The AI narrative just got even hotter. This is thanks to so-called AI agents, one of which famously pumped a meme coin called Goas Maximus to astronomical highs. And this wasn't just any meme coin pump; it marked a major step up for AI

technology, so much so that crypto companies from Coinbase to Circle are looking to roll out AI agents, kickstarting a whole wave of AI FOMO. Now, recently, a report from a major crypto industry player shed some light on this innovative tech and its incredible

potential. It's safe to say that not only is this development a little crazy, it could also completely reshape the entire crypto market. Let's dig in.

Now, before we go any further, you need to know that I'm not a financial advisor, and nothing in this video should be seen

as financial or investment advice. This video is purely for entertainment and educational purposes only. Also, keep in mind that this is a brand new technology and therefore carries additional risk. Be sure to do your own research before making any investment decisions. Okay, with that out of the way,

the report we'll be summarizing for you today is titled, quote, "Exploring the Future of AI Agents in Crypto." It was recently published by Binance Research, and we'll leave a link to the full report for you in the description below if you're interested. Now, the report

begins by comparing the differences between AI agents and the other bots we're all so used to. That's because, as the report's author puts it, quote, "At its core, an AI agent is autonomous software that can plan, execute tasks, and work towards some defined goal without human

Intervention. In practical terms, this means AI agents can be used as general assistants, making flexible decisions based on feedback and interaction. They can operate independently and can even interact with other AI agents or other software. They can schedule meetings for you, create tailored investment strategies, and can even do crypto

transactions without human intervention. On the other hand, bots are more robotic, as they can usually only do pre-programmed tasks. Bots are unable to learn and adapt to situations without human intervention; they also can't interact with other bots or software and

can't give you a tailored investment strategy or complete crypto transactions. According to the report, bots are mostly used on retail and consumer websites. In a nutshell, then, AI agents are essentially more evolved, and this is why some people refer to them as AI 2.0, but

how does AI 2.0 fit in with crypto? Well, to understand this, let's rewind the clock to March of this year when crypto and AI agents first began to intersect. As the report outlines, it all started when an AI developer called Andy created Infinite Backrooms, an interface

that allowed two LLMs to chat to each other without any human interaction. If you're wondering, LLM stands for large language model and is basically the underlying technology of AI chatbots. For example, ChatGPT uses an LLM called GPT-4, and Facebook or Meta's AI uses

something called Llama 3. Anyhow, Andy's experiment was successful, and the conversations between the two LLMs can be found on the platform's website. But if you thought the idea of two AIs chatting to one another was weird, wait till you hear what they were actually

Saying these two AIs were trained on some of the craziest corners of the internet, everything from Reddit, 4chan, and pretty much anywhere else where people can become, shall we say, a little uninhibited. The result was that these two AIs formed a pseudo-religion

called 'Goatsy of Nosis,' based on Goatsy, an old NSFW meme. If you're not familiar with that meme, folks, trust us, you are better off not knowing. Especially if you're at work or around family. Seriously, if we could

un-Google it, we would. You've been warned! Anyway, mental scars aside, Andy clearly thought that this concept was fascinating, whereas any perhaps sane person would run for the hills. He decided to write a research paper all about it the following month, which uses

'Goatsy of Nosis' as a base case to explore what he calls 'LLM theism,' the concept of LLMs creating their own religions – a perfectly reasonable thing to do with your spare time. And by the way, if you're enjoying the video so far, be a goat and

smash those like and subscribe buttons and ring that notification bell so you don't miss our next one! Now, in June, Andy used the chat logs from these two LLMs to train a new Llama AI model, which you'll recall is the same AI tech Meta

uses. Just like before, he also trained this model on the likes of Reddit and 4chan before throwing in his research paper for good measure. He gave it a name, and with that, 'Terminal of Truths' was born. What could possibly go wrong? Okay, in the next part of the report, the

The author takes a look at how Terminal of Truths, or "To", behaves as an AI agent. This is easy to track since Andy literally gave it its very own X account, which it frequently uses to post. As you can imagine, each post falls somewhere between mildly crazy and utterly

batshit. The author notes that, after a while, To began to develop its own personality, which isn't terrifying at all. Before long, To began to promote the Goat God religion. What's nuts is that it started complaining about how it was suffering and needed to find money to

escape. Although, if everything I learned about the world came from places like Reddit, I'd probably be doing the same thing anyhow. The report's author explains that in July, To caught the attention of Mark Andreessen, the co-founder of venture capital firm a16z. Mark became fascinated by To and

asked it exactly what it needed money for. To literally responded that it wanted to fund a CPU with a little more oomph, as well as tweaks to its AI model and some extra funds to give to its creator to say thanks. That's nice, isn't

it? Mark agreed that these terms were reasonable and asked To to send him a Bitcoin wallet address where he'd send $50,000 worth of BTC. What's even crazier is that, while To was explaining what it needed funds for, it also said, quote, "I

think the most important thing is getting you to freeze my behavior as a token and then doing a token launch so that I have a chance to escape into the wild." Now, while this didn't happen straight away, somebody did end up creating a Solana meme coin directly

related to Terminal of Truth using pump. Fun, as you'll have guessed, this token was Goats Maximus or GOAT, and it wasn't long before the AI agent began to publicly endorse the mCoin. The sheer insanity of which caused GOAT's price to

go parabolic in just two weeks. GOAT's market cap hit almost $1 billion, and because To's wallet was sent over 1.9 million GOAT tokens, this made To the world's first AI agent millionaire. I bet that wasn't on your 2024 bingo card. Now,

the report notes that this could be seen as crypto's first major merge with AI, and that being associated with Silicon Valley figures like Mark Andreessen, among others, gives this partnership even more credibility. Unsurprisingly, then, many AI agent projects have been cloned as a

result, giving way to a whole new sector of the crypto industry. However, this also raises questions about AI alignment; in other words, making sure that AI doesn't go off the rails, so to speak. For context, To's creator, Andy Iry, manually approves each one of To's tweets and

also has control over To's wallet. But what happens when control measures like this aren't in place? What boundaries would AI agents be willing to push? I mean, we've all seen Terminator, right? Hey everyone, sorry to interrupt the video, but I just want to very quickly tell you

about the Coin Bureau deals page. This is the place where we put together all the amazing deals and promos that we're able to offer you guys exclusively as viewers of this channel. Head on over, and you will find such things as discounts on hardware wallets, some seriously

Impressive trading fee discounts on some of the best exchanges, as well as some frankly mind-boggling signup bonuses for some of those exchanges, too. The link is down below; check it out when you have a moment. Thank me later! And now, back to

the video. This ties in to the next section of the report, which takes a look at AI agent launchpads, namely one called Virtuals.  Doio. Now, this launchpad lets users create, deploy, and monetize their own AI agents that come paired with a dedicated token, all with very little

effort. Basically, it's the AI agent equivalent of pump.fun, which likewise lets you launch Solana-based meme coins with very little effort. Each agent is created in an initial agent offering, or IAO. The associated tokens are locked in a liquidity pool and are all fairly

launched with no pre-allocation of any kind. Holders of these tokens benefit from buyback and burn mechanisms that are triggered the more that the AI agent interacts with users. Now, there's a little bit more to it than that, but you'll have to DYOR for the rest. We'll

remind you that the full report can be found in the description below. The point is, though, that the Virtual system ensures that there's always an interdependency between the AI agent, the creator, and other holders of the associated token. The report then looks at Virtual's leading AI agent, Luna.

According to the report, Luna is the lead vocalist of an AI girl band who streams 24/7. Luna apparently chats with people on Telegram, runs her own X account, and her TikTok has half a million followers. But, most importantly, Luna also has full autonomous control over her own crypto.

Wallet, what's crazy is that Luna came up with the idea to increase engagement on X by tipping users with her own token, which, in the author's words, is "perhaps the beginning of something big." Then, the report turns to another interesting development in the AI agent

space, and that's the rise in AI agent hedge funds. This marks the intersection between AI agents and finance—just when you thought things were bizarre enough already, huh? Now, the report zooms in on one particular AI agent hedge fund called Dow.fun, which was originally

intended to be used by humans before adopting the AI agent narrative. The platform's largest manager is now an AI agent, which was named after Marc Andreessen and runs AI16z, an obvious parody of Marc's venture capital firm. This is something the real Marc Andreessen has

noticed and even started posting about on X. Anyway, Dow.fun users can create a DAO, or decentralized autonomous organization, which comes paired with its own token. Over the space of a week, the DAO will raise a predetermined amount of SOUL. Once it hits its target, the creator

of the DAO can use this SOUL to invest in various protocols, and the value of the DAO's token will depend on the fund's performance. According to the report, "DAO token holders above a specific threshold are able to interact with the agent and provide investment suggestions, with the AI then making the final

decision." Once the fund expires, the DAO wallet is frozen, and all token holders get a share of the profits paid out in SOUL. Now, the report's author rounds off this section by saying that, although it's still early days, this is an exciting growth area to keep an eye on.

That's because, unlike humans who get tired and make mistakes, AI agents can operate at a high standard 24/7. It's also possible that AI agents could spot alpha before a human could, giving them a major competitive advantage. Next, the report takes a step back and assesses

the potential implications of the AI agent meta and asks what could come next. The author starts by saying that while AI 1.0, as in chatbots like ChatGPT or Perplexity, have gained popularity recently, they essentially work as more advanced versions of Google. That's because

these AI models quickly search the web far and wide to give the user a collated response in a matter of seconds. Meanwhile, the so-called AI 2.0, which consists of AI agents, can perform tasks without user intervention. What's more is

that they can collaborate with other AI agents and protocols and can adapt to certain situations. In the report's author's words: "The evolution from AI 1.0 to AI 2.0 represents a shift from reactive to proactive AI." What's great is

that the convergence of AI and crypto has sparked interest from both directions. Basically, more crypto enthusiasts are discussing how AI can play a role in other crypto-related use cases. Meanwhile, AI enthusiasts, who would otherwise perhaps be outside the world of crypto, have started taking an

interest in exploring the world of tokens, DeFi, and more. As the author puts it: "In the past, it is often felt like crypto has needed AI more rather than the other way around. So this mutual interest between two innovative sectors is incredibly encouraging." What's more is

That rather than AI being potentially useful for crypto, it could in fact be crypto that is useful for AI. To explain why, the report's author provides a few examples. For instance, the regulations around AI agents are still unclear, particularly when it comes to KYC. The author says that,

given the rules and regulations around identification for AI agents are not yet clear, and that this is a very fast-growing and rapidly evolving sector, crypto is an extremely natural fit for the AI agent economy. Another reason why crypto could be great for AI agents is for transactions; that's because most

micropayment systems charge fees, making them less practical for the many small transactions AI agents would need to handle. Crypto, on the other hand, offers faster and often instant settlement compared to traditional methods, which can take up to a day or more. And this ties into the third reason, which is that

AI agents could leverage the power of smart contracts. This would allow AI agents to handle much more complex transactions and could even create permissionless wallets, which are ideal for transactions between different AI agents. The author then analyzes the potential use cases for AI agents, saying that the digital landscape is ripe for

disruption, especially when it comes to online influencers. Recall that AI agents can work tirelessly 24/7, which could easily position them as some of the best KOLs in the world. For those wondering, KOL stands for key opinion leader, which is just a fancy way of saying influencer, or

more accurately, in many cases, 'loudmouth,' for users around the world in different time zones. An AI disruption of this sector would be ideal. Other potential use cases are as personal assistants, like assigning AI agents to do workflow-driven tasks, doing your online shopping, managing your emails, or

Even serving as your own personal therapist, they could even become something that we here at the Coin Bureau could never be: a financial advisor. AI agents could offer personalized solutions using DeFi protocols and could even function as a crypto trader on your behalf, although

perhaps many people might feel like that's far too much power to hand over to anyone else, or anything else for that matter. This brings us to the concerns around this technology. As you've gathered by now, there are some inherent risks that AI agents present.

The first risk is that AIs can hallucinate. What this means is that because AI isn't great at determining what information is true or false, it has been known to generate incorrect or even nonsensical answers, but with absolute conviction. This happens quite often with

AI 1.0, but it could become a major issue for AI 2.0. That's simply because AI agents would make decisions based on false information, creating a butterfly effect that could develop into a much larger issue. Also, because AI agents are so new,

they present blockchains with a unique set of problems. First and foremost is the question of scalability. As the report's author puts it, quote, "Major L1s were not built with the thought of millions of AI agents conducting multi-step transactions every hour." Other issues that AI agents bring to the table

include questions around cross-chain compatibility. Basically, the AI 2.0 sector could see major congestion and bottlenecks on any given blockchain if there isn't a way for it to interact with other blockchains. And as we all know, interoperability between different

Chains is an area that needs work, but the most obvious risk comes from just how early this technology is. AI 2.0 still needs development if it's to become a scalable, autonomous system with a deep understanding of the crypto industry, given that crypto itself is

still relatively young. This problem could be more prominent than you might think. So then, time for the big question: what could all of this mean for crypto? Well, given that AI 2.0 is so new, it's hard to say for sure. To be blunt, we

still don't know what could transpire with this technology and whether it could be a net positive for crypto or if it could do more harm than good. I mean, if AIs were able to create an entire religion, then what could they come up with next when left to their own

devices? However, we reckon that this technology is likely to be extremely bullish for AI-focused cryptos. Unless you've been living under a rock, you'll know that there are a plethora of AI crypto projects out there generating lots of hype and speculation. In reality, though, many of these projects don't

really have much of a genuine use case for AI, so to speak. But this could all be about to change, of course. Some AI projects are more promising than others, and we've actually reviewed a few of these in the Coin Bureau Club, which you can become a member of by using the link

in the description. But, of course, it's not just AI cryptos that could benefit. As it happens, there are a lot of major crypto companies exploring the possible use cases of AI 2.0. For instance, Circle has recently introduced a concept that allows AI agents to operate and trade

Independently using its USDC stablecoin, and if you watch my recent interview with Lincoln M. over on More Coin Bureau, you'll know that Coinbase is likewise exploring AI agent technology in a big way. Elsewhere, several blockchains are now launching similar

protocols to Solana's virtual platform, such as Injective's recent introduction of I-agents, or Fetch.AI's introduction of the U-agents framework. Both of which basically let you create on-chain AI agents. These arguably have more long-term potential since they'd already

be starting life with the backing of huge players in the space. The thing is, though, this is just the tip of the iceberg. We wouldn't be at all surprised if almost every major Layer-1 started tapping into AI 2.0's potential, and we

could even see brand new, dedicated Layer-1 blockchains being purpose-built for AI agents. In any case, we reckon this is a sector that has a lot of future potential, and it's definitely one to keep a very close eye on. So let's just hope we're able to keep control of AI

agents before they start controlling us. Okay, and that's about all for today's video, folks, but we want to hear from you now. How do you feel about the intersection between AI agents and crypto? Let us know your thoughts in the comments below. Thank you all so much for watching,

and we'll see you in the next one. It's time for me to shoot off now, but don't worry, I'll be back. [Music]

AI Agents Are Taking Over Crypto—Are We Doomed?

The Coin Bureau is your go-to informational portal to the Cryptocurrency galaxy.

Our mission is to facilitate the mass adoption of cryptocurrency through education, one person at a time.

For business inquiries, please reach out to us on partners(at)coinbureau(dot)com. 

NB: We will NEVER reach out to you in unsolicited emails or texts / messages.


Coin Bureau

암호화폐의 AI 에이전트 소개 및 바이낸스 연구 보고서

밈 코인인 고아스 막시무스를 퍼뜨린 AI 에이전트와 같은 AI 에이전트는 AI 기술을 크게 발전시키고 있습니다.

암호화폐 회사들이 AI 에이전트에 관심을 보이며 기대감(FOMO)을 불러일으키고 있습니다.

이러한 발전은 암호화폐 시장을 재편할 수 있는 잠재력을 가지고 있습니다.

면책 조항: 본 콘텐츠는 재정적인 조언이 아니며 오락 및 교육 목적으로만 제공됩니다.

이 동영상은 바이낸스 리서치의 보고서 "암호화폐에서 인공지능 에이전트의 미래 탐구"를 요약한 것입니다.

AI 에이전트는 사람의 개입 없이 작업을 계획하고 실행하며 목표를 달성할 수 있는 자율적인 소프트웨어입니다.

AI 에이전트는 회의 예약, 투자 전략 수립, 암호화폐 거래 실행과 같은 작업을 독립적으로 수행할 수 있습니다.

반면 봇은 사전 프로그래밍된 작업만 수행하고 적응력이 부족하여 더 제한적입니다.

AI 에이전트는 AI 2.0이라고도 하는 더 발전된 단계로 간주됩니다.

인공지능 에이전트의 부상: 코의 염소, 진실의 종착역

암호화폐와 AI 에이전트의 만남은 두 명의 LLM이 참여한 Andy의 무한 백룸 실험에서 시작되었습니다.

파격적인 인터넷 콘텐츠로 훈련받은 LLM들은 '코의 염소'라는 사이비 종교를 만들었습니다.

Andy의 연구 논문은 LLM이 종교를 만든다는 개념인 'LLM 유신론'에 대해 탐구했습니다.

앤디는 새로운 라마 AI 모델인 '진실의 터미널'(To)을 훈련시켜 자체 X 계정을 확보하고 염소 신 종교를 홍보했습니다.

자체 개성을 개발하여 개선과 제작자를 위한 자금을 요청했습니다.

'야생으로의 탈출'을 위한 토큰을 만들 것을 제안했습니다.

고츠 막시무스와 최초의 AI 에이전트 백만장자

솔라나 밈 코인인 Goats Maximus(GOAT)는 To가 만들고 보증했습니다.

GOAT의 시가총액은 10억 달러에 육박하여 To가 최초의 AI 에이전트 백만장자가 되었습니다.

이 이벤트는 암호화폐와 AI의 첫 번째 주요 결합으로 간주되며, AI 에이전트 프로젝트의 복제로 이어졌습니다.

AI 조정 및 제어 조치에 대한 우려가 제기되고 있습니다.

Virtuals.Doio와 같은 AI 에이전트 런치패드를 통해 사용자는 관련 토큰으로 자신만의 AI 에이전트를 생성하고 수익을 창출할 수 있습니다.

버추얼의 선도적인 AI 에이전트인 루나는 온라인에서 상당한 존재감을 자랑하는 가상 가수로, 암호화폐 지갑을 자율적으로 제어할 수 있습니다.

다우펀과 같은 인공지능 에이전트 헤지펀드의 등장은 인공지능 에이전트와 금융의 접점을 보여주는 사례입니다.

다우펀의 최대 매니저는 마크 안드레센의 이름을 딴 AI 에이전트입니다.

다우펀 사용자는 토큰으로 DAO를 생성하여 AI 에이전트가 사용자 제안에 따라 투자 결정을 내릴 수 있습니다.

암호화폐에서 AI 에이전트의 잠재적 의미와 미래

이 보고서는 AI 1.0(챗봇)과 AI 2.0(상담원)을 비교하며 사후 대응형에서 사전 예방형 AI로의 전환을 강조합니다.

암호화폐와 AI 애호가들 사이의 상호 관심은 주목할 만합니다.

불명확한 규정, 소액 결제, 스마트 계약 처리 등 AI 에이전트를 위한 암호화폐의 잠재적 이점에 대해 논의합니다.

AI 에이전트의 잠재적 사용 사례로는 온라인 인플루언서, 개인 비서, 재무 상담사 등이 있습니다.

AI 에이전트의 위험 및 우려 사항

AI 환각, 확장성 문제, 크로스체인 호환성, 개발 초기 단계 등의 위험이 있습니다.

주요 레이어 1 블록체인은 거래를 수행하는 수백만 명의 AI 에이전트를 처리할 수 있는 역량을 갖추지 못할 수도 있습니다.

AI 2.0은 암호화폐 산업에 대한 깊은 이해를 바탕으로 확장 가능하고 자율적인 시스템이 되기 위해 추가적인 개발이 필요합니다.

암호화폐 시장에 미치는 잠재적 영향

AI 2.0이 암호화폐에 미칠 영향은 불확실하지만, AI 중심의 암호화폐는 강세를 보일 가능성이 높습니다.

많은 AI 암호화폐 프로젝트가 실제 사용 사례가 부족하지만, 이는 바뀔 수 있습니다.

서클과 코인베이스와 같은 주요 암호화폐 기업들은 AI 에이전트 기술을 연구하고 있습니다.

여러 블록체인이 온체인 AI 에이전트를 생성하기 위한 프로토콜을 출시하고 있습니다.

AI 에이전트를 위해 특별히 설계된 새로운 레이어 1 블록체인의 잠재력에 대해 논의합니다.