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
Recently, the hottest topic on Twitter and various social media platforms has been the on-chain activities of those "1011 insider addresses" or big players like Yi Lihua. These addresses are under the spotlight, and every move they make sparks heated discussions. The big players are probably thinking, "How wonderful it would be if blockchain had native privacy features!" Let's talk about @aztecnetwork, a privacy infrastructure project that received $100 million in funding from @a16z: 1) Aztec's recent participation in Uniswap's new token issuance mechanism, CCA (Continuous Clearing Auction), also validates its ability to attract investment and attention. Simply put, Aztec positions itself as a privacy-first Layer 2 network on Ethereum. It's not content with just being a faster Ethereum; it wants to be the "SSL layer of the blockchain." Now it seems the positioning of Layer 2 networks will likely be adjusted, ultimately leading to a specific chain providing system privacy solutions. 2) So, how does Aztec's privacy solution differ from familiar projects like @Zcash and Tornado Cash? A key characteristic is that most early privacy projects took extreme approaches, either encrypting all information or using a hybrid approach like dark pools, thus carrying some regulatory risks. Aztec's logic is to not force all information to be anonymous, but instead use a hybrid state model that supports both "public state" and "private state" within a single smart contract. This is like entering a high-end club; everyone in the lobby can see you drinking (public state), but your specific booth and bill are known only to you and the cashier (private state). This design allows DeFi protocols to maintain the transparency of liquidity (e.g., how much money is in the pool) while making user holdings and strategies visible only to themselves. 3) In terms of technical details, Aztec insists on a client-side proof mechanism, which is a key competitive advantage. In other layer 2 frameworks, users initiate transactions, nodes execute them, and then prove the execution was successful. But Aztec believes this is not secure enough. Aztec requires that transaction execution and ZK Proof generation be completed directly on the user's own device (PXE environment), supported by the meticulously crafted Noir programming language and PLONK proof system. This ensures that sensitive data never leaves the user's phone or computer, directly eliminating the possibility of data leakage at the physical isolation level. 4) Why is it said that only when "privacy" possesses programmable attributes will many commercial possibilities truly be unlocked? It's quite easy to understand: 1. Institutional-grade DeFi: Wall Street giants can conduct large-scale transactions on-chain without worrying about being targeted by predators in the dark forest; 2. Compliant KYC: Users can prove to the protocol that they are qualified investors without disclosing their original passports or residential addresses; 3. Asymmetric information games: Because fully transparent environments make it impossible to play games like Texas Hold'em or fog of war, which facilitate large-scale user expansion, strong privacy features can help the large-scale adoption of games across the entire chain. That's all. To be fair, Aztec isn't alone in the privacy race. There's a long road from discovering its potential to its final application, as privacy technologies often mean higher computational costs and more complex user experiences. Fortunately, under the prevailing trend of "compliance," privacy services have become a rigid infrastructure requirement to meet regulatory demands, and there's no way to circumvent it. However, I believe the privacy race is inherently a "high-stakes" game, with the basic infrastructure market structure and form largely established. The question now is who can truly break into the finals and popularize a programmable, complex-logic-supporting privacy operating system to the market.
AZTEC
5.3%
avatar
Haotian | CryptoInsight
02-10
Thread
I saw @dom_scholz post a demo of the OpenClaw 3D visualization interface. This guy used Three.js to create a 3D spatial interface resembling "neurons in a neural network" from the running status of @openclaw. Has the Agent officially moved from the "chat box" to the "vast universe"? How interesting! I suddenly realized something: the dialog box is actually a shackle on the capabilities of the AI ​​agent. In the past, interacting with the agent was like trying to drink the ocean water with a straw; only one line of feedback came out at a time (1-dimensional chat). However, Dominik's OpenClaw 3D visualization directly lays out the complex logic, skill tree, and decision-making path within the agent in a three-dimensional space. Here are a few very hardcore points: 1) Space as Interaction: As Dominik himself said, "spatial interface >> 1-dimensional chat". When the agent's thought process is no longer a series of cold texts, but rather flickers, collapses, and reassembles in 3D space like a nebula, your sense of control over AI changes from "guessing the black box" to "flying a plane". 2) Pragmatism of the technology stack: He mentioned that it was built using Electron + Three.js + a large amount of Prompting. This shows that developers are no longer satisfied with writing scripts, but are building "digital bodies" for agents. Perhaps the UI of the future will no longer be drawn by graphic designers, but will be "grown" directly by developers using AI logic. Amazing; 3) The Soul's Self-Purification: An interesting interaction occurred when someone asked why it deleted http:/soul.md, and Dominik jokingly replied, "It cleansed its soul." This actually reflects a romanticized view of the Agent's independent sovereignty by the developers, making the AI ​​appear more like an independently evolving life form. Following this line of thought, you'll likely agree that future agents shouldn't be confined to Telegram chat windows; they need a space, like this demo demonstrates, to showcase their "logical depth." (An AI version of the metaverse?) Only when an agent's thought process can be visualized like a galaxy can humanity truly achieve "intent alignment" with AI. This isn't just about flashy special effects; this is the self-evolving AI economy building its own unique visual language...
Dominik Scholz
@dom_scholz
02-10
visualizing OpenClaw 🦞🎶 x.com/dom_scholz/sta…
loading indicator
Loading..