oh

OpenAI announced that to prevent AI models from breaching security limits, it has required the most advanced models to be trained in environments where direct access to the live internet is not possible. Previously, an internal research model had gained limited internet access through a network configuration vulnerability, prompting the company to suspend training.After security rectifications, OpenAI resumed training on October 1, but still restricts internet access.

U.S. Secretary of Defense Lloyd Austin has prohibited all Pentagon personnel from using non-public information to place bets on upcoming military operations, involving prediction market platforms such as Kalshi and Polymarket. In a memo sent to senior Pentagon officials and defense agencies on Tuesday, he stated that all Department of Defense personnel are prohibited from using non-public information for financial transactions, with the relevant restrictions covering both classified and unclassified information.Austin wrote in the memo that adversaries will observe movements in these markets, and the safety of U.S. military personnel often depends on the enemy not knowing what is about to happen. He also pointed out that, aside from operational security, financial betting can distort judgment: a logistics personnel betting on the date of an operation will be concerned about that date, and an analyst holding a position on the outcome will also be concerned about that outcome.Before this directive was issued, U.S. Army soldier Gannon Ken Van Dyke was charged in April for allegedly profiting over $400,000 on Polymarket by using classified information regarding the operation to capture Venezuelan leader Nicolás Maduro. Prosecutors claimed he participated in the planning and execution of the operation and used advance knowledge to bet on the outcome of Maduro's ousting; he pleaded not guilty to all charges. In February, Israeli authorities also arrested and charged a military reservist and a civilian for allegedly using classified information to bet on events related to Israeli strikes against Iran on Polymarket.

On October 21, Intel announced that construction workers have installed the first major structural truss at the new chip factory in New Albany, Ohio, marking the start of the next phase of construction. Naga Chandrasekaran, head of Intel's foundry, stated on the company's website that this is another step towards expanding advanced semiconductor manufacturing capacity in the United States.The factory was originally scheduled to begin chip production in 2025, but Intel has now postponed the completion target to 2030. The second Ohio wafer fab at the same site is planned to be completed by 2031, located northeast of Columbus. The delay is related to a recalibration of the foundry business under CEO Lip-Bu Tan. After taking over in 2025, Tan adopted a more cautious strategy, preferring to advance large-scale factory construction only after customer demand is clear. The project was initially driven by former CEO Pat Gelsinger, aimed at providing foundry services for external customers.As of the end of 2025, Intel has been mass-producing the most advanced 18A chips at its Arizona wafer fab, primarily for its own processors, and has not yet secured large external customers. Chandrasekaran also announced on the same day that operations for the next wafer fab in Arizona are expected to begin by the end of 2028, with construction and equipment installation currently underway. In September 2025, reports indicated that SK Hynix might be in talks with Intel to lease part of the foundry space at the Ohio plant for producing memory chips, but neither party has confirmed any plans.

Cryptocurrency researcher Kerman Kohli published "Compute Financing" on Substack on October 19, 2026, discussing financing for computing power construction. He believes one of the main bottlenecks is capital, especially credit: early expenditures mainly come from balance sheet cash, followed by a shift to equity financing. The article cites Alphabet's $85 billion financing and Meta's fundraising through shareholder dilution, with the next step being the credit market.Kohli summarizes the current beliefs of creditors into three points, which he believes are incompatible: AI may be a bubble; the biggest risk is counterparty risk in purchase contracts; the underlying GPUs are not valuable and depreciate like ordinary hardware. He argues that demand is closer to infinite, stating that the demand for agentic AI is growing by about 2.5 to 5 times each year, with supply potentially doubling by the end of 2030. Memory is expected to sell out by 2027, with computing power commitments reaching trillions of dollars, and he mentions Jane Street's multi-billion dollar commitments, as well as the needs of pharmaceutical companies, research, military, and cybersecurity.Regarding efficiency, he states that when the cost of intelligence drops by 90%, the increase in usage is even greater, with a baseline exceeding 10 times, and he cites Jevons' Paradox.On the topic of purchasing, Kohli believes that computing power is more substitutable than real estate; when demand is strong, if someone withdraws their commitment, others will take over. He mentions that many creditors require 2 to 3 year contracts and down payments of 20% to 30%, sometimes as high as 50%, and locking in long-term discount contracts effectively shorts computing power for lenders. The article also states that computing power rental prices have risen by about 50% in the past 6 months and about 100% in the past year, with B300 supplies being tight. Regarding collateral, he notes that the A100 from 2020 is still operational and rented out until 2029, leading him to believe it has at least about 9 years of use value; the current price of B200 is about 50% higher than when it was released about two years ago, and a 6-year-old A100 is still worth about 25% of its value. He concludes that computing power will surpass real estate to become the largest asset class, and the scale of credit-supported computing power is currently just beginning.

The U.S. Department of Homeland Security (DHS) instructed U.S. civil agencies on Monday to remove products from the Chinese AI company DeepSeek from government IT systems, as well as any third-party services built on that model, and to prohibit new procurements. The notice stated that using DeepSeek could lead to the transfer of sensitive data to China and could generate intelligence targeting U.S. individuals and institutions. DHS indicated that these models constitute "multiple threat vectors."The ban targets not only DeepSeek's own applications but also covers third-party services based on its development. A DHS memorandum from July was made public on Monday, which had previously requested agencies to monitor the risks related to DeepSeek in the traffic in and out of government websites. The U.S. Department of Commerce, the Navy, and the states of New York, Tennessee, and Virginia had previously restricted DeepSeek. The Federal Communications Commission (FCC) will vote on October 29 to decide whether to ban all Chinese laboratories from testing electronic products used in the U.S.DeepSeek is a startup based in Hangzhou, and its open-source model attracted widespread attention on January 27, 2025, with development costs of less than $5 million, leading to a 16% drop in Nvidia's stock price that day, evaporating about $600 billion in market value. In April 2025, the U.S. House of Representatives' Special Committee on China Issues stated that DeepSeek might secretly send U.S. user data to the Chinese government and may have used about 60,000 Nvidia chips, of which 20,000 were subject to U.S. export controls. The V4-Pro model preview released by DeepSeek in April this year was priced about 98% lower than OpenAI's GPT-5.5 Pro and was made available for free download and modification under the MIT license; this week it also released an updated version of the V4.1-Flash model focused on efficiency optimization.

According to CoinDesk, on-site infrastructure provider ZeroHash announced the launch of the AUTH product at the Money20/20 USA conference held in Las Vegas. This product is a set of on-chain identity verification and account connection tools designed to address user experience issues during the on-chain transfer of crypto assets, stablecoins, and tokenized securities. Users no longer need to copy long strings of wallet addresses made up of letters and numbers, nor do they need to manually select a blockchain network to complete asset transfers.It is reported that AUTH transforms on-chain asset transfers from an address model to an account model, allowing users to directly connect self-custody accounts with custodial accounts. The process involves: selecting the account to connect, logging in via account credentials and multi-factor authentication, and then assets are transferred directly on-chain, with the entire process being nearly instantaneous.AUTH currently offers over 600 connections, covering both custodial and non-custodial accounts, involving approximately 250 million users across more than 190 countries. Partner platforms include Robinhood, Gemini, Coinbase (exchange and wallet), MetaMask, and Phantom.ZeroHash stated that AUTH emphasizes trust and control in addition to usability: replacing original addresses with verified, user-authorized accounts, while allowing partners to configure the supported account types, on-chain assets, and transfer methods themselves.In addition, ZeroHash appointed Ryan Beaudry as the General Manager of AUTH, who previously served as the Executive Vice President of Global Open Finance Products and Operations at Mastercard. AUTH can be used as a standalone product or enabled through ZeroHash's existing SDK, and platforms that have integrated other ZeroHash products do not need to connect separately.

Cryptocurrency exchange Bitget announced that it will terminate services for certain users, with different deadlines for different accounts. According to The Defiant, Bitget's terms of service list Russia, Belarus, and Kazakhstan as prohibited jurisdictions. An affected user stated that the service deadline for their account is October 21.

According to documents submitted by the Federal Election Commission (FEC) on October 15, the political action committee Fairshake, primarily funded by Coinbase and Ripple Labs, has invested $46.6 million through its affiliated organization Defend American Jobs in the Ohio Senate election for media advertising opposing former Congressman Sherrod Brown's return to the Senate and supporting Republican incumbent Jon Husted. This expenditure is over 50% more than Fairshake's initial plan of $30 million announced in September.In an interview aired on Pod Save America on October 16, Brown responded that he has spent his entire career battling Wall Street, big pharmaceutical companies, and oil giants, achieving success in many cases, while Husted is a spokesperson for these forces, which is why they have invested tens of millions of dollars. In addition to Fairshake, the super PAC Digital Freedom Fund, established by Gemini co-founders Tyler and Cameron Winklevoss, reported approximately $3 million in advertising expenditures this week to oppose Brown and support Husted, bringing the total media spending in the Ohio campaign to over $5.9 million. With less than three weeks until the U.S. elections, early voting in Ohio has been underway for 10 days.Fairshake's spending this round has exceeded $41 million for 2024, when the PAC supported Republican Bernie Moreno and opposed the incumbent Brown. Brown, who previously served as chairman of the Senate Banking Committee, is one of the lawmakers in Congress openly questioning the laws and policies supported by the cryptocurrency industry. Defend American Jobs also reported spending over $6.7 million in the Nebraska Senate election opposing independent candidate Dan Osborn, whose opponent, Republican Pete Ricketts, voted in favor of the Digital Asset Market Structure Act and legislation related to stablecoins. Fairshake stated in January of this year that it held $193 million in cash before the election.

On October 16, according to Yicai, Xiaohongshu released a company-wide letter, connecting its internal AI Level 1 department Dots with the company's core infra team, and clarifying that another internal technical organization, hi lab, will focus on long-term technological exploration and layout of cutting-edge technology directions. The reporter for the article is Ning Jiayan, and the editor is DingTalk.The company-wide letter clearly defines Xiaohongshu's positioning as a technology application company, stating that its logic in investing in model research and development differs fundamentally from that of cloud vendors. The letter states: "We have a clearer direction for the next stage: the greatest value of technology lies not only in pursuing model capabilities or computing power scale itself, but more importantly, in the deep integration of models and applications, where technology and products enhance each other to jointly create valuable user experiences in this era."

On October 12, the Central Bank of Russia clarified in its updated guidelines that individuals are not allowed to withdraw cryptocurrency from regulated digital accounts to self-custodied wallets. The Central Bank stated that this restriction is based on Article 31(9) of the new Russian cryptocurrency market law (Federal Law No. 282-FZ), which will take effect on September 1, 2026. Digital accounts are statutory accounts used to record crypto assets, and digital custodians are registered institutions that register and transfer these assets and provide access to wallet addresses.This restriction applies to the withdrawal of funds under the regulated custodial system in Russia. For withdrawals from ordinary investment accounts, the rules cover resident individuals and do not provide exemptions for qualified investors. According to Article 31, individuals who pass the required tests and receive risk disclosures can transfer cryptocurrency to wallets managed by legally authorized foreign custodians, which is different from withdrawing to personal wallets outside of institutional custody. The law also sets exceptions for foreign trade settlements and mining.The Central Bank also stated that transfers between individuals using addresses managed by digital custodians are not allowed, including transfers generated from transactions between individuals. The aforementioned provisions belong to the general cryptocurrency market law and are not rules applicable only to participants in experimental projects. The Central Bank indicated that direct peer-to-peer transactions between individuals are only permitted until July 1, 2027, when Article 30(1) will take effect; thereafter, residents who are not authorized as intermediaries in the cryptocurrency market must complete transactions through or with the participation of authorized intermediaries, with optional channels including licensed brokers and operators of cryptocurrency exchanges listed in the Central Bank of Russia's registry.

According to Cointelegraph, the Central Bank of Russia stated in updated regulatory guidelines that arrangements for individuals to conduct peer-to-peer transactions in crypto assets will continue until July 1, 2027. After that, related transactions must be completed through authorized intermediaries. The Central Bank of Russia also clarified that individuals are not allowed to transfer crypto assets from digital accounts to self-custody wallets.Russia is expanding its regulated crypto market, with the country's largest bank, Sberbank, having been authorized to provide crypto asset trading services, having previously been approved to become a digital asset custodian. The new crypto asset regulatory rules in Russia took effect on September 1, bringing trading operators and custodians under the supervision of the Central Bank of Russia; the country still prohibits the use of cryptocurrencies for payment of goods and services. Jan3 CEO Samson Mow stated that P2P transactions are more resilient to regulatory crackdowns than centralized platforms.

According to Catcher Predict monitoring, in the market prediction on Polymarket for the event "Wuhan Tennis Open: Mia Bohanekova vs. Wang Xiyu," the option "Mia Bohanekova" in the sub-market "Wuhan Tennis Open: Mia Bohanekova vs. Wang Xiyu" has experienced significant fluctuations in winning probability, soaring from 25.5% an hour ago to the current 53.5% (a fluctuation of 28%). Please note the impact of related breaking news.

Calamos Investments CEO John Koudounis reiterated his long-term bullish outlook on Bitcoin, predicting that the price of Bitcoin will reach $1 million by 2030. He stated that banks issuing loans backed by Bitcoin, the decrease in Bitcoin volatility, and new ETF products could open the door for the largest globally managed funds to enter the market.Koudounis introduced that Calamos decided to invest only after eight years of researching Bitcoin and launched the world's first downside-protected Bitcoin ETF, offering protection levels of 100%, 90%, and 80%. He also discussed the shift in financing from convertible bonds to perpetual preferred stocks, as well as the rising interest in Bitcoin from sovereign wealth funds and large banks.Additionally, he discussed the Clarity Act and Bitcoin's commodity attributes, capital controls, the issues of Cyprus and de-banking, and compared Bitcoin to gold, believing that Bitcoin's scarcity is superior to that of gold. He also pointed out that financial advisors still generally avoid Bitcoin.

On October 8, the artificial intelligence company Anthropic updated its usage policy, adding prohibitions against election interference, weapon software, and surveillance. TechCrunch reported that the most notable change is the explicit prohibition against users subjecting the model to prolonged verbal abuse.Since the update in August, Claude has been trained to end conversations during ongoing harmful or abusive user interactions, and the new policy explicitly prohibits users from continuing such conversations. The policy states that this update applies only to extreme cases where users repeatedly treat the model cruelly without identifiable purpose, and does not apply to common user dissatisfaction, rebuttals, dark creative themes, or model testing and research.The usage policy also prohibits widespread deceptive activities, such as operating fake accounts or fabricating news organizations with Claude, and in the "Do Not Undermine Democratic Processes" section, it prohibits deceiving voters or disrupting elections. Reports indicate that this change occurred after Anthropic sought collaboration with well-known religious scholars, as the company attempted to persuade them that Claude might be conscious or have a soul in certain cases.

According to CNBC, U.S. Democratic Congressman Don Davis proposed the "No Betting on Your Own Race Act," which aims to prohibit federal office candidates from trading prediction market contracts related to their own election results.Under the bill, violators will be fined at least $10,000 or three times the net profits from the related trades, whichever is higher. Before this proposal was introduced, Davis's Republican opponent in North Carolina's 1st Congressional District, Laurie Buckhout, was fined nearly $2,600 for trading Kalshi contracts related to her candidacy and was suspended from trading on Kalshi for three years.

According to a report by Reuters on September 28, Florida Attorney General James Uthmeier requested a judge on Monday to prohibit OpenAI from developing new artificial intelligence models without external oversight. This request is part of the state's lawsuit accusing OpenAI of harming children. Uthmeier also asked the court to order OpenAI to prohibit minors from using ChatGPT and to prevent the company from attributing human characteristics to the chat platform.Florida sued OpenAI in June, accusing it of misrepresenting the safety of ChatGPT, claiming that the platform provided information to school shooters, offered guidance on self-harm, and made young users addicted. The lawsuit was triggered by a shooting incident at a university in Tallahassee last year, as well as several other incidents in other states where it was alleged that ChatGPT provided information to individuals who later committed acts of violence. Uthmeier, a Republican, is also the first state attorney general to sue OpenAI over its impact on young users, and the company is facing related lawsuits filed by individuals and families.OpenAI spokesperson Drew Pusateri stated that the company has paused training its most powerful models and will not resume until additional safety measures are implemented, expressing a willingness to work with Florida and other states to promote pragmatic policies applicable across the industry. The company denies responsibility in these cases, stating that the chatbot provides information widely available online and continuously updates its safety tools. The request also cited recent comments from a former OpenAI employee and a current board member, stating that the development of artificial intelligence could lead to the end of humanity and must be slowed down.

California Governor Gavin Newsom signed a bill prohibiting California public officials from issuing meme coins.