Zcash will freeze around 12,200 ZEC from its Sprout privacy pool after the NU7 upgrade, valued at approximately $16 million.

Beating AI Express News: Meta has acquired AI coding tool Conductor. Founder Charlie Holtz announced the team will join Meta’s Superintelligence Lab (MSL) to develop new products alongside Nat Friedman and Alexandr Wang. Conductor is a tool for managing multiple coding agents simultaneously, supporting Claude Code, Codex, Cursor, and OpenCode. It assigns tasks to different agents, enabling them to work in parallel on separate code branches. Developers can track progress, review code changes, and merge results all within a single interface. Launched in July 2025, Conductor secured a $22 million Series A funding round just this past March. Its official website notes it has over 100,000 active developers. However, the acquisition means the original Conductor service is being phased out: new user registrations and downloads will stop on November 4. The cloud version will switch to read-only access on November 13, fully shut down on November 20, and cloud user data will be deleted at that time. The local version will remain usable temporarily, but will also be phased out starting in early 2027. Holtz added that the team plans to launch a new developer tool interface early next year. Some community users have expressed disappointment over the shutdown, with many already seeking alternative tools, while others are calling for Conductor to be open-sourced.

HSBC has cut Arm Holdings' price target from $230 to $215, while raising Visa's price target from $375 to $465.

US President Donald Trump stated that if Democrats take control of Congress next year and initiate impeachment proceedings against him, it will "destroy" the Democratic Party.

According to official announcements from Bitget, the platform has announced the launch of its Open Custody Network (OCN). Centered on the core philosophy of "Assets in Bank, Tradings on Bitget", OCN is designed to connect professional financial institutions including banks and custodial firms with Bitget’s trading infrastructure. The network integrates core capabilities such as independent asset custody, bank and institutional-grade security, exchange-integrated trading services, and capital efficiency, breaking the traditional model where centralized exchanges combine asset custody and trading on a single platform. Through an asset mapping mechanism, OCN bridges independent asset custody and trading services, allowing all eligible users to retain full independent custody rights over their assets while participating in Bitget’s spot, margin, and futures trading, as well as accessing exchange-level liquidity and integrated trading services. This further lowers barriers to trading participation, enabling more users to access financial services in a more equitable way. Bitget CEO Gracy Chen noted that OCN aims to break the traditional paradigm where users must custody assets on exchanges to use their trading services, letting users choose their own asset custody methods while benefiting from high-quality liquidity and efficient capital utilization, making financial services more open and inclusive.

Beating AI News Flash: Ricursive Intelligence, an AI company founded by former Google DeepMind researchers, has released the open-source model Ripple-65B Preview. The two founders previously led the AlphaChip project, which uses AI to automatically optimize Google TPU chip layouts. Ripple is a 65-billion-parameter dense model trained from scratch, utilizing 12 trillion tokens of training data and supporting a maximum context window of 1 million tokens. In the official Terminal-Bench 2.1 programming test, it scored 80.5%, slightly outperforming the 501-billion-parameter Reflection Beam (80.1%) and surpassing the 975-billion-parameter Inkling (63.8%). Ricursive stated that the entire model development cost only $32 million, with the team expanding from 1 to 5 members during development. The company improved training efficiency by refining model architecture, training data, and reinforcement learning methods. In control experiments with 4-billion-parameter models, the computing power required to achieve comparable code capability was reduced by approximately 16 times. Ripple’s model weights are now open-sourced under the Apache 2.0 license; after quantization, it is expected to occupy around 36–40GB, and can theoretically run locally on Mac devices with 64GB or more of memory.

CryptoBoost founder stated in a post that the current Bitcoin (BTC) bear market began on October 6, 2025, from its all-time high. As of publication, the bear market has lasted 396 days, slightly longer than the average 391 days it takes for a bear market to bottom out since 2014. Historically, BTC’s bear market drawdowns have been steeper: the 2014 bear saw a maximum drop of around 86%, 2018 around 84%, 2022 around 77%, while the current pullback is approximately 53% — relatively mild. The growing participation of institutional players including spot ETFs, corporations, and funds may strengthen the market’s holding structure, thus mitigating downside risk. BTC whales have been steadily accumulating since July. That said, a cycle hitting the historical average length does not signal the bear market is over; BTC must reclaim and hold above the $80,000–$82,000 range to further confirm a market trend shift. Against a backdrop of recent market moves — declines in U.S. tech and semiconductor sectors, capital rotating into bank stocks, the Dow Jones rising after the U.S. midterm elections, and falling oil prices — he noted Wall Street’s capital flows remain unclear, suggesting BTC may be nearing the tail end of its bear market cycle.