Justin Sun: Li Lin Once Concealed Due Diligence Materials, I Lent the Company $30 Million to Fill the Gap

Beating AI News reports: Microsoft AI has launched MAI-Transcribe-2, now available in public preview on Microsoft Foundry. The new model supports 60 languages, adding speaker diarization, word-level timestamps, and domain-specific term biasing. Developers can pre-load names, drug terms, or industry jargon to reduce mishearing of rare words. The most notable improvement is speed. In non-real-time transcription tests by Artificial Analysis, MAI-Transcribe-2 achieved approximately 410.7x real-time speed, meaning a 1-hour audio recording can theoretically be processed in around 9 seconds. GPT-Transcribe clocked in at around 40x, making MAI-Transcribe-2 roughly 10 times faster; Gemini 3.5 Transcribe hit about 89.8x. Accuracy is also among the top tier. In FLEURS’ 60-language test, it posted an average Word Error Rate (WER, lower is better) of 5.2%, ranking first; in another non-real-time English test by Artificial Analysis, it had a WER of 2.0%, placing second, only trailing Alibaba’s Fun-Realtime-ASR-preview at 1.7%. Pricing is currently set at $0.10 per hour, down from $0.36 for the previous MAI-Transcribe-1.5. However, this is a limited-time offer valid only through December 31, 2026; the official full price has not been announced yet.

According to official announcements, OKX’s AI Builder co-creation initiative has officially launched and is now open for applications. Developers can connect to OKX’s market data, trading, and account functionalities via methods including MCP, CLI, or API to integrate trading features into their own AI products. After users complete transactions through the Builder project, Builders will split trading fee revenue with OKX, with a rebate rate of up to 50% — transactions from both new and existing users are eligible for rebates.

According to Lookonchain’s monitoring, a newly created wallet just one day old spent 752,723 USDC to purchase 10.2 million MEME at an average price of $0.074 two hours ago, currently holding an unrealized profit of around $388,000.

Grayscale Zcash ETF (ZCSH) has attracted a cumulative net inflow of approximately $34.4 million since its listing on NYSE Arca on August 25. Launched by Grayscale, ZCSH is the first exchange-traded product offering exposure to ZEC spot. According to HTX market data, ZEC broke through $1,000 today, hitting a new all-time high; ZEC has risen by around 31% cumulatively since the Zcash ETF's listing.

This week’s top 5 most-watched stocks among users are MSTR, CRCL, DELL, AMC, and FAMI, with market focus covering sectors including Bitcoin (BTC) treasuries, stablecoins, AI servers, and cinemas. Among them, Strategy resumed its Bitcoin accumulation after two months, with MSTR and CRCL both rising over 8% this week. Dell gained around 13.24% driven by strong AI server orders and growing backlog demand; in contrast, AMC trended lower amid swings between box office recovery expectations and high short positions, while FAMI surged on heavy volume before retreating, with sharp short-term volatility. MSTR (up ~8.94% this week): Strategy disclosed it purchased 4,603 BTC for ~$369.7 million this week, marking its second Bitcoin buy in two months. Following Bitcoin’s rebound on rate cut expectations, MSTR, a proxy for Bitcoin leverage, regained investor interest. CRCL (up ~8.04% this week): As Federal Reserve officials revisited digital asset and stablecoin regulatory frameworks, Circle rallied alongside broader crypto stocks. CRCL’s sharp rise was supported by USDC supply growth and policy expectations. DELL (up ~13.24% this week): Dell’s financial report showed strong AI server orders and backlog demand, prompting the company to raise its full-year outlook. The data center hardware sector was revalued, driving two consecutive days of sharp gains in Dell’s stock. AMC (down ~3.79% this week): Cinema stocks swung between box office recovery hopes and high short positions, leading to volatile declines. AMC saw active trading this week but lacked strong catalysts, resulting in a downward trend. FAMI (up ~0.41% this week): Farmmi saw abnormal trading volume this week, surging at one point on September 2 before quickly retreating. Low-priced Chinese concept stocks were driven by liquidity, leading to significantly amplified short-term volatility.

News Flash from Dong察 Beating AI: Ant Group has officially open-sourced its financial model Ling-3.0-flash-Fin. The model boasts 124 billion total parameters, 5.1 billion active parameters, and supports a 256K context window. The officially released BF16 weights are approximately 255GB, licensed under MIT, and can be self-deployed using vLLM and SGLang. Alongside the model, the financial search benchmark FinFIRST has also been open-sourced. Developed by Ant Group with professional support from CICC’s Investment Banking team, over 50 financial professionals participated in question design and validation. Only 9.78% of candidate questions passed screening, leaving 123 questions total: 61.8% require calculations, and 32.5% demand cross-referencing multiple sources. Each question is broken into multiple scoring dimensions, assessing whether the AI correctly identifies data, uses appropriate sources, verifies timelines and statistical standards, and delivers accurate calculation results. The entire dataset includes 701 granular scoring points. All questions, reference answers, and scoring rubrics are now fully open, licensed under Apache 2.0.