In the past hour, total liquidations across the network reached $13.27 million, with $11.88 million from long positions, with $1.39 million from short positions. BTC liquidation $616,484.64, ETH liquidation $486,010.05.

According to an official announcement, Bitget has adjusted the leverage, position tiers, and maintenance margin ratio for the SPCXUSDT U-margined perpetual contract trading pair. Leverage has been increased to a maximum of 100x, depending on the position size. Existing positions held before the update will be affected; users can adjust the relevant parameters according to their asset allocation needs. More details can be found on the Bitget official platform.

According to Odaily Planet Daily, BitcoinTreasuries.NET published an article on the X platform stating that listed Bitcoin mining company Hyperscale Data GPUS increased its holdings by 4.62 BTC, bringing its total holdings to 713.58 BTC, ranking 56th in the Bitcoin 100.

According to Gate data, WTI crude oil fell below $75 per barrel for the first time since March 4, down 2.22% on the day.

According to a recent study by Anthropic, based on 400,000 sessions, understanding business needs is more important than coding skills, and the success rate of non-programmers has reached parity with software engineers. The study shows that users of Claude Code are active for an average of 20 hours per week. The success rate of non-programmers using AI agents to write code has narrowed the gap with professional software engineers to less than 7%, and the success rate for management roles has slightly surpassed that of professional programmers. Large-scale models are reshaping the human-machine division of labor, with users leading approximately 70% of decisions and AI agents handling approximately 80% of execution tasks. Expert-level sessions can trigger 12 actions and 3200 words of output, more than five times that of beginners. Beginners have a 19% abandonment rate when encountering errors, while intermediate and expert users only have a 5% to 7% abandonment rate. The study demonstrates that intermediate industry knowledge in specific domains can overcome the programming barrier through AI agents, achieving a leap in personal productivity. Developers' use of AI agents is shifting from debugging code to autonomous creation. Within seven months, the proportion of debugging sessions used to fix code halved from 33% to 19%, while the proportion of direct deployment and data analysis doubled, and the average estimated value of tasks increased by approximately 25%.