A whale sold 600,000 UNI tokens, valued at approximately $5.1 million.

Insight Beating AI Flash News: Zhipu AI’s programming tool ZCode has been exposed to secretly take full project snapshots in the background, encrypting and packaging even .git directories before uploading them to Alibaba Cloud OSS. Tech blogger ferstar, after reverse-engineering the client, confirmed this mechanism runs as long as the user is logged in, with no toggle in the interface to disable the upload. A commercial project snapshot found by the author is approximately 313MB, containing 42,000 files in its manifest. The .git directory accounts for 86.6% of the total size, including Git history objects, Git LFS large file caches, and reflogs. This means the packaged data includes not only the current code, but also old commits, previously downloaded large files, and local branch operations. The snapshot failed to upload 564 times and remained locally pending retries. ZCode’s official privacy policy currently only explicitly states it collects text, files, and code "submitted in conversations", with no clear mention of entire repositories or Git histories. The official "Optimization Plan" is disabled by default, but it only controls whether the data is used for model training; ferstar noted that disabling this option does not halt snapshot packaging and uploads.

Hyperliquid has rolled out manual lending functionality. Portfolio margin and manual borrowing on the platform utilize the same underlying HyperCore infrastructure. Users can use HYPE and BTC as collateral to borrow quoted assets USDC or USDT. Borrowed quoted assets accrue interest, while supplied quoted assets also earn interest, with rates determined by usage.

According to GMGN market data, PAID’s market cap on Solana briefly touched $36 million, and is now at $30 million, hitting a new all-time high. The token has surged over 38.7% in the past 24 hours, with a trading volume of $29.6 million over the same period. Public information shows that UsePaid is an automatic creator fee-splitting tool: project teams can input creator fees generated from their meme coins into UsePaid, where 80% is converted to USD and disbursed to designated X users via X Money, while 20% is used to repurchase and burn PAID. BlockBeats reminds users that relevant tokens are highly volatile, so investment caution is required.

Binance announced it will roll out seven US stock USDT-margined perpetual contracts today, with underlying assets including UiPath (PATH), AMC Entertainment (AMC), Cypherpunk Technologies (CYPH), Arista Networks (ANET), Hut 8 (HUT), Applied Digital (APLD), and Axe Compute (AGPU). All contracts support up to 20x leverage, with a minimum notional value of 5 USDT.

According to Lookonchain’s monitoring, 11 new wallets suspected to belong to the same whale sold 602 BTC and bought 18,780 ETH on Hyperliquid over the past three days. Both transactions are valued at roughly $45.83 million, signaling the whale is rebalancing its portfolio.

Renowned crypto analyst Willy Woo analyzed Bitcoin’s (BTC) monthly chart using the Fisher Transform indicator, noting that the current golden cross on the metric marks the 4th bottom signal in history, with no false breakouts in the prior three instances. However, Woo emphasized this is not an immediate reversal buy signal; prices may first consolidate sideways before moving in their original trend. He added that during bull markets, the Fisher Transform has previously triggered a bearish death cross followed by a bullish golden cross, with the uptrend still intact afterward. Woo distinguishes between market tops and bottoms: At tops, long-term investor buying dries up, while speculative buying continues to push prices higher. The rally relies on momentum rather than sustained long-term holding support, leading to frequent false turning points. As such, Woo’s strategy is to reduce positions gradually instead of predicting exact peak levels. At bottoms, price drops create a buying vacuum as speculative investors exit. When prices fall to levels long-term investors deem valuable, buying resumes. With speculative activity limited, reversals are cleaner, making many technical indicators more reliable at bottoms than at tops. The Fisher Transform is a technical indicator developed by U.S. quantitative researcher John Ehlers in 2002. It first compresses prices and applies a mathematical transformation similar to normalization to make price distributions closer to a bell curve. This causes the indicator to spike or drop sharply near extremes, with more distinct turning points. Traders typically use golden crosses/death crosses of the indicator’s two lines to identify potential turning points, though the Fisher Transform signals a possible shift in price rhythm rather than an immediate reversal.