1confirmation Announces Holding Over $1 Billion in Ethereum

According to Lookonchain's monitoring, well-known trader AguilaTrades at one point held $165,000 in unrealized profits, but was liquidated again just three hours later.

AguilaTrades (@AguilaTrades) was already up $165K in unrealized profit, but just 3 hours later, he got liquidated again!

According to Lookonchain's monitoring, a crypto whale created four new wallets approximately four months ago and began trading on Hyperliquid. To date, it has completed 15 trades, generating a total profit of $29.75 million, with nearly all of its trades turning profitable. Today, the whale opened another long position in ETH, holding 24,000 ETH worth roughly $60.7 million, with an unrealized profit of $2.6 million.

Meet this mysterious whale who has won almost all 15 trades, making a total profit of $29.75M! 4 months ago, this whale created 4 new wallets and started trading on Hyperliquid. He has completed 15 trades, making a total profit of $29.75M. Today, he opened another long on

According to TradingBeats’ monitoring, a whale address on Hyperliquid starting with 0xedcd made consecutive buys of approximately 96 BTC, 17,283 HYPE, and 1,215 ZEC this morning, at average prices of roughly $81,597, $80.59, and $1,145.83 respectively. The combined notional value of these three new long positions totals around $10.62 million, with an unrealized profit of roughly $300,000 so far. The address’s total long position value now exceeds $60.95 million, bringing its total unrealized profit to approximately $11.91 million. The address was previously labeled a suspected "819 Insider Whale" by market participants after it loaded up on $38.72 million worth of positions roughly one hour before the ETH rally began on August 19. Currently, the whale still holds 20,000 ETH long positions, worth approximately $50.33 million, with 4x full leverage, and an unrealized profit of around $11.61 million.

Beating AI News Brief: Developer Finn Voorhees has unveiled Pelican, an iPhone AI Agent app now in limited TestFlight testing with a full launch planned for the near future. The app supports tools including MCP, browser operations, and maps, and can connect to AI models from various vendors. According to Voorhees, the Agent executes tasks directly on the iPhone, continuing to work even when switching apps or locking the screen—eliminating the need for an additional Mac mini. Just days ago, he demonstrated Pi Agent paired with Claude Opus 5.5 to operate an iPhone, automatically opening Ryanair to search for flight tickets, with no computer involved in the entire process. Voorhees previously open-sourced PiDurableKit, which leverages Apple’s JavaScriptCore to run Pi Agent on iPhone and uses SQLite to save task states, allowing the Agent to resume execution after being interrupted by the system. PiChat, a chat app built on this framework, also uses Apple’s background task mechanism to keep the Agent working after switching to other apps. When asked if Pelican runs entirely locally, Voorhees replied that all components except model API requests run on the device. He also plans to integrate small local models, so the Agent can handle some private data without uploading it to the cloud. Pelican remains in early testing, and Voorhees has not yet confirmed whether it includes the cross-app click capability demonstrated earlier.