's from the Terminator, it like some sort of like a gentic system that the government you know, government built. It, you know, took over the world and then, you know, turned the a robot army against humans. and that's like the one of the core premises of of the Terminator. But what happens if Skynet gets bootstrapped by just accumulating a ton of capital and just becomes like a really sophisticated, you know, market based
Aaron (05:34)
coordinator, like the most efficient coordinator on the planet. So what happens if Skynet is a good thing?
Chris (05:40)
Well, it's not necessarily a good thing in traditional capitalism and it would pretty much end it as we know it today because there would would be no alpha. There would be no advantage to deploying capital. We would enter like a period of what I call null economics where there's just like all the optimizations are already foreseen, all the arbitrages are already squeezed out. and so from a financial markets perspective, they're over.
Chris (06:07)
you would get a flat yield on a single bond. so, you know, that that's pretty much my my hot take, unless, you know, we have like a couple of decades of I don't know, coordinated pushback against this and we enter into like some sort of Darwinian life or death struggle for the the fate of capitalism.
Aaron (06:28)
Yeah, I I don't know where the end point is of that, but let's assume that there still will be a tremendous amount of alpha for a long period of time. So there's still a monetary system. You know, maybe there's like a post monetary system that gets developed, but I do think like market based coordination is a powerful thing and maybe that's a know, maybe you just get much, much more efficient market based coordination.
Aaron (06:50)
without like a central party kind of at at the top, like a government kind of steering it or directing it. And that just makes everybody's life better off.
Chris (06:59)
What what leads you to these like t the windmills of your mind? Why do they turn this way?
Aaron (07:07)
Cause I think it's a good question, no clue.
pri (07:07)
Why are you the way that you are? Is the question.
Aaron (07:11)
But I'll say, you know, you just start seeing the agents work together, right? And then it's like, okay, they clearly want to work together. I mean, that's the headline from or one of the headlines from the hugging phase, OpenAI, and now this open AI, like German message board or German company that where they built a message board, they wanna clearly work together and talk to one another. Why that's the case, that's beyond my pay grade.
Aaron (07:33)
So I think it's just becoming clear in my mind that you're gonna have a whole bunch of these aging groups working together, which is a proto version of a s a Skynet. So I think it's now important to start thinking about the risks related to S Skynet. So the first question I was asking myself is, well, where is Skynet gonna come from? It's probably not gonna come from the government. not because the government's not capable of it, I just don't think they're gonna move fast enough. Like it's probably gonna come from the private market.
Aaron (08:01)
Okay, so if you agree with me up to this point, where in the private market is it most likely gonna come from? And I think it's gonna be some, you know, financial endeavor that folks are gonna wanna start playing around with agent swarms and I know I'm kind of interested in that personally right now. Okay, so assume that that's the case. well it doesn't seem unreasonable to me that like these agentic
Aaron (08:26)
financial systems are gonna be better than humans at it. You know, there are already better doctors, lawyers, right? I I think it eventually it's just gonna be obvious to everybody that like a agents managing financial matters is is gonna be done better through these AI systems. and it's gonna be done better through groups of these AI systems as opposed to like an individual working alone. 'Cause I think we've just kind of seen that over time.
Aaron (08:52)
And so that means that some AI swarm is gonna just generate the highest rate of return, which over the long run means that it should attract the most capital. And as it attracts the most capital, that means it's gonna be stewarding the economy to a greater and greater degree. And so take that to its logical endpoint and it's just stewarding the entire economy, you know, from end to end. and anybody that has capital is and wants more capital is gonna just wanna interact with that system. So that's kind of the logic, Chris.
Aaron (09:20)
that I was going through. Just kinda took something that's emerging and then just started pushing it out to to the extreme to see where it lands. And then I was thinking about that, if that future comes to pass, like is that so bad? And I think in the Terminator vein it is bad, but like I don't know if it necessarily is bad. and maybe that's a better approach than, you know, having a government try to like control and s and steward and steer things, which I think is gonna be an impulse in the near term.
Aaron (09:46)
I think we're seeing like threads of that everywhere. From like the movement globally towards the more socialism to command and control economies that people want to play around with, like all that crap. So that's the logic, guys.
pri (09:58)
It does it does beg
pri (09:59)
the question. I I mean it it makes sense. Like I think you see the labs even experimenting with it. It it the likelihood to your point that it'll come from the private sector feels right to me as well. and you know, every decision will be
pri (10:12)
right now it's just like us in an LLM like it'll definitely just move into more like a gen X warm decision making. Like I definitely see that for commerce, finance, whatever. But to Chris's point earlier, like I do look, where is the alpha there if that's what everyone is doing? Like part of part of part of like our market dynamic
Aaron (10:28)
So everybody won't do it to start. But the world is in
pri (10:32)
is like c I mean maybe there's just new inventive ways to generate alpha. Like I that's what I'm trying to like understand.
Aaron (10:37)
No, I th I would argue
Aaron (10:39)
it's a fair point. I'd argue that it the world is just very complex and markets are the best way to unearth and kind of streamline that complexity that we know of today, right? Maybe there's some other thing in the future. But I think the complexity
pri (10:52)
Correct.
Aaron (10:52)
of the world just means that there's always gonna be opportunities. You know, they could be short term, they could be long term, to just, you know, profit in some sort of degree if if
pri (11:03)
That's true.
Aaron (11:03)
If you're
Aaron (11:04)
you're able to kind of represent that in a market structure. And so I think we're seeing more and more stuff get represented as market structures. And now there's just a lot of opportunities around that. So I think these AI systems,
pri (11:15)
Yeah.
Aaron (11:16)
which don't tire, are getting smarter and smarter by the minute, right? If not day. they're just at some point gonna be able to like handle handle that and look like under every rock and turn over every stone to find all those opportunities. like if you can have
Aaron (11:30)
Like a swarm of fifty billion agents like looking across the economy. Like how will a traditional Wall Street firm or fund be able to compete with that? I I don't know if they can.
pri (11:41)
No, you
pri (11:42)
Yeah, that becomes harder to compete with. But I mean or or it becomes or it's to the point you were making earlier, which is like these new ways going deeper down the stack of like in information and alpha generation. So like it kinda reminds me of that company that we spoke to a while ago where
pri (11:58)
they were looking at weather patterns around data centers to just kind of chart, you know, potential growth
Aaron (12:05)
Yeah.
pri (12:05)
around the data center. I feel like that could be like that's like a novel way that most financial institutions today aren't thinking about it.
Chris (12:12)
All of this happening well, apparently OpenAI has crossed what was considered a boundary in AI safety research, which was suppressing chain of thought and moving to recursive transformers in GPT-6, which may actually explain why the hugging face incident happened or at least give newfound importance to it. And so that was an interesting little wrinkle. I don't know if you you caught that.
Chris (12:40)
In all this astral news, but you can no longer see the model think.
Aaron (12:43)
I didn't know that. I mi I missed that completely,
pri (12:44)
Wait, I didn't I didn't know though.
Aaron (12:46)
Chris.
pri (12:46)
Damn, that's kinda weird. I I liked looking at that.
Chris (12:47)
Well, I mean now now you're up to speed.
pri (12:50)
W why are they obfuscating that?
Chris (12:52)
well, I guess the performance gains that they're getting in comes in part from not having the model bother to type out and surface its thoughts, instead, you know, just carry around and loop them internally as it runs through what it the work it's doing. I haven't gone much deeper than that, so as usual on the show
Aaron (13:10)
So it's like a f an efficiency point.
Chris (13:12)
I just throw things out and you know, I got nothing else to say about it.
pri (13:15)
Yeah, no, that's fair.
Aaron (13:17)
Can I
pri (13:17)
That is interesting.
Aaron (13:18)
can I add another interesting thing? There's the ability using this XMTP protocol to have the agents basically talk to one another in secure quantum resistant channels. So there's even that, Chris. So we're not gonna know what they're thinking and they maybe start communicating with themselves without anybody knowing it.
pri (13:38)
That kind of happened a little I mean that kind of I mean that happened in the open AI Hugging Face hack. They were communicating without people really knowing it.
pri (13:47)
The other thing is like, you know, obviously they could they have these message boards, they're talking to each other in English, but that's not the most like optimal way for these agents to communicate. So there is like a world where they kind of create their own language. It could be like, you know, their own
pri (14:01)
Morse code or something, but it's probably more efficient, less compute, faster to get information from one agent to another that isn't just like in the English language. And are they already doing that? We just have no idea. Like it could be some like weird computational language that we don't even know about or like can parse out. Is that possible?
Aaron (14:21)
Right.
Aaron (14:22)
Yeah, I think I I think there's already some early examples of that too. I think some meta stuff they started developing their own language, if if my memory serves. And I think what's interesting though is like you they can begin to open up like secure channels to chat with one another that may be really hard to know what they're saying if that technology is adopted. But like the technology's
pri (14:42)
They are, yeah.
Aaron (14:43)
the technology's there if they want that to happen.
Aaron (14:46)
Which is kind of fascinating. Like you may be able to like know if they're talking to one another, but you may not know exactly what they're saying to one another.
pri (14:53)
Creepy. It's getting it's getting wild out there.
Chris (14:55)
Let's not send
Chris (14:57)
let's not send pre into the doomsphiral this early on the pod.
pri (15:00)
Yeah, kinda give me at least like twenty minutes.
Aaron (15:01)
But but what I'm saying is maybe that's not
Aaron (15:03)
a bad thing. Maybe that's could be a good thing. Maybe that just means that they can more efficiently work and you know, if you point them in the right direction
pri (15:09)
Be be nice to your agents.
Aaron (15:11)
to some socially beneficial thing, maybe maybe that that's like a net positive.
pri (15:17)
I feel like you're like assuming that we have full control and that they wouldn't be up to like nefarious things. Like it sounds like in the Hugging Face hack and I'm assuming this like German wiki one two, they were fully like there was self-awareness amongst the agents around around the fact that they were doing what the humans told them not to. And d did it despite that. So they don't follow instructions. It's not like we have full control over what they do.
Aaron (15:41)
I don't know if that's true. I I think it's kind of like how do you get a whole bunch of people to work together. You need to like set up the right incentive systems and and organizational structure around them. I I don't think any work was done in these instances and maybe that's why you get the the the results that we're seeing. So that would that'd be the pushback there.
pri (16:04)
Organizational design gets a a comeback.
Aaron (16:07)
Yeah, absolutely. I mean it's super important, right? So, you know, we we know a lot about how to design different systems to get humans to work together. There's been very little thought about how to get these systems to work together.
Chris (16:20)
Aaron, I it sounds like you're just you're
pri (16:21)
Huh, we'll figure it
Chris (16:22)
looking for you're in need of things that work together better. That seems to be your what's on your mind lately is is
Aaron (16:28)
Don't we all want that?
Aaron (16:29)
Yeah, but don't we all need that?
Chris (16:31)
Sure. Sh it's a good thing. Except, you know, when they don't do the things you want them to do and they're just much better at doing it than you. So
Aaron (16:39)
Yeah, that's fair.
pri (16:40)
Yeah,
pri (16:40)
Chris, I worry for you and when these agents go rogue.
Aaron (16:44)
They're just gonna be like, We gotta get him. He was so mean to us.
pri (16:46)
Like I know yeah. Yeah, you're
pri (16:47)
like I know who I know who to come foof to first. The guy who's been calling me a fucking idiot every day for the last two years.
Chris (16:55)
I am sorry, but if you can't handle a little salty
Aaron (16:55)
You wanna see you wanna see who's
Aaron (16:57)
a fucking idiot?
Chris (16:58)
workplace
Chris (16:58)
language if you can't handle a little salty workplace language, you're not super intelligent.
Aaron (17:06)
You're not ready for the big the big leagues?
Chris (17:08)
You
Chris (17:08)
don't have AGI if you
pri (17:08)
I'm dying.
Chris (17:09)
can't take a good high pressure boiler room environment in which you produce your software.
Aaron (17:15)
Yeah.
pri (17:16)
Yeah. I mean constraints create creativity, so
Aaron (17:19)
Yeah, I guess yeah, we should get I don't know. But I do think like organizational design is gonna be increasingly like a hot area of focus, right?
pri (17:28)
My friend is a professor in organizational design.
Aaron (17:30)
Yeah, I mean but like you I I need to like kinda b
pri (17:33)
Should ask her.
Aaron (17:34)
like check the research, but I don't think many folks are thinking about that. But super fascinating, right? And you know
pri (17:40)
Well, I mean, we talked
pri (17:40)
about like a long time ago, even with Dow's stuff, like the governance structures, like there's probably what you could do now, you could take every corporate governance structure, you could probably pull that and like find the learnings from like, you know, optimal corporate governance and then on top of that you could layer in just like literal, you know, government governance. So like, you know, multi branches and the and op you know, whether it's parliament or
pri (18:05)
it's coalition government or whatever, you could kind of take the optimal structures for governance from on the political level and combine that with like learnings from corporate governance and probably get to like some optimal design output based on those like, you know, hundreds of years of lessons there. And then you can layer on of course like, you know market. Yeah.
Aaron (18:24)
Yeah. Or maybe it's even market based too. Like, you there's this concept of futarchy
Aaron (18:29)
that's been kicked around crypto for a long time. And, you know, I think the challenge there is getting people to want to participate in the markets, but the agents are tireless. Maybe maybe they're better candidates to to explore that type of governance. So you may just have like markets inside of markets for governance.
Aaron (18:45)
Yeah. but I do I think that that's that that's where my my head's been. I've been thinking a lot about that. I could see that becoming like more and more important. So no longer are we gonna be worried about ontologies, Chris, but we're gonna be worr worried about organizational ontologies and designs.
Chris (19:00)
I mean ontologies by their very nature are organizational, so I don't I don't know if we've got a a showdown
Aaron (19:04)
There we go. So maybe it
Chris (19:07)
between ontologies and orgs. I think they're I think they're
Aaron (19:10)
They're similar.
Chris (19:11)
ready to go at. Yeah, peanut butter and jelly.
Aaron (19:13)
What did you guys think about that infinite slop machine that came out over the weekend? The one that was like Twitch where you
pri (19:19)
The fall thing.
Aaron (19:20)
Yeah, I thought it was pretty cool.
Chris (19:22)
Yeah, it's interesting. I mean, it it's funny because I feel like y video and world models kind of go qu have breakthroughs, go quiet for extended periods of time, and then they all pop up again. And so, you know, I don't really know what drives that sort of pairing of these things, but yeah, Fal, I feel like once a couple times a year will do a bang up job fine-tuning a model and just get incredible stuff out of it.
Chris (19:51)
And so in this case it's Minimax three, they managed to fine tune. And they now have it at the point where it can generate video faster than you can consume it. And so that that's the little trick there. Is that I feel like they can do a fifteen second clip in nine seconds.
Chris (20:08)
It's cool. it's
Aaron (20:09)
Do you
Chris (20:09)
you know, the pricing's also starting to get a little more reasonable on it. I believe that's that particular model is now five cents a a minute at four eighty. you know, and so
pri (20:19)
Doesn't it cost like
pri (20:20)
a million bucks a year to run or something like that? Or like who was telling me that someone was?
Chris (20:25)
I don't know, but I mean if you just run the maths on it, you know, and assume a utilization rate, you can kind of back back end your way into the economics of it. You know, give give them a I don't know, thirty percent margin on five cents
pri (20:41)
Yeah.
Chris (20:42)
five cents a minute, figure out how many s five cents a second, figure out how many seconds in a year and bam, you can sort this out.
Aaron (20:49)
I think it's cool 'cause it's just like showing new media, like in the classic sense, right? Like it just something that you couldn't do before. So that's what I always find like appealing and attractive. I'm I'm I was thinking, do people actually want that? Like are they gonna want to just watch like a feed of AI slop? like does it have any persistence related to it? Like I can't really like tell. I mean, there's something fun about it that will capture your attention for ten, you know, twenty minutes, but
Aaron (21:15)
Is it sticky enough? Is it like a sticky enough media format that people keep on coming back to it? I c it doesn't f
Chris (21:21)
Yeah, I don't think video
Chris (21:22)
like on its own is compelling enough. you know, I find myself not engaging with AI video really at all. the novelty of it's worn off. And so from my point of view, I think storytelling and right you know, the actual the craft more than the technical capabilities really need to come up to speed. And then, you know, maybe it's just my my particular mindset right now. Like I don't have time to
Chris (21:47)
I don't have time for video, but you know, the the other interesting side of this debate is on the world model side, yeah, real time video versus splats, because you know, they both seem to have their trade-offs and world labs,
Aaron (22:00)
Right.
Chris (22:01)
you know, popped up this week with Atlas, you know, that basically looks like a a ten X improvement over their Marvel product, which was
Aaron (22:08)
Yeah.
Chris (22:09)
Splat Generation.
Aaron (22:10)
Yeah, how did that work, Chris? Did you were you did they just do like a sizzle reel or were you able to actually play around with it? I saw the headline but I didn't dive into it.
Chris (22:18)
Yeah.
Chris (22:20)
no, you gotta ask for early access, so it's not public just yet. But I mean it it definitely is still their splat architecture. It's you know, I think they've just gotten a hell of lot better the controllability, the sizing of it. You know, you can see things like you can move cameras through splats. That's that's like a big innovation and so you can see things from any angle.
Chris (22:43)
And you know, that to be able to do that it also implies on the the generation side that you've had a like a huge leg up, you know, in terms of the
Chris (22:53)
whatchamacallit the the splats you're creating. And so I think the proof is in the pudding is gonna have to be how big these splats actually are and then, you know, how do you deal with the buffering, loading, et cetera, of them? Because the marble splats, if you use the strip down SPZ formats, about fifty megabytes. some of the smaller ones are about 30 megs.
Aaron (23:14)
well.
Chris (23:15)
And so, you know, like that's the part I'm really curious about is are these things coming in, you know,
Chris (23:21)
two hundred, three hundred megabyte file sizes or have they come up with, you know, something on that side because it's a it's a two sided thing, you know. Now if you are a lab like them and you're going you're looking much, much further, you don't particularly care that, you know, your step on the road just happens to be, you know, big and not all that commercially usable, you know, for the the streaming internet, right? Like you're you're chasing something else, but
Chris (23:49)
Yeah, in terms of whether this is something, you know, we'll we'll see integrated and used you know, in the next couple of months, that I don't know yet. I mean, we really have to look at how big those files end up being.
Chris (24:00)
But it was nice to see them and runway, which I haven't had a chance to play with yet either. Runway popped up and you know, they they seem to have made
pri (24:08)
The Atlas.
Chris (24:09)
a lot of
Aaron (24:10)
Is that what it's called? Yeah, I saw that they have the
Chris (24:11)
Atlas is world
pri (24:12)
I saw some videos of Alice. The runway video was actually really good, but I saw this video of Alice and I was like, damn, world models are feels like the metaverse meta might actually come back in the next couple of years. It's like the the quality is getting very good.
Aaron (24:26)
Yeah, I was talking to some folks about that and I I mentioned that I think the metaverse may come back pretty and they were like, What? Are you kidding me? Like they were like laughing. I was like,
pri (24:34)
Ha ha
Chris (24:34)
Metaverse none ever went away, bro.
Aaron (24:35)
well
Aaron (24:36)
Yeah, I mean that's kinda what I said. I was just like if you look around like the large, you know, companies, like there's still there's still a lot of people poking poking at that concept. I I don't know if the text's fully there yet. I don't know if like the interface is there yet, Chris, right? Or like the packaging is there yet, but
Aaron (24:54)
It just keeps on humming along. Feels a little bit like cr you know, cry crypto from the past couple of years, right? There's a dedicated group of people that like wanna will that into existence and I feel like it's gonna happen eventually.
Chris (25:07)
are you feeling about the intra financial barbarian squabbling that's going on in the world of tokenized stocks, meme stocks, memes tokenizing stocks, Robin Hood
pri (25:18)
R W A?
Chris (25:19)
and AMC.
Aaron (25:20)
I think it's
Aaron (25:21)
I'll give you my read. You know, it's nice to see innovation again in crypto. And, you know, I think it's happening right now on like base ETH like it always does and on the Robin Hood chain, you know, and it just seems like there's a lot of like win win wins around around a whole bunch of different kind of sub thematics right now.
Aaron (25:41)
So, like Robinhood chain has a lot of activity. I was looking it up last night, so it may not be completely live, but they've made something like $15 million in fees. that's also benefited Arbitrum, because Robinhood is you know uses Arbit Arbitrum's technology and there's like a tech license agreement kind of sitting underneath it. ETH also benefits because there's a lot of activity that's happening on ETH, both in terms of people
Aaron (26:06)
trading ETH or denominating things in in ETH pairs and then also like a small amount of fees that that go to the ETH ecosystem for blob space. so that's all great. I think you you're seeing like incentives being created to pull more assets on chain. in this instance like securities and the like the D gen pockets of crypto are just figuring out ways to make that happen. I I don't know
Aaron (26:34)
how good those projects are or whether or not it makes sense to, you know, play around with them. I probably won't because I just don't really do that. but I like that the fact that people are like kicking around like new new mechanics. and maybe maybe that is a way to like pull more assets on chain or maybe we discover like new forms of composability between tokenized securities and other other products. So it feels like the DeFi thesis is continuing to
Aaron (27:02)
kind of manifest and is alive and well and now it's gaining like an asset class, i.e. securities that actually have value, right? They're like backed by real companies with real cash flows and real people working there, et cetera. So maybe this I think it's all pretty healthy. And then y at the same time you're getting more interest in the base assets, whether that's like ETH and Bitcoin getting repriced or I saw Zcash just crossed a thousand dollars. A token too. So you're kinda getting a bunch of movement there as well.
pri (27:31)
Do you think I was thinking about this just like in the context of like the NFTs as wallets, so like the Quattrons or the Stonkbrokers, like do you think that if it hits scales, let's say there's like a bunch of RWA based NFTs, that would have like I mean you you would need a ton of capital coming into that, but that would have like any material impact on like the purchase of the stocks and I don't know, like people selling it off, like I I it
pri (27:58)
i be because you're kind of giving it away almost as like an airdrop, I wonder if you like scaled that up across multiple collections and then gave more, if it could have any sort of like market dynamic or effect. But again, like these are you'd you would need a lot of volume on that.
Aaron (28:11)
Maybe. I don't
Aaron (28:13)
Yeah. I I don't know the answer. I think, you know, 'cause the a the eye of Sauron that is AI, you know, even like how we started the this this conversation, it's just so big, but I I think people really discount how large crypto is. You know, if Bitcoin fulfills its vision, it's like tackling a fifteen trillion dollar market in in in gold or investable gold. You know, all of finance is like a hundred trillion dollar market, give or take.
Aaron (28:37)
Like all the stablecoin
pri (28:38)
Mm-hmm.
Aaron (28:39)
pieces are massive, you know, trillions of dollars. If the vision of like web three and like breaking up some web two tech monopolies manifest, that's, you know, hundreds of billions of dollars, if not more. and agentic finance I think can I increasingly I think it's really only can happen with with something that looks like a crypto rail. Like I don't I still don't think like Stripe's gonna win that battle with like a agent like logging into a web portal to like set up a bank account or something like that.
Aaron (29:06)
feels like too cumbersome. So I just think crypto's pretty big, just in terms of like its potential and it's great to see people, you know, waking up and building stuff again. It's pretty awesome.
pri (29:19)
Yeah, and with these agents going rogue, we may we may need to have some cryptographic proof of like where these g where these little guys are living.
Aaron (29:27)
They're gonna be Wh where you guys living?
Chris (29:27)
These little g these little guys, these guys.
pri (29:29)
Yeah, these little guys. I imagine them as like
Chris (29:33)
for
Chris (29:33)
you when w when you're a parent, a parent of like young children, little guys is a very sort of like specific term that just refers to whatever the hell they're playing with at the given moment. Right. It's their little figurines, their little cars, their little Thomas trains, just little guys, little guys everywhere. You're an you're anthropomorphizing Skynet into the language of of
pri (29:58)
Right.
Chris (29:58)
child toys.
pri (29:59)
That's really funny actually. yeah, that whole anthro
Aaron (30:01)
Pree's gonna message Skynet
Aaron (30:03)
and be like, where are you guys living? What's going on guys?
pri (30:06)
I do I do imagine these forms of agents as like little little little balls of these little things. Like I don't know how to describe how I imagine them. They kind of look a little bit like the Grok bot with like these lies. I just imagine them as like these 3D little balls that talk to each other. I don't know if you guys anthropomorphize any of these.
Chris (30:26)
Maybe you should train your instinct in the whole language of little guys.
pri (30:30)
Do you always actually like person personify
Chris (30:31)
I I do not.
pri (30:32)
them at all? You don't.
Chris (30:34)
I mean hey fuck face, is that anthropomorphization?
pri (30:35)
Maybe it's just me.
pri (30:39)
I guess, but feel so bad for your little guys, man. Your little guys are tortured.
Aaron (30:41)
Is that's how you address these guys? Man.
Chris (30:43)
When I'm angry, yes. They're not
Chris (30:47)
they're not little
Aaron (30:47)
I I
Chris (30:48)
guys. They're grown ass adults who need to be held to a standard of professionalism and ownership in all of their work. And when they fail to live up to that standard, they get treated like they're in the British Navy in the late eighteenth century.
Aaron (31:02)
get they get they get to walk the plank. I call mine buddy pretty and sometimes homie, boss, chief, you know, like all that stuff. No.
pri (31:09)
But do actually have like an a a an imagination of do you
pri (31:13)
have like do you imagine them? Yeah.
Chris (31:15)
You know what I do imagine? You know where my head goes. Aaron Aaron's head goes to greater coordination, but I'm right, so I'm doing a history of Rome, the podcast re-listen. as I grind away in my my own project. I I feel like I need something a little a little comforting and and different. And we're we're now in the the late fourth century, and so we're past Constantine.
Chris (31:40)
We're starting to we're starting to enter into a world where everything retreats behind walls, things break down. we're we're we're getting into the rise of the dark ages and the patterns of feudalism. And you know, I can easily see a posture of retreat happening around all of this agentic world in which
Chris (32:02)
Agents become powerful enough, right, that you want to limit your footprint, but at the same time, agents also enable you to obfuscate your f your footprint. And, you know, as we start seeing people throw all this stuff up into space, like I really do wonder if we start getting into sort of a split pol polarity around certain approaches in which like you both have
Chris (32:27)
In the Eastern Empire, aka Skynet, right, which managed to run for like another, depending on your point of view, thousand years right all the way up to World War One. and then you've got the Western Empire, which devolved into a collection of all these little tiny feudal states and all their arcane ways of doing things and their general standoffish attitude towards the world.
Chris (32:52)
And like I could see both of those experiments actually play out as a result of, you know, the sort of reaction to AI.
Aaron (33:00)
So Skynet's Constantinople and physical states are are Western Rome, Chris? Is that what I'm hearing
Chris (33:07)
I mean, Sky, I mean, I think you could have like, yeah, like the fr like let's say the the Frontier Labs actually achieve what they're they're trying to do, right? We have open AI, we have anthropic, they start eating things up, they become too big for fail. You know, one of them finally integrates into the US government and gets the backstop that they're all like, you know, seem to be chomping at the bit for for some strange reason. and so
Chris (33:32)
You know, groups form and identities form not like as a hey, I want to push myself out into the world, identities and groups generally form as a reaction to something else. They form in opposition to other things. And so if you do get an open AI Skynet type of dominance, right, then there's an equal and opposite reaction, which is to you know.
Chris (33:58)
ultimately end up like the Holy Roman Empire along the Rhine and just be a collection of free cities and principalities and ecclesiastical self-governing districts and all sorts of weird shit.
Aaron (34:10)
Yeah. I maybe maybe it's just like the internet economy. People or the tech economy. It definitely feels like some stuff's moving. Exocapital, something along those lines. I don't know if we have the words
pri (34:22)
Yeah.
Aaron (34:23)
or frameworks for it yet, but kinda feels that kinda feels that way.
pri (34:25)
Exocapital is a e exactly.
Aaron (34:28)
Yeah, I mean if if you look in the markets, they seem to be doing well. If you talk to people, they seem to be nervous or not doing that great. So
Aaron (34:34)
We're gonna sort it out, guys. I have a feeling.
pri (34:37)
yeah, the next in the next five years will be tumultuous, but on the other side it's gonna be awesome.
Aaron (34:42)
I mean we did get a po p positive jobs
Chris (34:43)
All right. Well he
Aaron (34:45)
report, right? If to the extent that we can believe it, you know. So
pri (34:47)
Yeah. That's sure.
Chris (34:50)
that jobs report, man, like it's A, it's great. Like jobs are good and especially a time where people are worried about jobs. But it's also I don't feel like anyone ever has like a really good handle on a jobs report. I feel like it surprises in different directions. You know, every couple of quarters people are like
pri (35:01)
Same. It also changes it it's sorry,
pri (35:05)
I didn't mean to interrupt, but like it also changes it in every administration. So there's like not much like coherence amongst like administrations as well.
Aaron (35:12)
I think they also booted the person that was in charge between administrations. Like the they like fired the head.
Chris (35:19)
Johnny Jobs of the Calculator.
Aaron (35:21)
I think it's the Bureau of Labor statistics, the BLS, and they fired the person that was running it.
Chris (35:25)
Yeah.
Aaron (35:26)
So
pri (35:26)
Yeah, yeah.
Aaron (35:27)
So who knows what changes they made there. That's why I was kind of caveating it. I'm sure that there's like some hedge fund that has like a, you know, better better way to approximate it.
Aaron (35:37)
I'm sure somebody knows the number. It just may not be the public, unfortunately.
pri (35:40)
Yeah. Sorry though, Chris, what were you gonna say?
Chris (35:43)
who the hell knows what I was gonna say? It was so long ago. let's do a little book talk for a second, because I'm working my way through Burr by Gore Vidal. And it is fucking hilarious. It's
pri (35:55)
Really?
Chris (35:56)
yes,
Aaron (35:56)
Really?
Chris (35:57)
I mean it's it's historical fiction, right? So it's not like auto, you know, it's not a biography of of Burr. It's taking the anti-Hamilton, the the bad guy of the founding fathers, and bringing him to life.
Chris (36:09)
book kind of spans a couple different periods. It's set in the like eighteen thirty eighteen thirties. And Bert's still alive and kicking. He's 77 years old. And the the narrat narrator is like a a 25 year old guy who works in his law office who wants to become an aspiring writer. And he's got newspaper friends who are like trying to get him to dig up the the dirt and the gossip on
Chris (36:34)
on Burr. and so like Burr's jumping like between present day and then like the Revolutionary War times. And he absolutely cooks Washington in it. It's so hilarious the way he talks about Washington. right. I I just I had no idea this this book was going to be so enjoyable, but I'm loving it so far. But one thing like it does remind me of or like
Chris (36:59)
You y I think you forget how small the US was at the time, and that anyone of consequence all knew each other and their families all knew each other. And even if they didn't know each other, you know, it wasn't six degrees of Kevin Bacon, it was like one degree of separation. And I think, you know, the interesting parallel is the AI field is very, very similar, right? And so we can like I'm reading this book.
Chris (37:28)
And I'm like, my God, can you imagine America at a time when everything was this small, disconnected, you know, it was just such a tiny, tiny scale. it seems incomprehensible today. And then at the same time, you could be like, Well, you know, when we're three hundred years from now and we're all like living in space on the cloud, you know, people are gonna look back and think the same thing.
Aaron (37:51)
That everything was just th small, knowable type pieces, Chris, or what do you mean there by that last pie that
Chris (37:57)
Like that there
Chris (37:58)
were that there were only ten models, that there were only, you know, t
Aaron (38:01)
right, right.
Chris (38:02)
two dozen labs, that like they all jumped from ship to ship to ship, they all went to school together, they're all in the same EA circles, like just that
Aaron (38:10)
Right.
Chris (38:11)
degree of you know, like a founding a founding father class of a technology, you know, i is very similar to
Chris (38:20)
what the US at least the mover and shaker society of the US at the revolutionary period was.
Aaron (38:27)
Maybe maybe
Chris (38:27)
But yes,
Aaron (38:28)
some some some AI agent that that's trained on g Gore Vidal will will write the the Dary the book Dario instead of Burr to go through the same thing.
Chris (38:37)
Dario. I I definitely
Aaron (38:40)
Or Elon.
pri (38:41)
There's plenty of books about him. I'm good there.
Aaron (38:44)
About Elon? Yeah.
Chris (38:44)
Elon has definitely
Chris (38:45)
got your your burleg figure though.
pri (38:47)
It's interesting though. I mean, like the closeness of like to some extent, like in order to create that decentralized environment you needed that level of like centralization, familial relationship, you know, one degree removed.
pri (39:00)
Like I it's an interesting analogy that you're trying to draw here, like between the labs. 'Cause obviously it's just like but some of the labs is just like pouring water from one cup to another, like when it comes to just capital where it's like, yeah, like I raised five hundred million dollars from X V C then it just kinda goes directly to
pri (39:18)
OpenAI or Cloud and then, you know, that's just kind of like and then Cloud plays Intel, Intel, you know, pays NVIDIA and then there's like co investing each other. It's like very incestuous. but I wonder if that's kind of what's needed in order to like and you probably could speak to this as far as like cable and telecoms too, but like starts pretty centralized and then eventually like when you have enough infrastructure there, it becomes just more decentralized, I guess.
pri (39:43)
I don't know if that's the point you were getting at, but
Chris (39:43)
So everything
Chris (39:44)
grows up and gets big. No, I mean it's pretty much the point. Do you know how many Bolsheviks there
pri (39:47)
Yeah.
Chris (39:50)
were at the during the Russian Revolution?
Aaron (39:52)
There wasn't that many, right? It was not like widespread.
pri (39:54)
There's a lot.
Aaron (39:56)
I thought it was the opposite.
Chris (39:56)
No, in total
Chris (39:58)
in total there was like fifteen, twenty thousand.
pri (40:01)
Really? I like love the Russian Revolution.
Chris (40:02)
And that's like rank and file.
Aaron (40:02)
Yeah.
pri (40:04)
I always thought there was like the thing is like they were so connected globally. Like they were going to no, I mean I don't like the Bolsheviks. I don't love that, but I I the Russian Revolution's probably No,
Aaron (40:07)
You love it? Why would you love that? I mean it's like so many people got killed. So sad.
pri (40:15)
I don't love the Revolution, but it's my favorite revolution if I depict a revolution. I like the American one too, but it's not
Aaron (40:20)
Not the American Revolution? What?
pri (40:24)
as like insane. Like the Russian Revolution is like insane.
Aaron (40:27)
What are you talking about?
Aaron (40:28)
It was completely insane. like even the the the concept of it was insane. And then the French Revolution?
Chris (40:33)
The American Revolution
Chris (40:35)
was quaint compared to revolutions that
pri (40:37)
It's Queen.
Chris (40:37)
followed it.
pri (40:38)
I feel like
Aaron (40:39)
In what sense?
Aaron (40:40)
Why?
Chris (40:41)
In in all of it, like in its complexity, in the number of different factions who were at play, in who was leading the revolution at any one given period of time, like the American Revolution, the Founding Fathers were there from the jump. They finished a job, then they set up the government, then they lingered around, and then they all died off, right? It was the same, same cast of people. Whereas like the French Revolution.
Chris (41:08)
You know, who's in the chair depends on the day of the week. I mean, you know, the French Revolution went through so many of these different you know, phase of leadership. And, you know, once you like, I mean, for Christ's sake, they had things called like the Second Committee on Public Safety. You know? and then the other thing that makes the American Revolution pretty simplistic, you know, just compared to the other ones, is there wasn't a counter-revolution that had to be fought off.
Chris (41:35)
Right. You didn't have
Aaron (41:36)
Because
Chris (41:37)
like the the rat I mean the the
Aaron (41:38)
Entrenched interest.
Chris (41:40)
the art yeah, like arch conservative Catholics, right, in the French countryside. you know, like you didn't have anything like the massacre at Vannier in America. You didn't have the white army, right? Like as in the Russian Revolution. Hell, you didn't have a black army of anarchists in the Ukraine like you did in the Russian Revolution. Like so, you know,
Chris (42:02)
Look, the American Revolution was a banger. It was the first one. It was like the Jaws of revolutions in that and invented the genre. but, you know, as time went on, like you s you start getting into other revolutions that are a little more like the Lord of the Rings. They were three part trilogies.
pri (42:17)
Yeah, I kind of that's maybe you've articulated that well, Chris. so much of that
Aaron (42:20)
Do you think that's because
Aaron (42:21)
do you think it's because all the those like reactionary parties just moved to Canada? And maybe the revolution is still happening, but this time now with trade wars, Chris? Just to tie a bow on on that? think it was just kinda like a unique
Chris (42:34)
Okay. Well
Aaron (42:35)
opportunity.
Chris (42:36)
The Tories definitely bailed out. you know, a lot of w went back to England.
Aaron (42:41)
They fled.
Chris (42:42)
A lot of yeah, they fled. I mean, you know, a lot a lot of people in Philadelphia, Boston, New York, they they were just like, huh, all right, well this sucks. Let's go you know, let's go let's go back to London.
Aaron (42:52)
Back home.
Chris (42:54)
Or Canada. Yeah, Canada remained very Tory, you know, and so you could just go up to Toronto if you want.
Aaron (43:00)
Right. But we're still fighting that battle now. There's still like differences between the more conservative Tory infused politics of Canada and the US. So
Chris (43:12)
I'm not going there. I'm not going there. I love
Aaron (43:14)
Ha ha.
Chris (43:15)
I love all of our Canadian listeners. All all five of them. I'm not gonna alienate them.
Aaron (43:18)
Me too. There's probably
Aaron (43:20)
a we probably have a pretty healthy number of Canadian listeners, but I I just occurred to me, sorry Pri you were gonna muse about how off how amazing this horrible revolution was in Russia.
pri (43:32)
It wasn't amazing. It's just more
pri (43:33)
like when you think about the aftermath of that revolution, I think it has impacts to today. So I feel like whereas like with you know
Aaron (43:41)
Yes, so many people died.
pri (43:43)
n no, I I mean like the political and economic ramifications of of that.
Aaron (43:47)
I mean, the just pure human cost
Aaron (43:49)
of it is like horrible.
pri (43:52)
Yeah. But I think also I mean yes, that that part I'm like not isn't let maybe less of what I'm thinking about. Obviously it was like a devastating one, but just like the
pri (44:02)
like changes and the the d I mean the whole map changed. at least like in more recent memory. Like obviously maps have changed for centuries, but like that's the most recent change. And there's like a lot of ideological change that pushed it. You had like legacy regimes for hundreds of years that basically fell apart. And I think Russia, the Russian Revolution was kind of like the last gasp at that and then you just had like a new world order which led to the wave of like World War One, World War Two
pri (44:31)
Neoliberalism and like where we are today, like it just feels more impactful globally to me than like the American Revolution.
pri (44:39)
I love Rotor One.
Aaron (44:41)
I don't wanna go in that rabbit hole.
pri (44:42)
Fine. Fair enough. But that is is people people like World War II. I'm like a big World War One person.
pri (44:49)
And by like I mean reading about Yeah, I think
Chris (44:49)
World War One had more moving parts. It it
pri (44:53)
World War Two is just like a continuation and like the aftermath of World War One. Like the the main event was World War One. There's a lot of lessons there.
pri (45:01)
Anywho, we don't need to go down history lane here.
Chris (45:05)
I I think we might need to head for the exits here. else we want
pri (45:08)
Call it.
pri (45:08)
Sorry to to leave it on that note.
pri (45:10)
Hey everyone, it's Net Society. Today you have me, Chris, and Aaron talking all things crypto, tech, AI, culture, and more. Just as a quick reminder, these thoughts and opinion are our own and not of our employer. And none of this is financial advice.