这一轮周期见顶了吗?| 1000x
- Jonah 的顶层判断是,这轮加密周期尚未见顶,BTC 在9.2万–9.8万美元区间震荡,意味着“观望模式”,而不是自动触发卖出。 如果 BTC 向上突破至12.5万或15万美元,他预计热钱会强力抬升市场已经筛出的领涨项目:Hyperliquid、可能还有Sui、AI16Z和Virtuals。
- 宏观威胁是CPI重新走高,而不是华盛顿政权更迭前8天的一份强劲就业报告。 Avi 承认NFP可能预示通胀,但称这并非必然;Jonah不会因为这一推断在9.2万美元卖出BTC,在Trump治下,他拿10英尺长的杆子都不愿碰债券。
- Sui 可能是下一轮周期的优质资产,但在没有杀手级应用的情况下,两位主持人都不认可今天以约500亿–520亿美元估值买入。 它狂热的社区和开发者信誉或许能熬过加密寒冬,使其从熊市低位成为“另一个100倍标的”;但目前Jonah只看到技术优势,仍无法回答“买家为什么要选择这条链的区块空间?”Avi 对Sui技术的看法来自二手信息:他没有亲自基于Sui开发,但听到过开发者的正面反馈。
- Hyperliquid 单靠交易所业务、按约12–15倍远期倍数计就显得低估,而成为L1后又拥有“自带杀手级应用”的稀缺优势。 Jonah倾向于现在小额积累、BTC突破后再大举买入;Avi在20美元和18美元下方更加看多,因为此前在30美元附近卖出的巨鲸已经开始回补。
- Crypto AI 可以大幅增长,但今天市值最大的代币未必继续胜出,因为框架相比DeFi的内生网络效应弱得多。 Virtuals 从约2.25美元涨至5.20美元后又跌回2.70美元;据Avi判断,回落原因是竞争者出现,他会考虑在2–2.30美元附近重新买入,而Jonah已经放弃了过去持有AI市值前10代币篮子的想法。
- 两位主持人都认为,Agent 是Crypto AI走向大众实用性的最清晰路径,但真正持久的优势来自领域专长,而不是通用代码。 “要把一个Agent训练到擅长某件事,你首先得擅长这件事”:即便市场输入相同,在不同语境下也可能代表不同含义,因此必须由专家先提供判断逻辑,Agent才能成为有用的交易助手。
- 他们从代币发行中得到的教训是,优先配额会把影响者变成卖家,而自然价格发现才能带来利益一致的贡献者。 Avi称,发行前以100万美元估值获得私募额度、随后以1500万–4000万美元估值上市,对内部人而言就是“白捡的钱”;1000X则在约10万美元市值启动,没有私募轮,并通过更广泛的社区找到开发者和品牌贡献者。
1. Sui 是熊市观察标的,不是500亿美元买入标的
Jonah 对Sui最强的看多逻辑,最后竟变成了“谩骂与仇恨”(the vitriol and the hate)。2023年他在Solana接近20美元时发起抨击,相关回应带来约70万次观看,也显示许多人的生计仍押在这条链上;代币跌到10美元时他转为看多,但也承认自己“显然买得不够多”。
Sui的市场反应弱于Solana,但依然暴露出一批坚定的社区成员。据称,Blockworks对Solana开发者的调查显示,Base和Sui远远领先于其他替代方案、并列第一,说明开发者认真看待这项技术;但Jonah始终无法得到“区块空间买家为什么会偏好这条区块链?”的直接答案。
Avi将Sui与Aptos作了比较:Sui的技术团队更强,Aptos的营销和商务拓展更好,但“做Sui的人学会了营销”,Aptos却没能把早期热度转化为用户。不同于NEAR据称空洞的8亿美元生态推动,Avi认为Sui确实有开发者兴趣,且仍在增长,只是除了Suilend借贷协议等熟悉产品外,还没有杀手级应用。Avi提醒称,他本人没有基于Sui开发,判断依赖的是开发者的二手反馈。
这笔交易仍然过不了估值这一关。Sui市值一度达到约520亿美元、与Avalanche持平;Jonah 2023年抨击Solana时,后者尽管社区更大,市值却只有约150亿美元。更理想的做法是等加密寒冬到来,确认Sui的开发者仍在,再行积累,押注其走出类似Solana从8美元涨到200美元的“另一个100倍”行情。
2. Hyperliquid 已经拥有多数L1花数年寻找的应用
Avi喜欢这种倒置:Hyperliquid是一款“反向长成L1”的应用,而不是一个估值昂贵、还在寻找需求的L1。交易所自带杀手级应用,即使再增加几个应用,当交易所本身按约12–15倍远期倍数估值时,也足以支撑更高估值。
Jonah尚未解决的问题是,除了交易所之外,Hyperliquid的低延迟还能为谁带来独特价值。它使用Tendermint和Cosmos SDK,但两人都同意,这并不能构成有意义的ATOM投资逻辑:Avi称Cosmos是“公共基础设施”,对整个行业有价值,却缺乏代币效用,也没有能承接Hyperliquid成功的领导层。
两人的入场策略不同。Jonah会逐步建立小额、高确信度仓位,但会把“大笔资金”留到BTC突破区间之后,因为HYPE还可能再跌50%;Avi则认为,30美元附近完成派发的大户已经开始在20美元和18美元下方回补,仍在卖出的主要持仓者变少,而许多投资者在错过第一轮上涨后正等待机会。
3. AI市场增长不会保护今天的代币龙头
这个赛道的上行空间仍然很大,但边界并不清晰:Avi记得一组AI与DeFi市值对比约为100亿美元对500亿美元,Jonah听到的则是150亿美元对DeFi峰值1500亿美元。两人都没有独立核算这些数字,因此共同结论是方向性的——AI可能还有很大增长空间——而不是对具体规模的确认。
Avi的关键区分在于,DeFi的使用会强化既有龙头:流动性越高,交易所和借贷协议越好,因此投资者可以买入市值前10的项目作为行业篮子。给AI框架增加用户,却不一定能让框架本身变得更好,所以“我认为AI会增长”并不意味着当前前10项目会捕获这部分增长。
Jonah的反驳保留了一个重要细节:在无代码系统之间切换,可能像“从Uber换到Lyft一样顺滑”;但如果要替换他们完整代码实现的Eliza,“会是一场英雄式的痛苦折磨”。Virtuals也有自己的代码框架GAME,只是团队选择Eliza,是因为它在GitHub上的人气更高,更便于试验。
市场已经先后轮动过Bittensor的TAO、GOAT、Virtuals和AI16Z,而不是不断强化某一个持久龙头。Virtuals从约2.25美元涨到5.20美元后又跌回约2.70美元;据Avi判断,回落源于竞争者出现,他可能在2–2.30美元重新买入。Jonah则说,自己“过去以为它会像DeFi一样”,如今更偏好小额、带有期权性质的仓位,而不是重注。
4. 有用的Agent需要交易员编码开发者不知道的东西
Jonah认为,Agent是Crypto AI“跨越鸿沟”的最佳机会,因为一个用独特材料训练、能够互动的机器人,即使在加密圈外也容易理解。他仍不明白为什么每个Agent都需要代币,但承认代币可以类似股权凭证,用来观察项目进展和增长势头。
Avi认为核心约束在于专业知识:“要把一个Agent训练到擅长某件事,你首先得擅长这件事。”随机找来的开发者无法判断哪些数据重要、应该如何合成;同一个输入,在语境A中可能意味着“X”,在语境B中却可能意味着“Y”,这些区别必须由两位主持人手动编码进去。
他们的路线图是先做一个加密交易助手,判断一条新闻是否重要、为什么重要、重要到什么程度;再借助外部专家扩展到大宗商品和其他市场。更长期的设想包括接入交易终端、提供结合仓位的辅助,以及服务非交易决策的Agent——但他们也承认:“我不知道这会走向哪里。”
5. 优先配额制造的可能是卖家,而不是伙伴
讨论AICC时,Avi指称Bankless拿到了一大笔配额并立即卖出,同时强调AICC创始人未必是坏人。Jonah说Bankless发行了这枚代币,Avi随即纠正:“他们没有发行代币,只是拿到了配额。”两人批评的是配额和出售行为,而不是代币的创始归属。
Avi对影响者天使轮的更大批评是:以100万美元估值获得额度,等代币以1500万、2000万或4000万美元估值上市时,就成了“白捡的钱”(free money)。内部人的优势在于低价入场,而不是长期信念;这会制造出先对外展示信仰、随后立刻卖出的压力——“这就是我他妈最恨这个行业的地方”。
两人将AICC事件形容为声誉上的“自我毁灭”。Avi提到约200万美元,Jonah用粗略的分摊和税费计算后认为,每人可能只剩约50万美元。这些数字只是他们的估算,但原则是明确的:如此变现受众,可能摧毁后来开展项目所需的信任。
他们给出的替代方案是1000X:项目在接近10万美元市值时启动,没有私募轮,也没有提前配额,最初只是玩笑,后来才认真推进。自然价格发现带来了没有被利用感的持有者,也带来了开发者和品牌贡献者;Avi还补充说,在“98%的情况下”,有名的KOL投资者只出钱,随后便消失。
6. 他们关注的宏观威胁是CPI,而不是就业数据
Jonah认为“市场已经亮牌”:BTC在9.2万–9.8万美元之间震荡时,HYPE、SUI、AI16Z和Virtuals持续走低,暴露出热钱的去向——那些自我感觉变富的人,正在把现金砸向这些项目。若BTC涨向12.5万或15万美元,他认为上涨潮会重新抬升所有资产,也将迎来他更愿意增加山寨币敞口的时点。
Avi把NFP后的恐慌称为“事后资本”(hindsight capital)。30年期美债收益率已经在12月1日至1月1日期间从约4%升至5%,却几乎没有引起关注;等到强劲就业数据公布、收益率大幅上行后才把利率称为“尖叫着要做的交易”,意味着“那时候……行情已经被定价”。NFP可能预示通胀,但并非必然;他称需要关注的关键数据是CPI。
Avi认为,Trump“很可能是一位通胀型总统”,因为他偏好经济保持火热,施压要求降息,支持关税,并希望减少非法劳动力;这些政策可能迫使Fed暂停降息,甚至重新加息。Jonah的结论很直接:在政权更迭前8天、BTC位于9.2万美元时卖出,意义不大;他看空债券、预计收益率走高,并称自己拿10英尺长的杆子都不会碰债券。
Imagine selling Bitcoin at $92K because you're worried about a nonfarm payrolls print 8 days before regime change in Washington. And not just any regime change—this regime change, right? So, to me, I think we're in wait-and-see mode.
Jason
Hey, everyone. Jason here. There's a reason why Ledger is trusted to secure more than 20% of the world's crypto assets today. Check out their newest devices, the Ledger Stax and the Ledger Flex, featuring the world's first secure touch screens. Take your security seriously. Hit the link in the show notes of today's episode to check out their latest devices. Trust me, you won't regret it. Okay, welcome back to another ThousandX podcast.
We got a lot to talk about this week, don't we, Avi? A lot is going on.
We do have a lot going on. The market has been very volatile—up and down, left and right. And we got a lot of hate for our last podcast because we talked about how much we hate Sui. We got attacked pretty mercilessly by the Sui community. They remind me a lot of the XRP guys, actually.
1. Sui Has A Powerful Community
You know, it's funny. I took another crack at trying to understand why Sui matters after all the vitriol and hate, and the best bull case I could come up with for Sui was the vitriol and hate, right? There was this moment in late 2023 when I put out this tweet. I'd had a bottle of wine, I was in Venice having a good time, and I put out a tweet with a little bit of free time about how Solana is fucked. I talked about well-articulated, but not that well-researched, reasons why Solana was ultimately doomed to fail.
The token was trading at around $20 at the time, and it immediately got around 700,000 views. Everybody started hating on me. That bald guy who's the cheerleader for the community started freaking out, calling me names and insulting my parents and stuff. I thought, “Okay, wow, I guess this ecosystem isn't dead. In fact, the opposite: it's vibrant.” People really care. They've invested their livelihoods in this thing.
Then we got Toly on the podcast the next week, and I got bullish at a $10 token, higher than where I was maximally bearish. Obviously, I didn't buy enough, but I bought some. I think something similar is going on with Sui here: if you bear-post and immediately have people coming out of the woodwork to tell you how stupid and ugly you are, that's a sign that people care.
People are really invested in this chain. They're invested in working on it. The community is there, trying to pump this thing and make it go up. It doesn't feel like a bunch of people who are just long at the highs and unsure why. There are some clear technical advantages that Sui has. That was also reflected in the Blockworks Research survey where they asked a bunch of Solana developers, “If you were working on another chain and weren't allowed to work on Solana, what would it be?” The vast majority picked Base or Sui.
Those 2 responses, Base and Sui, were tied. Nothing else even came close. So, to me, the rabid community is the reason why we might see a self-reinforcing rally. Beyond that, I kept trying to debate people on Twitter and ask, “What is the actual business case for this chain?” People would say, “Oh, you're so stupid. You don't understand the tech. It's just better than everything else.”
I would keep prodding and saying, “What's the business case? Blockchains sell block space. Why would buyers of block space prefer this blockchain to any other? Is there any unique product or service that can only work on this chain, or that can't work as well on the more established chains?” No one could answer my question with a straight face. They could only insult me and tell me that I'm a nontechnical idiot.
To me, it reminds me of previous blockchains that have been technologically superior to the incumbent, but that technological superiority hasn't necessarily translated into upward price action. Sometimes it does, sometimes it doesn't. I don't think there's enough there. If you aren't holding Sui bags from the lows and rolling in money right now, I don't think there's a compelling reason to invest here.
I agree with all those points. What's very clear to me is that Sui has managed to build a pretty crazy community, but not just of retail investors. There are a lot of VC funds and hedge funds that have started to take it seriously. I think it stems from the fact that these guys were building at Facebook and Meta. They were a serious team. It was Sui and Aptos, and Aptos kind of shit the bed. Nobody really talks about it anymore.
Yeah, what happened there? They were the more vaunted product when they launched, and then the founder just quit. What the hell's going on there?
When we were diligencing them way back in the day, what was always said to us was that the Sui guys were actually better tech people, but the Aptos guys were better marketing and business-development people. I think what ended up happening with Aptos is that they just never really got the product to where it needed to get to, and the Sui guys learned how to market.
They're good tech people, right? They did build good tech, as far as I can understand it. I'm not a developer. It's not like I've gone on and actually built something on Sui and can verify it, but I'm going off secondhand information from people who have built on it, saying, “Okay, yeah, it's pretty good. Developers seem to like it.”
I've not heard anything of that sort from Aptos. So I do think that they sort of took Aptos's niche away from them. Aptos wasn't able to capitalize on the initial hype that surrounded it. They weren't able to turn that into real usership, and I think what's happening right now is that Sui has just cannibalized it completely.
Now, is it worth $52 billion? No, and I think it's kind of nuts to think that it'll stay up here. But is it something that might actually stick around? Yeah. I think what they've managed to do, unlike a lot of these other chains, is get to a place where people are taking them seriously. Developers are taking them seriously, right? Nobody ever—
That's what it takes.
Nobody ever really took NEAR seriously, despite their claims of having great developer tools and a better ecosystem for building. They raised an $800 million ecosystem fund, but that was all fugazis, right? The people who were building on NEAR weren't actually building anything real. They were just porting over copycats from the rest of the world.
Oh, yeah.
I do think what it does say is that, if you go look at the top assets now, Sui has reached parity with Avalanche, and Avalanche had such a huge head start on Sui. I think what it tells you is that people don't want subnets. People don't want L2s. People want a fast base-layer chain, and that's kind of what you're seeing with the fracturing of Ethereum as well.
People just don't want interoperability between all of these different chains. They want 1 good chain, because otherwise it's too confusing, and nobody has figured out the UI to make all of these subnets very easily usable.
2. Sui Faces The Valuation Test
That brings me back to my original question, though. Solana serves that purpose, right? I understand—I've been educated by the online community of anti-Jonah people that Sui can do things that Solana can't. It's faster. It provides better, to use a nontechnical term, tagging for asset ownership than Solana, perhaps. Maybe it's less buggy. Who knows, though? It hasn't been battle-tested, but let's just assume that it is.
Still, that alone isn't enough. I think Sui is the kind of product that you buy during the next bear market. It will go through a winter. All of crypto will go through a winter. All these tokens will shank. Sui is probably one of those things that you're going to want to accumulate for the next cycle, when everybody writes off all of these alt-L1s as dead, but the Sui builder community is still strong, still grinding, still trying to put apps together and achieve product-market fit.
Then some smash hit will come out during the bear market that you can capitalize on in the next bull. I don't think it's a good trade to be buying it here at $50 billion, and that's nothing against Sui. That's just markets and valuations. It's overpriced relative to what it is right this second, and it'll probably be a good dip buy for the next run.
I think that's an amazing framing. I think that makes a ton of sense. It's one of those things that, if it exists through a bear market and maintains its community, that's going to be like another 100x-er after the next bull.
You could kind of do this with Solana, right? Everybody thought Solana died, but they maintained their core community and their developers.
People never fully gave up on it, even post-FTX, and that led it from $8 to $200, from the low to the highs in the bull just recently. So I think the same thing could happen with Sui if they manage to keep that hype.
Looking at the TVL on Sui and looking at the applications on Sui, they don’t really have a killer app yet in any meaningful way. I’d say their biggest protocol is the Suilend protocol, which just looks like you go to the website and it looks like any other borrow-lend protocol that’s ever existed. But the one thing that I’ll say is it’s growing. Sui is growing quickly.
Yeah. The amount of hate that we got for suggesting, God forbid, that Sui might not be good value for your money at this particular point in time was not even close to the amount of hate that I got for suggesting that Solana was torched, right? When I suggested that, there were literally 100x as many Solana diehards coming out of the woodwork to engage. Whereas this time—
Yeah, and—
Yeah.
And actually, at that point, Solana was actually lower in price.
Yeah. It wasn’t at the highs. It was close to the lows.
Let me actually see. Where was Solana when you decided to shit on it and call it stupid? That was last— that was October.
That was November 23.
Yeah, that was— No, I think it was October.
October, yeah.
Yeah, it was October of ’23, and it was at about $15 billion in terms of market cap. So actually about a third of the size that SUI is today, and it had a larger, more active community around it.
I mean, look, I agree. The way to play this is kind of just to ignore it until you get to the bear market, and then it’s a freaking amazing trade if you get into the bear market and it still has a good community surrounding it. So I’d definitely keep an eye out.
3. Hyperliquid Moves Toward L1
One thing that we actually haven’t really talked about on this podcast is Hyperliquid and its plans to become an overarching L1. What’s kind of cool is that very rarely do you see an application go backwards into an L1 because normally the L1s have a massive valuation premium. So if you’re going to launch any project, you’re going to launch an L1, right? And if you’re launching an application, you’re probably launching it on an L1, and you don’t have any ambitions to become an L1.
But now Hyperliquid has sort of backed into this, which has massively raised their potential valuation. It’s also cool because you now have an L1 that has a baked-in killer app, which is very rare. Normally, L1s are searching for their killer app instead of the killer app becoming an L1.
So we’re off about 50% from the highs. I think people have generally stopped talking about it in the way that they were talking about it 2 weeks ago, when the entire timeline was just bull-posting Hyperliquid, which is almost never a good time to buy.
When the biggest holders of this thing feel so good about themselves that they can end every tweet with “Hyperliquid” and go after people left and right mercilessly, I think you should be cautious. Flood is, I think, the biggest proponent of this product. Flood was having way too much fun on Twitter when Hyperliquid was around $30. He was having way too much fun, and I like Hyperliquid. I’m a fan. I’m an owner. But when somebody’s having that much fun on Twitter with their number-one bag at the highs, you should probably start chipping out a little bit, at least.
Easy to say. We’re all having fun when crypto’s ripping, and it’s hard to know when it stops because you’re not in it for a 10% or 20% rally. You’re in these tokens for a 5–10x at a bare minimum because, at the end of the day, that’s the kind of risk that you’re taking to the downside, too.
I don’t know about Hyperliquid. Well, first of all, just to call a spade a spade, Hyperliquid is built using the Cosmos SDK, right? I looked into this when I did my research a while ago on Hyperliquid and ultimately concluded not to buy it at the lows, which was stupid in hindsight. It uses Tendermint. It’s part of the Cosmos SDK. It’s part of their ecosystem.
So if they become an L1, is that bullish for Cosmos? What else is going to get built on Hyperliquid other than their—
Cosmos is done. Cosmos is now public infrastructure. Their token has no real use. They did great stuff for the crypto community by basically building an actually pretty good framework, but the token itself is done, right? It’s over. That’s in the past. I don’t think it’s ever really coming back. They have no real leadership.
Now, that being said, we are always going to be indebted to them for what they did for the industry, which was build a really phenomenal framework and come up with new consensus mechanisms to drive the industry forward. So we appreciate them for that, but I don’t think buying the ATOM token makes any sense in any meaningful way.
Yeah, I agree. How ironic would it be if the most successful Cosmos outcome is Hyperliquid, which is basically built with Cosmos’s technology, and they benefit not at all from it?
I think that with Hyperliquid— Okay, so let’s say they go backwards and become an L1. You already have a community of enriched bag holders sitting there. They want to keep the train moving forward. They’re going to build all sorts of other interesting stuff on Hyperliquid, or uninteresting.
I think that’s the next move for Hyperliquid. You have to see whether this is just an exchange, or whether they’re figuring out other unique applications for this latency-sensitive platform beyond the same crap that’s on every other chain, including Sui, Solana, Base, and ETH. If it’s just a bunch of Uniswap-style DEXs—
Forget about anything else, Jonah. These guys, just based on being an exchange, are undervalued. If they figure out a way to even get a few other applications built on them as an L1, they deserve to be valued higher. That’s my take.
Right now, I think they’re trading at a 12-to-15 forward multiple, so it’s legit.
Yeah, that’s pretty low. Yeah, I agree.
I see a downward trend channel in the midst of a market where Bitcoin, the benchmark, is ranging sideways. My take on this is that the market has shown its hand for what’s going to perform during the next phase of this bull cycle. And yes, I do believe there will be a next phase.
There’s a debate about whether we’ve peaked or not in this cycle. I think we’re going much, much higher. When we do, the market has shown its hand in terms of what’s going to rip. Hyperliquid is clearly going to rip, right? Those cards got shown a couple of weeks ago. Maybe Sui, as much as I hate it, will rip. The AI coins, like ai16z and Virtuals, will rip.
But in the absence of sector-wide momentum, in the absence of a rising tide, I think it’s interesting that all those coins are just kind of spiraling lower and lower. They’re grinding. They’re not tanking, but they’re grinding lower every day or every week. And to me, that’s relevant because it shows you that that’s hot money. That’s where people who feel rich are throwing their cash.
It’s hard to feel rich when your Bitcoin’s just knocking around like a pinball machine between $92K and $98K, right? But if we take the next leg higher to $125K or $150K Bitcoin, the hot money’s going to flow right back in and power-lift those tokens.
So, I have a more medium- to long-term trading style than you do, but I wouldn’t be buying Hyperliquid, for the sake of example here, at $20 a token right now, trying to catch that falling knife. I would rather buy Hyperliquid at whatever price it’s at when Bitcoin breaks through the top of the range and keeps sending.
By the way, let me clarify one thing. I'm not talking about how you accumulate your first position, right? Yes, it would be good to have some Hyperliquid here, but it could go down another 50% before it starts to rally.
So I think with your conviction buys, maybe you scale in and buy for now with small size, and then if you're going to do a big slug on high conviction, it's when the market's rallying again. And that goes for Virtuals, ai16z, and some of the other stuff that performed recently as well.
I think the thing with Hyperliquid that is a little bit different to me is that there are so many people that miss this. What I watch the most when I look at a chart like Hyperliquid is actually the flows of the token. Are the big buyers buying or selling right now? Basically, once it hit $30, a lot of the big wallets started moving and actually offloading and taking a ton of profit, and a lot of them are just starting now to buy back below $20 and $18.
So what I know is that there's a significant amount of money coming back in to buy this thing below $20 and $18, and the longer that we hold above that, a lot of the people that missed out on that first rally are going to come back in. I just don't see that many large holders offloading anymore. They're just not really selling.
That's valid.
Whereas with Virtuals, just talking about another chart that looks completely different, from the time we first talked about it, it was trading at, like, $2.25—or when we launched our 1000x token, it was trading at $2.25. It goes all the way up to $5.20 and is now all the way back down to $2.70, and people are just hammering this thing. They're absolutely hammering this thing.
The reason that they're hammering this thing, as far as I can tell, is that there are a lot of competitors that have popped up that are better. Look, I love Virtuals because it's where we launched our token, and they were very good to us in the beginning, but we actually migrated off their framework.
To be fair, their no-code framework is what we migrated off of, and we migrated to Eliza, which is ai16z's framework—a you-write-tons-of-code framework. The reason why we did that was because a no-code framework just isn't good enough for most bots. However, to be completely fair to Virtuals, they do have their own you-write-a-bunch-of-code framework. It's called GAME. We just didn't choose to go down that road because Eliza was so popular on GitHub that it seemed like a no-brainer for us to explore that in the limited amount of time that we had.
4. AI Lacks DeFi Network Effects
Well, yeah. That's kind of what I'm saying. I talked about this a little bit on The Steady Lads podcast that I went on Friday, and somebody pulled up a chart of AI market cap versus DeFi market cap. We weren't even halfway there, right? There's still so much more for the AI crypto space to grow. Now, the issue is—
I've seen $15 billion estimates, right? Can you just state the numbers that you saw on the chart for the listeners?
I think I saw $10 billion versus $50 billion.
Okay.
What did you see?
I've heard that AI is $15 billion right now. It's probably $10 billion since I saw that, and at the peak of DeFi summer, DeFi market cap was $150 billion. I don't know. I haven't run these computations myself. At any rate, I just wanted to get the numbers out there for the listeners. Please continue your thought.
There's a difference. There's a huge difference. The difference between these AI things and DeFi is that DeFi inherently has network effects built into it, right? The larger an exchange like Uniswap gets, the more liquidity it has, the better it is as a product. The more liquidity a borrow-lend protocol has, the better it is as a product. So inherently, as winners get larger, they actually become harder to displace, not easier to displace.
Whereas there are basically no switching costs, relatively, in AI. It's so much easier. You don't lose anything by switching from one framework to another, really, in the same way.
I disagree with that. I mean, you're right in the sense that we could technically clone our agent on GAME or any other framework—
Mm-hmm.
—or just build one ourselves. And then when we get our bot to a sufficient level of competency—
Wow.
—we just turn off our Eliza bot and migrate to the new bot, but that would be a heroic pain in the ass. What I will say is, for a no-code framework like what Virtuals has promoted to get us on board, it is very easy to switch. You literally just copy-paste a few questionnaire boxes from one framework to another.
For the no-code stuff, it's literally as seamless as switching from Uber to Lyft for a given ride. For a full-code framework like Eliza, it would be pretty heroic for us to switch. We'd have to do a lot of work.
Let me reframe this. The more people who use a framework doesn't necessarily make it better. Obviously, if you have better or smarter people contributing to a framework in greater quantity and improving the framework, then it's better. But just the usage of the frameworks doesn't make it better.
And so I think that the network effects are a lot smaller. What I'm trying to get to at the end here is that in DeFi, it was pretty reasonable to assume that once you hit $20 billion in market cap and you think you're going to $100 billion in market cap, it's very reasonable to just buy the top 10 DeFi protocols and say, “If DeFi does well, these top 10 DeFi protocols will do well.” Whereas in AI, I think it's a lot harder to say, “I think AI is going to grow, therefore I'm just going to buy the top 10 biggest AI projects,” because it's so much easier to disrupt them.
Dude, it's so interesting. I like that take. I hadn't thought about it that way before, but we've seen this now with 3 different protocols, right? We've seen first it was Bittensor—TAO went to the moon and then just kind of stalled out, like, plateaued. It was weird.
Okay, maybe that's a unique ecosystem that's a little bit complicated and hard for people to understand. Maybe the space is going in a memier direction. All right, so then Virtuals shows up. Virtuals goes to the moon. We're all feeling pretty happy.
Dude, you forgot about GOAT.
Yeah, GOAT was cool. I put it in the meme coin basket, but—
There are so many different cycles already.
Yeah. These things, though, I used to think that this was going to be like DeFi, where it's like, okay, AI mania is coming to crypto. That's my conviction trade. I'm just going to be long the top 10 call options that position me for that, and that doesn't look like it's going to work. These things just rip and then they fizzle.
And to your point about network effects in DeFi, the analogy isn't really working in AI. I think the more people that use the Eliza framework, the worse it becomes, almost, because your bot is going to sound like all the other Eliza bots. The more people use a certain framework, the more commoditized it becomes, and the attention you're trying to farm with these bots—you’re not trying to be a commodity to farm attention. You want to stand out. You want to be unique to farm attention.
So I take your point there. I still think AI mania is coming to crypto, and I still think that all of this stuff is going vertical. But you've definitely made me more cautious and made me want to sell some. I'm deeply in the money on Virtuals, deeply out of the money on ai16z, which I added on the highs, and then I've got a constellation of other little coins that I'm experimenting with that don't really move the needle in my portfolio.
I don't know. I think you're right. I think it's probably like you hold small percentages of those things just in case they come back, but you don't place big bets on them here—Virtuals and ai16z.
Yeah. That's a little bit of what I'm thinking. Virtuals is like, if it really washes out, I'm probably rebid between $2 and $2.30. I think that's a reasonably good level to get back in. But right now, it's just a full downward trend, and the conversation is that it's being outpaced by every other framework out there.
And so, therefore, it doesn't really make sense to get in and buy it.
5. The 1000x Agent Vision
I don't know. We should probably talk about our token, podcasters launching tokens, and where some of these agents can go in terms of vision and the future. I personally think agents are going to be the part of this AI sector that really sticks and achieves mainstream appeal in the next 12 months. Would you agree with that?
Yeah. Yeah. I think that's very reasonable. I don't really know where else AI fits in, to be completely honest.
Fits into crypto?
Like, where else other than—yeah, like, where? Where does AI and crypto fit in besides these types of protocols?
Maybe DePIN, decentralized compute, but I don't think that's decentralized GPUs or something. Render's been trying to do that, and they've dipped their toes in.
It's not really getting traction. To me, an agent is really the only thing where you could cross the chasm from crypto nerds to the mainstream. We could talk about what we want to do with our agent and how it could potentially cross that chasm at this particular juncture. I don't know.
Basically, to me, an agent—a bot that posts content, is interactive, and is trained on some unique training set—has universal appeal. It doesn't matter whether it's Tesla's Full Self-Driving, trained on their millions of hours of driving videos, or a little Twitter bot that's trained on our podcast. Agents have universal appeal and can be understood by people outside of the crypto world.
I think I'm still trying to understand why these projects need tokens, but it is sort of a form of equity that people can use to monitor the progress and traction of an agent. So it's useful in a certain sense. Tokens aside, though, our 1000x agent has begun tweeting interesting crypto updates, synthesis, and takes on those updates. It incorporates momentum and technicals into its takes.
It tries to understand the backdrop for that news and synthesize a bullish or a bearish opinion with degrees of bullishness and bearishness. It understands what's neutral. It understands what's not. We've built in a character file that references our podcast transcripts. We've built in crypto market niche knowledge that you and I have accumulated over years. The bot understands how to think about certain types of developments in the crypto market.
Which, by the way, is actually a completely underrated portion of this: in order to train an agent to be good at something, you need to be good at it. You can't just have random devs building trading agents because they don't know how to trade, and so they don't know how to even teach the bot: What parameters do you put in? What data do you ingest? What do you look at? How do you synthesize it? A lot of that is manual.
Oh, it means X in context A and Y in context B. So you and I have literally had to input that into the bot and train it, and building that for crypto is something you and I could do. Building it for oil is something I could do. Building it for distressed credit, you and I might struggle, right? We'd need to bring on an outside expert.
Right, but we can bring— But I think one thing that's becoming very clear to me is how useful these things are going to be in a broad variety of contexts and how, actually, once you dive into it, Jonah, you've spent a lot more time on prompt engineering than I have, but there are some core tenets that hold true across whatever you're trying to do, whatever market you're trying to teach this thing on. There are some core tenets.
And so once you build one, it becomes a lot easier to build the next one because you understand the process. At the end of the day, this 1000x bot is going to start on crypto, but it's going to expand. It's going to become a trading personal assistant for whatever market you want it to be a trading personal assistant for, which I think is going to be pretty sick.
It's just going to require a lot of—we're going to grit our teeth and work through it—but it's going to be pretty broad-based and broad-reaching by the time that we're done with it. I think it all started as a little joke.
Yeah. It was a pet project. You launched it while I was sleeping, then we started jamming on it together. It's so fun to work on this.
Basically, the roadmap would be: we're going to make this bot excellent at its current mission, which is your crypto market sidekick that keeps you up to date on everything that's going on and helps you understand whether or not it matters, and, if it matters, why and how much, right? So we're going to make that as good as we can for crypto.
Then I think it's a pretty light lift to copy-paste that into commodities, and we can just crowdsource the relevant commodities market context. Then we can put it in every other market as well.
But even before we do any of that, outside of crypto stuff, we have sponsors, partners, and people in our network that run interesting crypto platforms where we could incorporate the 1000x bot into their terminals and their user experience. This thing could look at your positions and be your little sidekick, your voice in your ear that talks to you about what you're doing, in a way that you couldn't necessarily rely on ChatGPT to do for you. So I think there's utility there.
And then the big vision, once we've got this thing going for all markets, is to replicate the process outside of trading, incorporate it into other businesses, and have this agent be present in other categories of marketing or areas where there's decision-making and user engagement.
If you're deciding what movie to go see, maybe there's the Gladiator III agent reply-guying to politicians as Maximus. Maybe there's the shoe bot shilling Nike in replies to Adidas's tweets. I don't know where this is going, but this journey feels like it has legs. It's one of those good gut feelings that's fun to work on.
Yeah, and it's a great learning experience for sure, regardless of how things actually end up. What's cool is when you get to interface with technology that is genuinely doing something new, it's pretty good stuff. And I think it's only going to get—
I was going to say, yeah, this feels good, but AICC or some other podcast launched a token and it all went to diddly-squat pretty quickly. What did they—
I mean, first of all, it's still trading at 60 mil, which is kind of nuts.
That is nuts.
But I think it speaks to—look, it speaks to a level of grift that exists within crypto. I think that the actual founders of this AICC thing are not necessarily bad people. I actually think that they're good people. They were trying to build a good project.
But you have to be very careful with the types of people that you allow into your project. I think what ended up happening is that they just got too close with the Bankless guys, and the Bankless guys kind of took advantage of them. They ended up getting a pretty massive allocation, then immediately dumping.
But I think, to take a step back for a second, I've been in this industry as a quote-unquote KOL for 8 years now. And I don't do a ton of angel investments, kind of for this reason: the unspoken thing about all these angel investments—and this is going to piss off a lot of people—is that you get access to projects that you're guaranteed to make money on, not because they're going to be good long term, but because you get in at an unbelievably low price.
I have tried to avoid this at all costs. But a lot of people, when they get these opportunities, it's a home run. It's free money, because you're getting in at something like a $1 million valuation, and you know it's going to go live at at least 15 or 20 or 40.
But you also understand that your entire edge is getting in at that low valuation, so of course you're going to sell right away. You don't believe in the project long term. You don't want to hold this thing. You think it's nonsense. But you have to pretend like you believe in it long term, because otherwise nobody's going to buy it.
But you know in your heart that this thing is making you money because you got in at a cheap price. And that's what I fucking hate about this industry, and that's why there was no round for 1000x.
A lot of people, by the way, got upset with me over this. A lot of people were like, "How? Adi, you didn't fucking tell me that you and Jonah were launching? You didn't tell us? What the hell?" I'm like, "Why would I tell you?"
You didn't even tell me.
I love you—
I just woke up.
You might be a really good friend of mine.
Yeah.
But, to be fair, I was just fucking around. I genuinely had no intention of launching this thing as a real project. But even once you woke up and we talked and we were like, "Okay, maybe we should take this seriously," I still didn't tell anybody, because why? Why would I, right?
Then what it looks like is that it's just a pump and dump. And the real way, if you have an audience and you want to launch a project, I think the best way to go about anything is organically. Let it be discovered organically.
Because that creates a strong base, creates trust between you and your community, creates a strong base of holders that don't feel like they were taken advantage of, and it absolves you of accusations of grift. You know?
No, that's true.
And this is very key. You can't even seem like—right, there are a lot of actions that you can take.
For example, if you raise a round because you think raising a round will allow you to get in the right people, and they'll be able to help you the right way, what I've found is that in 98% of cases, unless you're working with somebody who has a genuinely stellar reputation, most people who are KOLs will put in their money and then disappear. They won't talk to you, they won't help you, and they won't do anything. They just put in their money and then they're gone.
People just want names. They want name brands. They'll say, "I want to stick my name on this. I want to stick my name on this. I want that guy's brand. Let me go raise." It's not necessarily the right way to do things. I don't think it's the right way to do things.
No, I agree.
What you really want is people who are bought in, and that's, at the end of the day, our community. The Bankless guys—I mean, I've never really had any respect for them in any meaningful way.
Oh, shit.
But I think that this is an obvious outcome.
They produce good content. They've been in crypto for a long time. What I don't understand is why you would get yourself involved with a project to pump it up and then dump all of your tokens on launch day onto the general public.
Because they made money.
How is that possible?
If you love Ethereum, you're probably poor. All I can say is that torching your reputation for $2 million is the funniest thing I've ever seen in my whole life.
If you've been in this industry for as long as they have, you've been running a podcast, and you've been in all these angel investments, and you're not worth at least 10, 15 bucks, something is wrong. You're not cut out for this. You're not cut out for it.
Maybe they're not even trading. To me, it's like $2 million split between the 2 of them, so that's $1 million each. They probably live in high-tax jurisdictions, and they're Americans. So that's $500K.
You torched your reputation for $500K? Are you kidding?
$500K. Golf clap. I don't know. To me, it makes no sense. I'm so shocked by this.
We launched a token at the same time as Bankless, and they rugged their entire community at the same time.
Just to clarify, they did not launch the token. They got an allocation.
Sorry. Yes.
We launched our token at, I think, a $100,000 market cap.
Yes.
Our community is frigging awesome. We found our developer through the community, and we found our branding guy through the community. It's completely different.
I think what crypto is great at is incentivizing groups of people who don't know each other, all across the globe, to help each other out without gigantic reams of documents, startup equity, lawyers, and startup expenses. If you do it that way, you have to get your community involved at an attractive price. You can't let them in on the highs and then hammer it all the way down in their face and expect a good outcome.
I wouldn't have pegged the Bankless guys as doing that until this, though. But now they've torched their reputation. Nobody's going to deal with them anymore. That's fucking crazy, what they did to themselves. That's epically stupid.
It's pure, pure, pure self-immolation.
For not that much money. If they'd done it for $500 million apiece, I would've been like, "Okay, well, hat tip." But this is small potatoes.
6. Trump Changes The Macro Regime
Should we talk about NFP, macro, and Bitcoin? To me, very quickly, I want to talk about it. I feel like the whole market got itself all worked up over an NFP print, and it didn't matter.
I've seen this a lot in young traders, and I disagree with it. Maybe there are some interest-rate guys out there who are really good at trading around these economic events, or FX traders or macro traders, but for the most part, the only thing that's going to torch crypto, Avi, macro-wise—macroeconomic-data-wise—is a screaming-high CPI print. If inflation is back with a vengeance, then we can all get worried about crypto again like we were in 2022.
But jobless claims and nonfarm payrolls—who frigging cares? These prints come out regularly. They don't matter. The Fed isn't going to change its policy because the economy is going well. Ninety percent of Washington is about to turn over in a week and a day.
You also have to remember that a lot of the talk you see is massive hindsight talk. It's hindsight capital.
There was an article in Bloomberg that came out after the NFP print that said, "This is a huge shock. Rates are going to go through the roof now because of all of the uncertainty around Trump, inflation is going to come back, rates are going to go through the roof, and it's a screaming buy."
All I could think to myself was, "Where were you when the 30-year went from 4 to 5 from December to now?" If you look from December 1 to January 1, the 30-year rate went from 4 to 5. That's a huge move. It's a big move.
Yeah.
Nobody was talking about it. People only talk about it when it really hits you in the face, and by that point, the move is priced in. It's really hard to make a bet on that. You can't extrapolate.
The right bet would've been to say, "Okay, I think there's going to be a rate move," before all of this goes down. But after you get a huge move, talking about it just screams hindsight capital.
The NFP came in strong, and it could be an indication that inflation is going to come back in a big way. But that's not a given. The only thing that we have to watch now is CPI.
What I will say is that Trump is likely to be an inflationary president. The things that he wants to do and the things that Trump cares about are inherently inflationary. He wants a screaming-hot economy. He wants to pressure the Fed to cut rates.
No more illegal laborers, high tariffs.
No more illegal laborers, exactly. There is an argument that with Trump in charge, inflation will come back, which means the Fed will have to pause, and we might go into a rate-hike cycle.
If the Fed pauses, Trump is going to go ballistic on the Fed. What he wants is for the Fed to cut rates while inflation goes up, pump the stock market, and maybe pump Bitcoin this time, too. But he can't really have that because the Fed is independent.
All I'm saying is, imagine if you sold Bitcoin at $92K because you're worried about a nonfarm-payrolls print 8 days before regime change in Washington. And not just any regime change—this regime change.
To me, I think we're in wait-and-see mode. The 1 thing I would take away from a Trump administration, macro-trend-wise, is that I wouldn't touch bonds with a 10-foot pole. I'm bearish on bonds. I think yields are going higher. You do not want to be earning fixed income in Trump land. That is crazy.