Hyperliquid's largest gold long position holder has unrealized profits exceeding $6.35 million, while holding a flash swap short position.

Beating AI Express: OpenAI begins limited resumption of the $200/month ChatGPT Pro 20x, but has not yet reopened the plan to all users. The official new rule states only users who recently used Pro 20x get one chance to resubscribe. Regular new users and those upgrading from other plans remain suspended. Specifically, users who held Pro 20x as of September 10, or whose subscription ended in the prior 30 days, may be eligible. For those who canceled, failed to renew, or downgraded, the re-subscription window is capped at 30 days and can only be used once. The core differentiator of Pro 20x remains its usage quota, which is 20 times that of ChatGPT Plus. OpenAI suspended the $200 tier on September 10 due to surging demand for Astra. Tibo Sottiaux, who leads ChatGPT and Codex, called Astra’s demand “unprecedented” at the time, noting that Pro 20x exerts the most strain on the system.

Beating AI Insight News Brief: OpenAI has added Chrome extension support to ChatGPT’s desktop built-in browser. Users can now directly install, pin, and use extensions, with the official demo using 1Password. Enterprise administrators can also centrally deploy and manage extensions. This built-in browser operates as a separate environment from users’ own Chrome, and does not automatically inherit Chrome’s login status, cookies, or installed extensions—separate configuration is required within ChatGPT. OpenAI’s documentation confirms it now supports standard browser functions including login, multiple tabs, downloads, and extensions, though security boundaries remain restricted. OpenAI engineer James Sun noted that agents cannot currently directly operate extensions or access data within them. For instance, while humans can use 1Password to fill passwords, ChatGPT cannot open the extension to retrieve passwords on its own. This is part of the ongoing integration of the Atlas feature into ChatGPT. When OpenAI announced the shutdown of Atlas in July this year, it explicitly stated it would transfer browser agent capabilities to ChatGPT and Codex. With Chrome extension support now added, ChatGPT’s desktop version is increasingly resembling a full-fledged browser.

Beating AI Express, a news brief from 动察 – According to a Financial Times report citing recent OpenAI investor presentation materials, OpenAI projects cumulative negative free cash flow of $278 billion between 2026 and 2030. Its revenue will grow from $36 billion this year to $350 billion by 2030, with cumulative revenue over the five-year period reaching an estimated $840 billion. OpenAI plans to invest roughly $856 billion in computing power and infrastructure by 2030, which will be the company’s single largest expenditure. The firm just completed a $122 billion funding round in March, but at its current spending trajectory, it is expected to deplete this capital by 2028. The report notes that OpenAI is seeking a new funding round at a valuation above $1.2 trillion to support its expanding computing power requirements. Meanwhile, the company is competing with rivals like Anthropic for market share and has implemented price cuts to counter competitive pressure from low-cost open-source models in China.

According to EmberCN’s monitoring, trader Boomer, after taking profit on 14,300 ZEC for a total gain of $5.237 million, added 10,000 ETH long positions at an average price of $2,610. His total profit on Hyperliquid now stands at $22.28 million.

Polygon co-founder Sandeep Nailwal announced in a post that the Polygon network is now prepared to permanently burn 100 million POL tokens. The relevant contract has been deployed on the testnet, and will launch on the mainnet once the Security Council completes its final signature. The 100 million POL will be permanently removed from circulation, with the community able to trigger quarterly POL burns moving forward. Polygon has currently accumulated 12.1 million POL in base fees. Nailwal noted that POL has been in a deflationary state since January 2026, network activity remains on the rise, and the network has scaled to 5,000 TPS.

Beating AI News Flash: Anthropic has officially launched "embedded evaluation". Faculty, Accenture’s AI subsidiary, will deploy personnel to work inside Anthropic to conduct model evaluations, red teaming exercises, alignment assessments, and safety checks. This differs significantly from standard external evaluations. The evaluators will receive near-employee-level access, enabling them to review model training processes, understand development and deployment decisions, and communicate directly with internal staff. Over the next five years, both parties will invest at least $10 billion each to build this evaluation capability. Anthropic also acknowledged that the industry currently lacks unified rules, and details such as what evaluators can access, how results will be disclosed, and who will cover costs are still being finalized. For now, Accenture’s evaluation work is directly funded by Anthropic. However, Accenture is a long-term strategic partner of Anthropic, having established a dedicated Anthropic Business Group that trains approximately 30,000 people to use Claude. One of the minimum requirements proposed by the AI Evaluator Forum on the same day is precisely that evaluation institutions should not have significant commercial ties with the companies they assess. Anthropic plans to introduce additional evaluation institutions in the future and is also in talks with non-profit organizations including METR (a long-time partner of Anthropic) for pilot programs.