Binance will delist AAVE/ETH Leveraged Token Tradable Assets

Per PeckShield monitoring data, the KelpDAO attacker transferred 75,700 ETH to two new addresses on April 21.

April 21st – During the “Decoding Web 4.0: When AI Agents Take Over On-Chain Authority” roundtable, MuleRun CTO Steven Hu dove into the “Agent Moat” topic, noting the traditional idea of an AI tech moat is quickly fading. The core driver? Faster convergence of model capabilities and exponential gains in development efficiency. Hu pointed out the performance gap between mainstream large models is narrowing rapidly—especially over the past year, when the capability gap between domestic and global models has shrunk significantly. Meanwhile, exploding coding capabilities have supercharged software development: features that once took weeks or months to build now take days. This means Agent frameworks and specific functional modules can be quickly reused or replicated via open-source tools, making product-level “feature moats” increasingly fragile. In this landscape, Hu argues Agents’ future core competitiveness will hinge on two pillars: 1. **Continuous high-frequency iteration*

On April 21st, per HyperInsight monitoring (https://t.me/HyperInsight), a whale address starting with 0x616 liquidated its $4.06 million BTC short position via a stop-loss order, with a liquidation price of $76,218. The transaction incurred a minor loss of roughly $2,614. Subsequently, the address re-entered the market above the liquidation price at $76,670 and opened a new BTC short position of 52.6 BTC—valued at approximately $4.03 million—indicating an intent to reinitiate a short position. Address: 0x616f5d5db5cf8d98a8554c84492bce764c7f09b1

On April 21, at the roundtable discussion *Decoding Web 4.0: When AI Agents Take Over On-Chain Governance*, Box—Developer Relations for Monad Foundation Greater China—delved into the “Agent Moat” topic. He noted that the technical barrier to building AI agents is currently fairly low: the so-called “moat” for AI agents hinges not on the product itself, but rather on a founder or team’s understanding of AI, their ability to steer it, and their external influence. He pointed out that many agent products today are built quickly using open-source code and off-the-shelf tools, slashing development timelines. It’s not unusual to launch a product in just a few hours—but this also means the product form is easily replicated. When cloud providers or major model vendors step in, they can often replicate or even outpace these offerings in a short time, causing early-stage startups to quickly lose their competitive edge. Box further emphasized that in this landscape, the survival logic for a

Binance announced on April 21 that it will delist and cease trading for the following spot trading pairs at 03:00 UTC on April 24, 2026: 1INCH/BTC, WIF/BTC, and XRP/MXN.

April 21: At the "Decoding Web 4.0: When AI Agents Take Over On-Chain Authority" roundtable, Kelvin Tan, Associate VP at the Hong Kong University of Science and Technology (HKUST), discussed the "Agent Moat." He emphasized that different AI agents rely on distinct model training trajectories and technical systems, leading to significant gaps in real-world user experience. Recently, some new models and tools have delivered stronger generation quality, faster execution efficiency, and a higher ceiling for development output. However, Tan noted these differences haven’t yet formed a decisive lead—they lean more toward "efficiency gains" than "paradigm shifts." In short, AI agent competition remains in a fast-evolving stage, with no stable, insurmountable tech barriers on the horizon. Current AI agents and large models are evolving at a blistering pace: new products or features launch almost weekly, fueling ongoing industry progress. But from the lens of real-world use and busines