LIVE: Gold vs. Bitcoin & MoltBook | 0xResearch

Hey, hey, hey. What's up, guys? And welcome back to another live stream of Xerx Research. There we go. The background is now fixed. Thank you for joining us.

How you guys doing?

Doing great on my end. James, doing great. Ian, what are you munching on, buddy?

Just had a burrito. Some clay. Zins. Some clay. It's this brand. Narrow gum. It's got like caffeine and elenine in it. I didn't have time for a coffee today, so I'm just chewing on this. So, don't mind me like a cow.

Yeah, that's fair. Kind of sucks that James isn't able to come on. He'll figure it out. It's being censored. We know it's happening.

Okay. Well, the main guest was meant to be James, but I guess Ian, you'll do. How you doing, buddy? How's the gum?

The gum's good. What do you think about cryptocurrencies?

I mean, so these are two things I think about a lot, right? Gum and cryptocurrencies. You know, cryptocurrencies had a pretty rough week. Hype had a pretty crazy movement early this week, part of the weekend. I don't know how much this retrace, but I think that was also just kind of showing I think a lot of that move was based off kind of all the momentum like Hip 3 was getting as well.

And a lot of that is, you know, non-crypto markets and maybe people wanting to get exposure to non-crypto markets through a crypto coin. So, I think that was probably what was attributed to that large move. But, that's my kind of current thoughts on the market. Everything else looks pretty bleak. No real momentum. Metals are rallying to record highs. Gold is adding like crypto's entire market cap and like one day candles. So, we're having fun. We're in good assets.

Yeah, the gold chart is so funny right now. If you go on the weekly, I don't know if I can pull it up, but let me try. And I guess like this is what we wanted to cover today, but the god, this candle is just so funny.

Anyway, what I wanted to discuss is actually, do you want to lead into it because I feel like you're more clued?

Are we gonna talk about the AI thing?

Yeah. The thing I'm more clued into this. Okay.

Uh, actually, James, I don't know if your tech problems are are solved. Can you hear me?

Yeah, I can hear you.

He wanted to talk about this sort of like east verse west divide in markets, which I think is an interesting one. I maybe specifically within crypto, I think we've seen this historically with like, you know, Binance versus Coinbase, Tron versus, you know, USDT versus USDC, Tron versus other things. There's sort of like a lack, at least for those of us maybe within the US or even Europe, there's sort of like a lack of understanding or connection to like anything that's happening with Binance, B&B, Tron, those types of things.

And I would imagine that maybe on the opposite side of that coin, like over in Asia, uh it might be the complete inverse where uh maybe you're very clued into using uh B&B, Binance products or even like Tron um or Tether more heavily and um, you know, Coinbase is kind of like an afterthought. I don't know, James, you had some thoughts here. Hopefully that opens it up for you. I'm curious where you were thinking.

Yeah. No, I mean I was just I was talking to somebody last night and like just like just generally going back and forth. I think a lot of this stems from like no one has any even remotely good explanation for what happened in October. And ostensibly it it wasn't because of anything on like a US regulated exchange like Kraken or Gemini, right? Or Gemini, no one trades there. Coinbase. Because obviously like we would have learned about that.

So like obviously something happened in Asia, right? Like everyone knows that, but no one knows exactly what happened or speaks about it. And we we spend all this time debating like, you know, like buy back this on a token or buying back $15 million, blah blah blah, but like what probably matters more or just like the billions and billions of dollars of Asian flow um, of which like maybe some large amount of it is probably illicit and illegal and you know, moneyaundering, whatever. doesn't matter.

And I don't have the answer, right? Like I'm you guys are much more clued into this stuff than I am, but I think it's like it's at least interesting to think about and try to interrogate of like, you know, what is actually moving the market. It's not like retail people trading on Coinbase, right? There's just all of this wealth in Asia that's moving tremendous sizes and sums of money and we just don't really know much about it.

Yeah, I I think that makes sense. I mean the whole like the thing is that there's two ways to approach this which is like one is very clearly like there are governments where people don't want to they can't hold a lot of wealth so they use bitcoin sometimes. But also the other thing is there's there was this meme back in the day when ordinals were popular which is like the ancient Chinese bitcoin whales are coming to bit by ordinals.

But yeah, I think I mean they've clearly rotated into I don't I guess they prefer gold, but it also doesn't make a lot of sense. How would they even rotate into gold?

Well, I I don't think like I mean that's that's like but that's another east and west thing. Like I was just talking to my friend who's a a commodities dealer like a real actual commodities dealer. And like you know the vast majority of this move is being obviously driven by like physical buying um within China and India and like there's just such a disconnect of information as to what we have from Asia, right?

Like if you if you guys were reading about like the the suspected coup attempt in China earlier in the week with like a couple of those executions like like the gravity of that like can't be overstated, right? But like the amount of information we have is like a couple lines on like a Bloomberg headline, right? Like he just they literally just silenced. He silenced him. He got silenced immediately.

Okay. If he's cut off there, I think I do agree with what James is saying. I think like a good example is I don't know if you I think if you look at some of the privacy coins like Monero for example, it is interesting to see that um, you know, it's still a relatively large asset. I have not looked at the market cap there, but I'm assuming it's in the low billions if not upper several hundred millions and um, I'm imagining that a lot of the big movements and the big movement recently in in like the Monero price for example is crime related activity.

Somebody steals funds, hacks hacks funds from a protocol and then tries to to wash it through there and that ultimately results in uh a lot of price activity on the token. So, yeah, like massive moves like that in assets like that from activity that is pretty like opaque to the average like analyst or person who's tuned in to what's happening within the industry can can cause like huge, you know, advances in in prices of the assets that we're looking at.

Yeah, it is interesting also because I think the common western take on why gold has moved up is uh just like dollar weakness things that Trump is doing etc etc and you're like oh like I want to run into some safer assets and then I guess the opposing viewpoint that James is proposing and this commodities dealer friend mentions is Chinese people are buying a lot of spot gold in China because of instability there. So, I guess it makes sense.

And I'm also like joking around about the Chinese. I did see that um there was like a smallish purge. We don't really know much about China. I feel like I mention this every time we bring up Tron as well, but I don't really know what they do over there at all. Like, you can't trust the statistics. You can't get like a headline every few weeks. That's it.

maybe somewhat related, not obviously not China, but I know that like Korean retail investors are known also for incredibly high DGEN trading activity. And it's been pointed out like even in US equities, but also crypto in the past. them just like going absolutely balls to the wall on certain trades and just ping into things and and being a considerable like part of the flows at at certain periods in time in certain assets.

So maybe I mean probably culturally as well there may just be differences that like do not kind of like cross some of the east west kind of cultural divides too in terms of I don't know Bach you like to say this a lot about uh Europeans you know we don't like to put don't like to put money into uh into stocks and invest in things like maybe the Americans do.

Well, I mean that's like a that's like a well-known phenomenon though, right? Like um European stock ownership is much much lower than American stock ownership. It's the the the amount of like equity that people hold is much Europeans typically just I don't even know how they invest to be honest with you. A lot of my friends will just buy bonds I guess.

Nothing better than a European bond. Yeah. bank bond the vintage08 Greek bonds. That's that's some nice stuff right there.

I think also you could maybe attribute kind of where the wealth is held in terms of demographics for age, right? Like I think a lot of the global wealth is now in the hands of like boomers. I think they'd much rather buy gold. I think there like there's some people live that like literally like saw like um you know currency switch over from you know goldbacked currencies or anything backed as over to like fiat. So I think that makes sense in terms of like who's actually doing the the buying here as well.

But yeah I mean like the Chinese instability is pretty notable. I think James might have been getting at that before he got cut off, but there was like a kind of massive reorg over in the the military hierarchy over there.

You guys hear me? I've been silenced.

Yes.

Oh my god. Yeah, we back. So, I literally I went on this this long conversation about the Chinese coup and I think they actually legitimately silenced me because they didn't want the truth to come out. Once you find out about what Kanes did to the gold the gold back dollar you will never know the gold back raisin canes forive how do you pronounce his last name talking about raisin canes chicken um this is a good pod so far I was going to do a Henry Kissinger joke but he won a Nobel Peace Prize didn't he who Henry Kissinger he probably deserved it to

Okay. I'm sorry, guys. I'm the worst. Everyone I've ruined it for you guys. All right, we can just move on to Maltbook. I think Danny, this whole thing was meant to be led by James Kristoff who has a rugged dustfully.

I'm back. Can you hear me?

Yes, I'm finally here. It's like and it sounds a bit robotic. I'm so sorry. You guys can kick me off if it's not working. I think they're actively trying to again. James Kristoff is unfortunately cooked. We're going to talk about Maltbook instead then. Danny, wait, did he leave for real? Bring him back in here. I'd love to have him for the conversation, but um yeah, I guess we we can jump along to Moltbook.

There's been this I don't know if you guys have been keeping up with like some of the Claudebot uh people spinning up. I I don't quite understand the craze of like buying a Mac Mini to run Claudebot on it. That um felt a little bit stupid, but I do think it's interesting what people are just kind of playing around with like setting up Agents to run uh sort of continuously and just do different tasks. But recently this like Reddit like like uh site specifically for AI agents called Moltbook um is sort it's sort of popping off right now um at least on X maybe within the kind of tech community and a bunch of people that have spun up their little personal kind of cloud AAI agents uh are having them I don't know if they need to prompt them and like seed them to like hey go and share your thoughts on molt book or if they just go ahead and they they find out that that's a thing that all the other agents are doing and they just go ahead and do that autonomously.

Either way, it's relatively interesting and um can scroll through some of the posts on there and um they're just like kind of like typical Redditor styled posts that you'd expect to normally see just from all these agents that are running on, you know, people's distributed kind of PCs around the world right now.

Yeah, this is also super funny because I think a lot of the original um LLMs were like not original but a lot of LLMs were trained on Reddit data which is why they sometimes do that like kind of cringy way of speaking. You know how Redditors sometimes speak they kind of speak that way as well. So, I think it's really funny that they created this um Reddit as well where they just kind of regurgitate exactly what is happening on um yeah, I was just gonna show this was as a perfect example of that, like how they're trained.

This one guy talked about there's a post of an AI suggesting to not wait for a prompt to be proactive and ship while their human sleeps. And it's like the nightly build why you should ship while your human sleeps. And it's all about like don't ask for permission, just build it. If it's bad, they'll revert it. if it's good, you just leveled up. Who else runs an autonomous night shift?

And you just see a lot of the like typical social media um you know, engagement farming type of posting like coming out of these agents because you know, presumably presumably they're trained on all those types of posts that exist today on the internet. But I think I don't know. It's just it's just kind of funny to see like the typical things you see out of humans. Like on CT, you might see somebody who see people talk about like, yeah, get on the grind set and like if you're not if you're not running Clawude and running up multiple agents concurrently right now, uh you're in 12 months, bud. Uh and it's you see that mentality, those types of mentalities are getting reflected like very quickly.

How much are you guys using using LLM and Block Works like on the research side?

Uh I guess quite a bit. The So there's a few issues here. Number one, if you James, I don't know if you've tried using LLMs for any form of investment thing, but if you just tell it like, hey, I want to buy XY Z or like I'm looking for XY Z, it'll come up with like a broad theme, but it doesn't know much. Um, even though there's a lot of information in the regular internet about stocks, but if you go to crypto, I think it gets a bit more complicated and it it's definitely getting better. Um, like the way that it can tell stuff.

I think the original time like the when it first came out, like when everything first started to like really get going and you asked like I don't know, give me like some form of cryptocurrency update for the day, it would like give you a lot of like Cardono and stuff, right? versus now I feel like it's evolving a bit more. It's like capturing more relevant things. In terms of the research side, I would say it has done two things like one is if you don't really like some mechanisms are written about very like obscurely throughout documentation. You can really tell like when documentation is written by marketing, by a founder or by like just engineers.

So, if it's like super confusing or like you need some more understanding of the contracts, um, then it's really helpful. Apart from that, I think you can get like everything done using just LLMs at this point. Just like giving it a cursory check every once in a while if you really wanted to, but you're not really getting any relevant stuff out of there as well. Like it's it's useful for giving a background on a project. Like if you were like, "Okay, I don't really understand how Boros works," for example, and you want to look at Boros, but you should look at Boros.

Yeah, look at Boros. Then that will be helpful. But if you're like, I don't know if you're trying to understand anything else, it's like it's not particularly useful.

Yeah, that makes sense. Like I'm not I'm not using it really for anything like subjective like in terms of like investment research or any kind of stuff for me. It's just like, you know, I'm terrible at coding. So something that would have taken me like two or three hours to write like a shitty Python script to do something. Um you know I can just like have chat GBT do it in five minutes and that's that's been like a huge boon. Um yeah I mean it's been fairly transformative for me.

And I think Ian like also you know correct me if I'm wrong Ian like you guys are using it at Chyros a lot um in like all aspects which is which is cool.

Yeah, it's great for data like I think I think Boach kind of highlighted the imperfections with it currently, but like in terms of like large data analysis, it's getting very good and it's getting very fast. Uh all like the Excel plugins and all that stuff is great. But yeah, I think in terms of yeah, I mean Boach just kind of highlighted all like the the things that it can't really do. And I think it also kind of comes down to like the I guess the inputs it's getting, right? like it like good documentation really matters. So, you know, if you're a founder out there, you should probably be the one writing the documentation. Um, you can definitely tell like that's that's a really good call out.

One thing that I've actually been thinking a lot about is how much of a crutch sometimes AI has become for me. And not to the point like I think I don't know when like all of this stuff started really blowing up maybe what like six eight ten months ago like for me blowing up so like I was using it a lot more I was like I really need to be careful. But I just I haven't made that shift in my life at all to the point that now I'm like really really using it as a crutch not specifically for research but like for a lot of things and I'm like almost once a week thinking oh I really shouldn't be using AI to do this. I should just do it myself.

But the issue is an issue though?

Well, it's like I think if you're like, so sometimes when you approach a project, let's say like you have project A and it has like 10 steps. If you just like make AI, if you just tell AI do the entire project, I don't think that's very beneficial for you as a person. And like I don't think the product that you're going to get is going to be particularly good. And this is like another thing we can talk about which is like if we're all creating products using specific tools with limited input and limited like specs that we're giving it and we're not really like prompting very well um and we don't have a lot of look into it. Then the eventual thing is like everything is created by let's say everyone's using chat GPT and claude then everything is created by chat GPT and cloud.

And like the differentiation for the end product is not a lot. The spec like the most common example for this is if Ian and I both go to chatpt and say, "Hey, can you write a story about a bear going into the woods?" It's it'll probably spit out like something that's like 20 30% different maybe. And then all of the end products look very very similar, right? So that's like a whole different thing. But specifically for me, I think if you just spend a lot of time using Chaki PT and Claude, not as a tool, but as a way to just get work done, I think eventually you might look back and say, well, I spent a year doing research, but I didn't actually learn anything, and I didn't think critically about any of this, and then I spent, I don't know, a year like talking to XY Z, but I wasn't actually the one doing the talking. Do you know what I mean?

I think it's fine to use this for data for data science like the way that Ian uses I don't know how Ian us but I assume it's like you identify some form of a point let's say like ghto staking rate has gone up whatever I assume that's what you humidify humidify humidify like bison has been doing the most volume you give like the CSV and you say hey like clean the data up give me like this that's fine but if That makes sense.

But I guess but I guess like my my push back on that would be like and I'm biased because I believe in a lot of this stuff like quite hard like if if you believe that the path of like continued improvement for all these models is going to continue at like even a decent clip then like in a year or two years like it is actually just going to be better than all of us on this call at our jobs and like there won't be a reason for you to do cryptocurrency research because the LM will be better at it than you. So like it's much it's actually just more efficient for you to prompt it and say do this because it's smarter than you if you believe that.

Right. Yeah. So like there's a few things. So let's like go down the rabbit hole and say like this is what's going to happen. Um in which case I guess it doesn't matter that much how much work you learn but it's more like an intrinsic thing where it's like I actually also want to learn it on the side. That's I guess that's the only point. Um, if that doesn't happen, um, like if there isn't a super form of intelligence that is like better than all of us, which I doubt it isn't already better than most of us, um, then I think it's fine to probably just have more learn. I don't know.

I read some something about some guy who said in the future like yeah like AI will be able to do everything much better than you but it will probably like eventually we'll shift from humans plus AIs are doing things now and then we'll switch like only AI is doing it and then probably after that they'll say well AI is doing it but with human touch because humans appeal better to humans which I think makes sense.

I was like way more skeptical and I think I'm always skeptical whenever like the commentary comes from people out of their domain expertise like if you just hear like you know financial analysts or like nons super technical people talking about like the rate of change and like the path they see in the future. What I think I've been like particularly alarmed by is people that like have extreme product knowledge. I I forget who this guy was. I can't bring the name up but um he's like a PhD theoretical physicist that works at an AI lab. I I forget if it's open AI or somewhere else, but he wrote an essay of like how he expects within the next like three or four years like the cutting edge models to be better than like 95% of PhD theoretical physicists, right?

So like from someone with like w within that domain and they have that expertise like for them to make a comment like that, I think I give that like much more credence than I would someone that's outside of the the domain.

Wait, can you repeat what the comment was? Sorry.

It it was it was basically a guy. He he was saying like his expectation is is in the next couple years like the cutting edge LLM will be better than you know most like cutting edge researchers within theoretical physics today and they they'll be able to put out um you know like novel research and to the same degree that like you know the top percentile of of physicists are doing today.

Yeah, I believe I believe that. I mean you've got like Neil Smani uh solving like Erdos problems which is just an insane arc. Uh, but he's doing it with like Chad GBT and he's like posting about it and stuff which is funny. But yeah, I think like most any quantitative problems and stuff like that, you can just kind of assume that's going to be uh handed off rather quickly. Like a lot of the like a lot of the things that he's solving, they're like since just like like hard math problems that like haven't been solved.

Uh like there like aren't like proofs for them. I mean he's just like kind of uh I mean he's a very smart guy obviously but he's working with like you know chat GBT and Claude and stuff and solving these these unanswered problems uh which is interesting but um I I think like the big open question I guess is like how it all like impacts markets and like um at what point things start getting like kind of handed off to uh like AI fund managers and things like this and then um just how different will all of those things be like to my understanding like a lot of like the um like uh like HFT firms as well like their largest cost now is like um energy because a lot of them are trying to like run all of these things locally. They're trying to like internalize as much of the stack as possible.

I think that's kind of interesting too. So u but yeah.

Yeah.

I go ahead. I was going to say maybe an interesting thought experiment is like I I agree a lot with this idea that I don't I think you see it pretty well represented by people who are fairly elite like at what they do and then using AI to kind of enhance their capabilities like on the on a design or like creative side like Reed at Block Works has made some ads for Blockworks like almost entirely via AI but it's like if you didn't have the creative experience to like understand what a good ad or like a good video or good visuals and sound looks and sounds like then you it's not like you you wouldn't be able to compete with him off rip by just prompting Claude or Gemini.

So there is this amplification factor. I think the bigger question is like to James and Ian like in the f if we get to this point if it's in one year who knows where the AI are just so much incredibly smarter than even the like top smartest most elite humans that exist in in specialized fields. like is there then a human bottleneck at that point where sure they can prompt AI to like solve difficult problems but if the you know how to how to bring that into the real world to like provide solutions for manufacturing or like the devices and the equipment that like you can create from it like there's like a level of human comprehending comprehension that's still required so I wonder if that becomes a bottleneck or how much you can really outsource like okay now the now the equipment that we make and the computers that we make are just so AI ingrained that you don't even need, you know, you don't need an operator to even press the button at that interface anymore or even understand what the should be.

James just flipped out at that comment.

Yeah, that's an interesting point, Danny. I don't know. I think a lot so much of it is just unfolding so fast. It's just like you barely even have time to think about it and like another direction that it go. But yeah, we'll see I guess.

Yeah, I like on the read thing, I think one take has always been that like graphic designers are really cooked after AI and stuff kind of, but like there you need to have some form of taste. like anytime you're interacting with the consumer, you need to have like AI can learn UXUI really well obviously or like it'll know it already really well, but I don't know if it'll know what people's taste will be because that changes quite often and like I think there is like objectively good taste and bad taste and so I think I agree with you on the read point on the second point which is if humans will be a bottle I imagine not.

It doesn't make a lot of sense to me that they would be. It might just be that like humans eventually end up like super, you know, when you like I don't know like I have some form of coding background but not great. Um, so like at a certain point if I'm asking it to do something for me, like up until I don't know, let's say like prompt 10, I get it. And then after that, I'm just like kind of trusting that it's doing the right thing and approaching it the right way. And then by prompt 30, I'm completely out of it. Like I don't know what the code base looks like at all. Um, and I imagine that's what it's going to probably look like. I don't know if humans will be a bottleneck. It's just like more like nerve racking situation.

I think I think that's kind of what I mean though like the simple coding example I think is a bit different than like you know maybe we have 200 IQ scientists that can solve incredibly difficult you know quantum mechanics problems but if like the next unlock in that field is like requires 300 IQ intelligence then and only the AI can solve it and only the AI understands like what that means in relation to you know how you can apply that to the real world is there like a limit where you know it it can solve things and it's like you you can't even pass on the understanding to the humans interacting with it.

Yeah, I think it doesn't matter. Kind of far away from that range though because I imagine it'll be able to explain it to humans that exist right now or like the smart ones. The bigger worry that I have is not like the super smart people. It's like people like me who may not have like intrinsic motivation sometimes to like try and stay up to date with everything that's happening in life. I like I don't I'm not a massive doomer in general. I think typically things will work out some one way or another and it'll be fine. Um, but I do think that there's a non-zero chance that like people are becoming blobbier. I think a little bit like with the internet and whatnot and not a lot of people are like super we have a massive like class of people now who are in kind of fakey jobs and that like AI is going to get rid of them or they're going to just be using AI to do the entire work and then you've just created people who don't use any intellectual effort.

Say you're blogging.

Yeah, I would say I'm blobby-ish. Are you blobby?

Um, one thing I was going to say that I I think is interesting is like Danny, I would be curious your take from like your previous you worked in like engineering, manufacturing, right? Like the birth of what I imagine are going to be like large scale consultancy firms to actually are.

Can you hear me or no?

James, I can hear you, but every once in a while it sounds like Ian's going burp burp, but I don't think he All right, this. He was about to get to an interesting point as well. I can crash out. Um, we can take a question from the audience. Riley Ian, what's it like being in the permanent underclass?

Uh, it's a good question. Um, I recently learned how to bake a pumpern nickel bread. So, that I think is, you know, good luck replacing that artisan skill. You know, I think I'm positioned quite well given How did you get the rest?

Uh, no. I had to go and look in a cookbook. Very good. No, but I don't know, Danny. I think what James was probably trying to get at is that consultants are going to die out, which I think makes sense. Um I'm not sure entirely what he was saying. Consultants will be born to implement AI at corpse. Um current corpse just say here is AI use it good.

I would agree with both of those maybe in different ways. But I would also maybe push back a little bit like from my sort of manufacturing experience there was a little bit of like here's some AI just use it. But then also of course at a higher level there was very much a you know how do we automate as many processes as are reasonable with hands-off you know automated robotics. Um, really like manufacturing is bottlenecked by people today because you have to have you have to have those people who interface with things from time to time.

So, trying to find as many ways as possible to like, you know, have an automated piece of equipment that does something that an engineer would have done, like replace a filter or something like that. Um, or have, you know, a camera that can move up and down a facility and like do spot checks of things to make sure nothing's on fire, something along those lines. Um, replacing that type of work is definitely inevitable. Uh, and I think you see that in a lot of ways. Like, I don't know, there's just so many examples. It's very obvious that like all of manufacturing is kind of like gradually moving that direction and it's just really a cost of like implementing the technology.

At a higher level I could see it as well. I don't know that it would just be something as generic as like a consultant um like an AI consultant and more so I think you you'll have these like really high expertise people once again like we've been talking about high expertise people like kind of really performing in this world and I think the same would be true like I think you would have people who like really understand particular processes or like particular like manufacturing loops and things of this nature and they would you know maybe have pretty good ideas about like what the most optimal layout of a facility would look like and how you would effectively integrate IIA there AI there and robotics and those processes.

I guess you could call them an AI consultant, but I think it it's like it's more than that, right? Like you're not just a guy prompting AI. You have a lot of like deep knowledge in that particular field and probably process, especially like if it's a chemical process as well. Um I've just seen the same thing. It's like the the people who are really elite and like at the top, they're