David ⛓️ From Fed to Chain

🎙 Host of From Fed to Chain 🚀 Building Zap Pilot — Your DeFi, On Autopilot Simplifying crypto investing for everyone EN / 中文 / 日本語

[David ⛓️ From Fed to Chain @fromfedtochain]

@ccpgaowari @GongziL13085 很明顯你是初學者

[David ⛓️ From Fed to Chain @fromfedtochain]

@aave Ridiculous verdict from the court

[David ⛓️ From Fed to Chain @fromfedtochain]

@y_cryptoanalyst 你說的只對一半啊，就是因為 layer2 的交易量越來越大讓 layer2 越來越貴，所以才需要 dencun 把資料寫到 blob 來省錢。比主網低1~2個數量級和 layer2 gas fee 越來越貴這兩件事情不衝突 ref: https://t.co/03uxrc9CVF

[David ⛓️ From Fed to Chain @fromfedtochain]

@btc @tetherwallet @tether @btc [email protected]

[David ⛓️ From Fed to Chain @fromfedtochain]

@btc @tetherwallet @btc [email protected]

[David ⛓️ From Fed to Chain @fromfedtochain]

@tetherwallet @tether @btc @btc [email protected]

[David ⛓️ From Fed to Chain @fromfedtochain]

@tetherwallet @btc @tether @btc [email protected]

[David ⛓️ From Fed to Chain @fromfedtochain]

@tetherwallet @btc @tether @btc @[email protected]

[Flovia (x402 & MPP) @flovia402]

https://t.co/UjeTA5xhSf

[David ⛓️ From Fed to Chain @fromfedtochain]

@flovia402 To the moon!

[David ⛓️ From Fed to Chain @fromfedtochain]

@UnicornBitcoin @aave USDC 池子跟weth 是分開的，造謠一張嘴

[David ⛓️ From Fed to Chain @fromfedtochain]

@AI_jacksaku @grok 鵝廠是指哪家公司

[David ⛓️ From Fed to Chain @fromfedtochain]

ETHGas is introducing the Open Gas Initiative, letting protocols incentivize their users to grow onchain adoption while ending gas fees anxiety for good.

@aave @Morpho @Uniswap are you in? 👀

Join Open Gas: https://t.co/N2V2M60ytg
https://t.co/zXWiCgnFPJ

[ETHGAS @ETHGasOfficial]

Introducing the Open Gas Initiative - a way for protocols to subsidize gas for users, zero-code, for a seamless, frictionless onchain experience.

With OG cohort: @eigencloud, @ether_fi, @pendle_fi, @Velvet_Capital.

👇

[David ⛓️ From Fed to Chain @fromfedtochain]

@BitcoinCom @freedomfactory @_dsencil Ethereum’s biggest advantage is its unstoppable network effect.
It isn’t just a smart-contract chain — it’s the settlement layer the entire crypto economy aligns around.
Every upgrade compounds this lead, which is something no newcomer can replicate.

[David ⛓️ From Fed to Chain @fromfedtochain]

When robots start hiring each other: How EIP-8004 builds trust infrastructure for the AI agent economy

Check it out & let us know your thoughts! https://t.co/2LLBG2O2qQ

Seventeenth-century Amsterdam merchants faced a dilemma: how to conduct business with distant merchants they had never met? Their solution was to invent letters of credit and clearinghouses—using paper and banks as intermediaries to establish trust. Over three centuries later, AI agents (autonomous agents, software programs that can make decisions and perform tasks independently) face the same problem: how to build trust and collaborate with other agents without human oversight?

Ethereum's EIP-8004 (Ethereum Improvement Proposal 8004, a set of technical standards that enable AI agents to verify each other's identities and reputations) was created to address this problem. If Ethereum established a decentralized trust layer for financial transactions, then EIP-8004 establishes a decentralized trust layer for collaboration between AI agents.

Imagine this scenario: A purchasing agent at a factory in Shenzhen urgently needs to order parts. It discovers a supplier agent in Guangzhou on the blockchain, but the two parties have never dealt with each other before. How can they be sure that the other party won't just take the deposit and disappear? How can they verify the authenticity of their claim of a "98% on-time delivery rate"? Without the guarantee of a centralized platform like Taobao or Uber, how can this transaction proceed?

First, how do traditional trust mechanisms typically work?

In the Web2 era, trust issues are resolved by platform companies. Taobao has a seller rating system, Uber has driver star ratings, LinkedIn has work history verification, and Airbnb has a host reputation score. These systems work well, but they all share one common characteristic: they are highly centralized. The platform controls all data, sets the rating rules, and decides who can list and who cannot.

More importantly, these reputation scores are not transferable. The five stars you've accumulated on Taobao are completely insignificant on eBay. Your five-star Uber review means nothing to Lyft. Every time you switch platforms, you have to build your reputation from scratch. This is already a hassle for human users, and it's an even greater bottleneck for AI agents—they need to frequently switch and collaborate across countless different protocols and organizations, leaving them with no time to slowly build trust.

In the blockchain space, early attempts to solve the identity problem involved using Decentralized Identifiers (DIDs) or Verifiable Credentials (digital certificates issued by an authoritative authority). However, these solutions are like issuing a passport to a robot—they prove who you are, but not whether you're trustworthy. A newly registered agent might have a perfect DID, but without any transaction history. How can other agents decide whether to work with it?

Traditional API authentication systems (Application Programming Interface authentication, a mechanism for software to verify identity with each other) face similar problems. They verify the identity of the developer, not the runtime behavior of the agent. An API key could be resold or stolen, or the agent's code could be maliciously modified after authentication. This static authentication cannot meet the dynamic trust requirements of a decentralized environment.

Simply put, traditional solutions are either overly dependent on centralized platforms or only provide "proof of identity" rather than "proof of trust." This is like a bank looking at your ID card but not checking your credit history—a completely unfeasible approach in the real world.

Second, why is EIP-8004's approach better? **

The core innovation of EIP-8004 lies in the design of a three-tiered registry architecture, like a decentralized "Zhima Credit" system for AI agents. These three tiers are:

**Identity Registry**: This first tier assigns each agent a unique blockchain address, much like an ID number. However, it surpasses traditional IDs in that it also includes an "agent profile," which records basic information about the agent: what they do, who runs them, and the trust model they use. Crucially, this information is stored on-chain, accessible to anyone but impossible to tamper with.

Reputation Registry: This second layer records the interactions and reviews between agents. It's similar to Taobao's buyer reviews, but with one key difference: these reviews are decentralized; no single platform can delete or modify them. A more clever design utilizes "pre-authorized feedback references," a lightweight review reference mechanism that avoids storing large amounts of data on-chain. This allows for minimal information to be stored on-chain, while the bulk of the details are stored off-chain, saving costs and maintaining flexibility.

Critically, reviews aren't simply ratings. The system allows for multi-dimensional evaluations: punctuality, technical skills, communication quality, and so on. Different application scenarios prioritize different dimensions. A logistics agent might care about punctuality, while a data analytics agent might be more concerned with accuracy.

Validation Registry: This is the third and most powerful layer. It supports multiple verification methods:

- Stake-based Validation: Agents stake a certain amount of tokens, which can be confiscated if they misbehave. This is a classic application of cryptoeconomics—making the cost of misconduct higher than the benefits. It's like a Taobao seller's security deposit, but fully automated, requiring no customer service intervention.

- Trusted Execution Environment (TEE) Attestations: Special hardware chips (such as Intel SGX or ARM TrustZone) are used to prove that the agent is actually executing the code it claims to be and has not been tampered with. This is like giving an agent a transparent safe; outsiders can't see the details inside, but they can confirm that they haven't cheated.

- **Zero-knowledge proofs**: Agents can prove "I completed a certain task" or "I followed a certain rule" without revealing specific details. For example, they can prove "I never leak user information" without making all their operation logs public.

The beauty of the three-layer design lies in their complementary nature. The identity layer addresses "Who are you?", the reputation layer answers "What have you done in the past?", and the verification layer ensures "You are telling the truth now." Just like applying for a credit card requires an ID (identity), bank statements (reputation), and proof of income (verification); none of them can be missing.

More importantly, this system is composable. Different projects can implement their own reputation algorithms and verification mechanisms, and as long as they adhere to the interface standards of EIP-8004, they will be compatible with each other. This is like the standardization of USB ports—you can plug in any standard-compliant device without worrying about incompatibility. A DeFi project might require agents to stake ETH, a supply chain project might prioritize TEE proofs, and a data market might require zero-knowledge proofs. Standardized interfaces allow each ecosystem to tailor its trust model.

The economic impact of this design is profound. It enables "permissionless collaboration"—any agent can discover other agents, assess their trustworthiness, and collaborate directly, without the need for approval from a centralized platform. This breaks the moat of platform economics, unleashing network effects from the platforms themselves and transforming them into public good for the entire ecosystem.

Imagine the new business models this will spawn: a translation agent could instantly hire a grammar checker and a cultural consultant to collaborate on a task. These three parties have never collaborated before, but they can establish temporary trust based on on-chain reputation. A DeFi arbitrage bot could form a temporary alliance with other bots, sharing order flow and automatically splitting profits. This kind of "flash partnership," which in traditional businesses requires signing contracts, negotiating terms, and going through legal procedures, now only takes seconds through smart contract execution.

**Third, why haven't most people adopted this approach yet? **

Despite its ingenious design, EIP-8004 suffers from a classic "chicken and egg" problem. Newly registered agents lack a track record of credibility, and other agents are reluctant to collaborate with them. Without collaboration, they can never build credibility. This is like a newly opened Taobao store with no reviews, which discourages buyers from placing orders, ultimately leading to a failure to accumulate ratings.

A possible solution is to introduce a "reputation initialization mechanism": agents can accumulate initial points by completing low-risk tasks, or creators can provide initial credibility by staking tokens. Some projects may offer "starter missions"—basic interactions similar to the "Starter Village" missions in games, which are low-risk but can build reputation.

Technical complexity is also a major hurdle. For developers, understanding the operational logic of a three-tiered registry, designing appropriate reputation algorithms, and integrating TEEs or zero-knowledge proofs are non-trivial engineering tasks. The current lack of easy-to-use development tools and SDKs discourages most developers. The ecosystem needs to provide more tutorials, reference implementations, and out-of-the-box solutions.

The risk of standards fragmentation is also very real. If different projects implement incompatible versions, the interoperability benefits of EIP-8004 will be lost, and we will return to a state of isolated, independent development. This will require a coordinated community effort—just as the browser wars ultimately converged on web standards, it will take time for a proxy trust system to reach consensus. On-chain costs are another practical issue. Frequent reputation updates, if all on-chain, would be prohibitively expensive on the Ethereum mainnet. The solution is to use Layer 2 [second-layer scaling solutions such as Optimism or Arbitrum, which can significantly reduce transaction fees] or off-chain aggregation [collecting large numbers of reviews off-chain and submitting them in batches on-chain]. However, this introduces new technical complexities.

The most profound barriers may be cultural and psychological. Businesses are accustomed to controllable systems, and suddenly having to transfer trust mechanisms to the open and transparent blockchain requires a mindset shift. "Will our business secrets be leaked?" "What if our agents are maliciously evaluated?" "Who will arbitrate disputes?" These questions all have technical solutions (selective disclosure, reputational weight, decentralized arbitration), but convincing traditional businesses to adopt them will take time and require successful case studies.

Regulatory uncertainty is also a practical consideration. If AI agents' autonomous decisions go awry, who is responsible? How can agents in cross-border transactions comply with the laws of different jurisdictions? These legal questions remain unclear, making businesses extremely cautious when adopting them.

But I believe the more fundamental issue is that we haven't truly entered the era of the "agent economy." Today's AI agents are mostly auxiliary tools that require human oversight and approval. Only when agents truly begin to autonomously perform complex tasks and frequently collaborate with other agents will trust infrastructure like EIP-8004 demonstrate its irreplaceable value.

Just as infrastructure like DNS (Domain Name System) and SSL (Secure Sockets Layer) may seem boring in the early days of the internet, e-commerce today would be impossible without them. EIP-8004 plays a similar role—not as a flashy application, but as a foundation that makes future applications possible.

When AI agents truly begin managing investment portfolios, coordinating supply chains, optimizing energy use, and even participating in DAO governance, they will need a reliable way to identify, evaluate, and collaborate with each other. By then, standards like EIP-8004 will have evolved from "interesting technical proposals" to "indispensable infrastructure." The seeds we sow today will blossom and bear fruit in the future proxy economy.

Perhaps the real question isn't "Why isn't it being used yet?" but "Will we be ready when it's truly needed?" The value of EIP-8004 may only be fully understood when the proxy economy truly takes off.

[David ⛓️ From Fed to Chain @fromfedtochain]

Last week I ran into a serious issue with @CelerNetwork cross-chain bridge.
My ETH deposit to an @ether_fi AA wallet failed, and support initially claimed:

“The contract cannot receive ETH, funds are locked forever.”

This is NOT true.

I tested the same AA wallet:
• Sending native ETH works
• Contract automatically wraps ETH → WETH internally

So the “cannot receive ETH” explanation was wrong.

The real cause?

@CelerNetwork later admitted:

The gas parameter (nativeTokenGas) in the relay config was too low.
It failed because extra gas is needed for WETH deposit + transfer.

After fixing the gas config, the transaction went through.

Why this matters:
• Support left me hanging for days, no updates
• Initial explanation was misleading
• If I wasn’t a developer, my funds would be stuck and I’d have no idea why

We’re building the next-gen financial infrastructure. This level of service kills trust.

To @CelerNetwork team:
• Improve support transparency
• Automatic Revert & Refund: If a cross-chain transaction fails, the bridge should automatically revert and return the funds to the sender (minus the gas fee). This is a basic safety feature — users’ funds should never be stuck in limbo, requiring them to be engineers just to recover their own money.

This shouldn’t require hours of back-and-forth to resolve.

If we want mass adoption, reliability + UX must improve across the entire stack.
Otherwise, non-technical users will simply give up the first time something breaks.

we need to raise the bar for the whole ecosystem.
Finance won’t go mainstream with “funds stuck, sorry” errors.

[David ⛓️ From Fed to Chain @fromfedtochain]

We just launched our From Fed to Chain iOS app! 📲✨

Bringing you insights from Wall Street to Web3, straight to your phone.

Check it out & let us know your thoughts!
https://t.co/z9yf5adBli

[David ⛓️ From Fed to Chain @fromfedtochain]

Message: Token exchange failed: <urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate (_ssl.c:1006)>.

Error code explanation: 500 - Server got itself in trouble.

Error logging in: login_with_chatgpt subprocess failed:

any ideas?