10 #Bitcoin ETFs inflows 2,586 $BTC(+$247.92M) and 9 #Ethereum ETFs outflows 20,884 $ETH(-$70.69M) yesterday.

NVIDIA (NVDA.O) announced today that its board of directors has approved an additional $150 billion in authorization for its existing stock repurchase program, bringing the total remaining authorized repurchase amount to $235 billion. The company expects to complete the remaining repurchase plan by the end of fiscal 2028. "NVIDIA's growth stems from the transition to the next-generation platform of artificial intelligence and accelerated computing," said NVIDIA CEO Jensen Huang. "Our cash-generating ability enables us to invest in technologies driving this transition and return capital to shareholders. This authorization reflects our confidence in long-term future opportunities." According to market data from BIT (bit.com), NVIDIA's US-listed stock rose 1.4% in pre-market trading.

Beating AI News (from Dongcha) reports that Synthetic Sciences, a Y Combinator Winter 2026 (YC W26) startup, has officially launched its open-source research agent OpenScience. The tool can search academic papers, process data, write code, and access research databases; its newly added Autoresearch feature allows AI to run experiments continuously on its own. For instance, when training a model, researchers only need to instruct it to “reduce the validation set loss”. OpenScience will first run a baseline, then propose modifications, and execute experiments iteratively. It retains improved results, rolls back those that worsen, and determines next steps based on prior outcomes. Users can also pre-set limits on run count, total duration, and stop conditions, or specify constraints like “only modify optimizers moving forward”. OpenScience automates the experimental iteration workflow that previously required researchers to monitor closely: proposing plans, running experiments, comparing metrics, discarding failed approaches, and proceeding to the next round. Experiments can run locally, or be deployed to remote GPUs, SSH servers, and Slurm/PBS clusters. Every round’s code, results, and judgments are logged for later review. The agent supports direct login to ChatGPT/Codex subscriptions, as well as integration with users’ own API keys, local models, or the official pay-as-you-go Ace. In official self-tests, OpenScience completed 53 out of 70 Terminal-Bench Science tasks, scoring 75.7%. By comparison, Codex + GPT-6 Astra posted a public score of 68.1%.

The Securities and Futures Commission (SFC) of Hong Kong and the Accounting and Financial Reporting Council (AFRC) signed a new memorandum of understanding (MoU) today, expanding their regulatory cooperation scope from financial reporting and related audits of listed entities to financial and compliance reporting of licensed corporations, SFC-licensed virtual asset service providers, SFC-authorized funds, and registered open-ended fund companies, as well as audit and assurance work conducted by relevant auditors. Under the new MoU, the two parties will establish a comprehensive cooperation framework covering information sharing, case referral, mutual assistance, and coordinated inspections and investigations. The SFC stated that the cooperation aims to further enhance regulatory coverage, strengthen the quality of financial reporting and market integrity, and boost global investors’ confidence in Hong Kong’s financial regulatory system.

As $NEAR drops, trader 0xec90 just had 748,091 $NEAR ($3.67M) liquidated!

After Chinese cross-border brokers including Futu, Tiger Brokers, Longbridge, and Huasheng Securities, some Hong Kong-based Chinese securities firms have also started clearing their existing mainland China operations. To comply with relevant regulatory requirements, Guotai Junan International has implemented partial adjustments to services for its existing mainland investor accounts starting from the 26th of this month. According to customer service staff at Guotai Junan International, these adjustments are similar to measures previously taken by brokers such as Futu Holdings and Tiger International: when existing mainland investor accounts log in from mainland China (with login IPs showing as domestic), they will be unable to deposit funds or place buy orders, only allowing fund withdrawals or sell transactions. If an account’s login IP is located outside mainland China, its fund transfer and trading functions remain unaffected. If a customer obtains a new identity, they must update their corresponding document information; as holders of overseas documents, such customers will not be subject to this rule.

Beating AI Insight News Brief: Developer David has open-sourced Jevgrep, a research tool built on Jev specifically for coding agents to locate code. Users only need to ask a question like "Where is the login validation implemented?" and it uses TypeSafe’s Jev decision model to search the codebase layer by layer, identifying relevant files and source code snippets before passing them to agents such as Claude Code and Codex for further modification and testing. It primarily addresses the token-intensive "code lookup" step that coding agents face. Jevgrep does not feed the entire repository to the model upfront, nor does it rely solely on a single semantic search. It first determines which directories are worth exploring further, then checks relevant files and code declarations, and finally returns source code snippets, line numbers, and follow-up reading references. The repository also includes a "Skill" that enables agents to know when to call `jg` to gather context. The latest SWE-bench experiment used 10 Python tasks. Both with and without Jevgrep, 8 tasks were completed, but the total cost for GPT-5.6 Sol dropped from $7.62 to $5.44, a reduction of 28.63%. The author initially stated a 40% reduction on X, but later updated the experiment results in the repository, revising it to "around 30%". However, this figure only accounts for Sol’s costs and does not include Jev’s expenses. The confirmed Jev call costs in the experiment logs amount to at least $1.57, while some calls lack complete billing records, so the actual total is unknown. Additionally, the experiment only included 10 tasks, with each run only once.