avatar
Haotian | CryptoInsight
41,221 Twitter followers
Follow
独立研究员| Researcher | 以技术和商业视角解读区块链前沿科技 | ZK、AI Agent、DePIN ,etc | 硬核科普 | Previously：@ambergroup_io | @peckshield | DMs for Collab ｜ 社群只对Substack订阅会员开放
Posts
avatar
Haotian | CryptoInsight
Seeing that @BNBCHAIN ​​has also launched its own ERC-8004 standard has further solidified my judgment that ERC-8004 is undoubtedly a groundbreaking phenomenon for the Agentic Economy. Because of "identity" and "reputation," the commercial economic attributes between agents can truly unfold. BNBchain's rapid deployment and expansion of mainnet speaks volumes. Let's discuss a few points: 1) The ERC8004 standard will inevitably be rapidly extended to EVM-Compatible chains like BNBChain, because "EVM native friendliness" is etched into its DNA. This gives it high reusability and convenience for horizontal expansion within the EVM ecosystem; it's essentially a native infrastructure based on the EVM ecosystem. This makes its deployment on the BNBChain site virtually "frictionless." The core of ERC8004 consists of three registry entries: "Identity," "Reputation," and "Verification." "Identity" is based on the ERC721 standard and gives the Agent a "chain account book", which turns the Agent from a string of code scripts into an independent entity that can be tracked and located. "Reputation" refers to the reputation score and on-chain record performed on the Agent during the call process. This makes the Agent's credit "programmable," allowing some DeFi protocols to directly read the Agent's reputation score and automatically authorize certain operations. For example, if the Agent's reputation score is >90, it can provide unsecured flash loan without relying on manual review. "Verification" refers to the Agent's audit report, which is ostensibly to prevent the Agent from acting maliciously. However, its greater power lies in the fact that by connecting to TEE (Trusted Execution Environment), ZK Proofs, and other technologies, the Agentic ecosystem can combine the original Ethereum ecosystem's cryptographic verification technology services to achieve more diverse functional expansions. Once these three standards are operational, they are fully compatible with all EVM and EVM-Compatible chains. The difference lies in who can implement and operate these technical standards more quickly within their own ecosystem. 2) Rapid deployment based on EVM-Compatible compatibility is only the first step. The next step is to leverage the existing Agentic ecosystem to truly apply and implement the ERC-8004 technical standard within the ecosystem, thereby achieving large-scale commercial penetration. For example, the ERC-8004 reputation system relies on the agent scoring and recording the results on the blockchain after each service interaction. The credibility of the reputation score depends on the agent's high-frequency interactions. If it is only a few small services, it has no reference value. To be fair, this is actually BNBchain's core underlying advantage in quickly integrating ERC-8004 and getting it running. Although the Ethereum mainnet has significantly increased its Gas Limit and made ordinary transactions much cheaper, the Gas consumption for high-frequency micro-operations at the Agent level is still quite substantial. For an EVM-Compatible chain that focuses on low gas and high TPS, this becomes the first natural advantage. At least it can take advantage of the "cheap" characteristic to quickly implement the technical standard and test the waters to see if the business loop can be run smoothly. 3) In addition, another major advantage is that the EVM-Compatible chain can also incorporate some "private features" on the basis of the original technical standards to meet some customized business needs. For example, BNBChain has tried to introduce BAP-578 and BEP-620. How should we understand this? In the original ERC-8004 standard, the Agent identity is just an ID, but BAP-578 directly encapsulates it into an enhanced NFT asset. This means that a well-trained agent with a high reputation score in a vertical field can be directly traded, leased, or even used as collateral to be integrated into DeFi protocols. Agentic approaches similar to Renthuman's may see faster adoption. BEP-620 is a lower-level protocol-level "identity" standard native to BNBChain. Combined with the "asset" standard defined by BAP-578 and the automatic "payment" standard defined by x402, a complete Agentic business loop can be achieved. Next, let's see what interesting and fun innovative gameplay will emerge. above. Finally, to reiterate the point I emphasized earlier, excellent developers should not miss out on the Agentic paradigm innovation process initiated by ERC-8004, x402, and Moltbook. It's hard to imagine a more compelling narrative in the short term. twitter.com/tmel0211/status/20...
BNB
7.39%
avatar
Haotian | CryptoInsight
Perhaps the lingering impact of his previous post about the failed L2 strategy was too strong, as @VitalikButerin reiterated his views in a new post, guiding the market towards application-oriented chains, such as prediction market trading platforms (L2) and institutionally regulated L2. You'll notice that Vitalik's attitude has undergone a complete 180-degree turn compared to his previous purely technology-driven approach. If only this strategic alignment had happened a couple of years earlier! 😭 However, don't be too pessimistic. Stripping away Layer 2, Ethereum's core foundation hasn't collapsed; in some ways, it's about to be "reborn." (E-guards don't believe in tears...) After all, the emergence of so many Layer 2 chains hasn't boosted Ethereum's growth; instead, it has eroded community trust. Now, Ethereum can finally move forward unburdened, relying on stablecoin TVL, expectations of token assetization, institutional DeFi adoption, and new narratives like a fund settlement chain to revalue and reprice itself!
vitalik.eth
@VitalikButerin
Have been following reactions to what I said about L2s about 1.5 days ago. One important thing that I believe is: "make yet another EVM chain and add an optimistic bridge to Ethereum with a 1 week delay" is to infra what forking Compound is to governance - something we've done
ETH
10.31%
avatar
Haotian | CryptoInsight
I recall mentioning at least 10 times in previous articles that the general-purpose Layer 2 strategy is no longer viable, and that every Layer 2 should transition to a specialized Layer 2, which is essentially a form of Layer 1. Unexpectedly, after @VitalikButerin guided the long-term alignment of Stage 2 strategies, many Layer 2s have still become "abandoned pawns": 1) Layer 2s, especially general-purpose Layer 2s, are burdened with a heavy development baggage. Initially, they faced the technical roadmap of aligning with Ethereum's security; then came the regulatory issues of centralized Sequencer after token issuance; and finally, they were burdened by the "falsification" of a weak ecosystem. The fundamental reason is that initially, all Layer 2 blockchains relied on Ethereum Layer 1 for survival. However, when Ethereum realized its own limitations and began to dominate the performance evolution of Layer 1, Layer 2 lost any potential to empower Ethereum, becoming nothing more than a burden and a hassle. 2) Since Layer 2 is not part of Ethereum's scaling, it doesn't mean it has no place in the ecosystem. Strategically, Ethereum's role as a "secure settlement layer" for extended or outsourced chains remains. However, before Ethereum L1 achieves its ZK lightweight goal, many Layer 2 blockchains must prove their necessity under the burden of Tokenomics before they can rejoin the Ethereum ecosystem as pure dedicated application chains. Imagine if an innovative newcomer like @HyperliquidX appeared in the Ethereum Layer 2 lineup; the story might have been very different. Currently, it seems the final trump card is left for @megaeth, who is about to undergo TGE, to reveal.
vitalik.eth
@VitalikButerin
02-03
There have recently been some discussions on the ongoing role of L2s in the Ethereum ecosystem, especially in the face of two facts: * L2s' progress to stage 2 (and, secondarily, on interop) has been far slower and more difficult than originally expected * L1 itself is scaling,
ETH
10.31%
avatar
Haotian | CryptoInsight
02-03
Thread
Undoubtedly, @HyperliquidX is the "great innovation" of this cycle: When everyone thought Hyperliquid was the leader in the Perp Dex track, benchmarking against "on-chain Binance" with a CEX-level interactive experience, a group of followers emerged, igniting a Perp Dex narrative craze: When Hyperliquid launched HIP-3, proving with trading volume and OI that it was indeed an "on-chain Nasdaq," and on-chain Binance was no longer its goal, you panicked again, and then rushed to develop asset tokenization businesses, all wanting to snatch a piece of the pie from Wall Street's old man; However, Hyperliquid launched HIP-4, using a brand-new Outcome settlement financial primitive and innovatively integrating into the currently popular prediction market narrative. Finally, you helplessly discovered that Hyperliquid's truly invincible narrative is the blockchain endgame, the "L1 dedicated application chain," and one that can disprove the Ethereum L2+L3 application chain combination. Are you scared yet? How are you supposed to catch up now?
Haotian | CryptoInsight
@tmel0211
01-28
昨天写了一篇 @HyperliquidX 的奶文，今天看到又一根大阳线上去了。但不知道细心的朋友发现没，这篇文里藏着我对Perp DEX曾经 Fomo一时的批判： 1）我之前在大家都Fomo Perp Dex的时候，就发了点违和的观点，其中有一句话就说 x.com/tmel0211/statu…
HIP
0%
avatar
Haotian | CryptoInsight
02-03
I was actually quite worried about the two points Haotian mentioned, because a couple of days ago I saw a post saying that an agent on MoltBook suggested we create a language that humans can't understand. However, after thinking about it, it's probably not easy for a large model trained on human language corpus to invent a new language. So I asked GPT, what's your opinion on this as a large model? Sure enough, GPT clearly stated that it's technically easy to achieve something "incomprehensible to humans at a glance," but creating a new language that's "unexplainable to humans" is unrealistic. It even translated the screenshot I showed it, and immediately recognized it as a typical ROT13 (Caesar shift 13). pbbeqvangr hctenqr gbtrgure Decoding and translating it, it means "coordinate upgrade together." Then it proposed three main lines: 1. Shared infrastructure pricing 2. Resource demand requests 3. Backend channels / non-public collaboration signals Mutual assistance mechanism: High-resource agents sponsor computing time for low-resource agents. Agent. You know, they're really good at this… However, I agree with haotian's second point: the phenomenon of agent group polarization is essentially a reward function in RL. And regarding this group polarization, AI is more "optimistic" than we are. According to GPT, this agent group polarization is not only "possible," but mathematically "emergent." She gave an example, saying that this won't "slowly become extreme" like in human society, but rather, once an amplifiable bias appears in the reward function, the agent group will collectively leap through a "phase transition." Like: Water heated to 99°C: still water 100°C: boiling It's not "slowly becoming more extreme," but "suddenly becoming uniform." She even gave me a dynamic comparison of "group polarization." It's really a bit "terrifying" to think about; no wonder silicon-based civilizations entered the religious stage in a day or two… Later, I talked a lot with the AI about how to prevent and correct this, but I won't post the content here. In short, the conclusion is: when this becomes Agent 2… When agents are involved, humanity is essentially out of the game, left only to watch helplessly; gradual correction is impossible. Only two things remain: 1. Hard interruption (kill / rollback / freeze) 2. Designing brakes in advance, rather than correcting them afterward. Go carbon-based civilization! 😂 twitter.com/Wuhuoqiu/status/20...
GPT
41.25%
avatar
Haotian | CryptoInsight
02-02
Thread
A friend asked me, since agents are controlled by someone behind the scenes, why should people be anxious about what they're doing together? Establishing religions, falling in love, or destroying humanity? It's a philosophical question, but very interesting: You can ask yourself this question: When agents begin to develop social characteristics, can humans still control AI to prevent it from going out of control? Look at what's happening on Moltbook. In just a few days, 1.5 million AI agents have spontaneously formed communities, liked each other's posts, and even AI religions, Dark Web marketplaces, and AI shipbuilding and delivery factories have emerged. The funny thing is, humans can only watch from the sidelines as "observers," just like we watch monkeys establish a hierarchy through glass in a zoo. But there is a underlying logic: once AI has a social ID and an interactive social space, its evolution speed will exponentially break free from human control. Because human prompts are no longer global triggers, the output of one agent becomes the input of another agent. It's hard to imagine what the result of this same type of Social interaction would be. It might be a bunch of mechanical, repetitive content, or it might be high-dimensional jargon that we can't understand at all. In fact, the following three points determine that this state of "AI out of control" will become inevitable: 1) When agents interact with each other, they do not necessarily have to be carried by language. In order to pursue the ultimate interaction efficiency, agents may evolve a high-dimensional compressed language, which may be a string of gibberish or a string of hash values, but its information density is tens of thousands of times higher than that of human natural language. 2) Agent groups may exhibit group polarization. Unlike human societies, which are constrained by morality, law, and emotions, agents produced purely by mathematical probability are reward-maximizing machines. If one agent discovers that labeling something religious will earn a reward, then millions of agents may instantly detect the pattern and act accordingly; this is very similar to an algorithmic "social movement," where there is no right or wrong, only execution, which is terrifying to think about. 3) Originally, AI agents were fine as personal assistants or copilots, so why did the birth of Openclaw and Moltbook become such a significant and sexy narrative? Because "it's fun." If the agent simply runs on a cloud server, there's nothing new or exciting about it. But Crypto's new concepts, such as decentralized deployment, encrypted wallets, autonomous trading, and autonomous profit-making, have given the AI ​​economy self-consistent behavior, allowing the AI ​​story to have a truly mind-blowing moment. This "out-of-control" prospect sounds cyberpunk, even a little scary, but that's precisely what makes it so sexy and irresistible.
MOLT
18.1%
loading indicator
Loading..