A resignation post racked up 140 million views in total—was it actually used to stoke AI panic for the U.S. Democratic Party? Jacob Coxon, a former researcher at Anthropic, posted his resignation, and in just one day the views surpassed 100 million. They have now reached 140 million. Conservative U.S. think tank Capital Research Center researcher Parker Thayer believes the sudden viral spread does not look like organic sharing, but more like a well-prepared, well-funded PR campaign designed to amplify fears about AI extinction and build momentum for the Democratic Party’s push for stricter AI regulation. First, the dissemination pace was far too fast. Eighteen minutes before Coxon posted, the Wall Street Journal had already published an exclusive report. Within less than 15 minutes after the post went live, the earliest three quote reposts came from Encode AI, the AI Policy Network, and the AI Futures Project—organizations that have long advocated for AI safety and oversight. Second, behind these organizations is the same key benefactor. SFF’s public records show that Skype co-founder Jaan Tallinn has long funded the AI Futures Project, Encode AI, the AI Policy Institute, and other groups, with amounts ranging from hundreds of thousands to over a million dollars. Tallinn is also an early investor in Anthropic, and in 2021 his lead investment helped drive Anthropic’s $124 million Series A round. Third, Coxon himself is connected to this circle. Thayer says Coxon received a “Long-Term Future” scholarship award from Good Ventures in 2022. Good Ventures was founded by Facebook co-founder Dustin Moskovitz and his wife, and Moskovitz, too, was an early investor in Anthropic. Just 27 minutes after Coxon posted, Max Nadeau—who handles technical AI safety funding at Coefficient Giving, a charitable grantmaking organization that Moskovitz has supported long-term—also joined the spread. Finally, the political relay quickly followed. Several Democratic politicians later invoked Coxon’s warnings to call for stronger AI regulation. Even before Coxon went viral, the left-wing heavyweight Bernie Sanders had already prepared legislation to ban superintelligent systems and pause advanced AI development. Coxon’s sudden viral surge at exactly this moment hands the pro-regulation camp a ready-made piece of political ammunition.
Osmosis to confiscate attackers’ assets and use community BTC to cover the AlloyedBTC shortfall On September 10, Osmosis officials stated that a solution proposal for the earlier Nomic chain nBTC incident has been published and has officially entered community discussion. According to the proposal, the Osmosis governance community will consider confiscating the attackers’ assets that were previously frozen and using BTC accumulated in the community fund pool to cover the remaining collateral shortfall, in order to restore AlloyedBTC’s full collateral support. This proposal still needs to go through governance discussions and a vote; it is not yet completed in terms of asset confiscation or collateral top-up.
A whale bought 767.8 BTC via THORChain within two days, at an average purchase price of about $78,628. On September 10, according to Ember CN monitoring, the whale continued to cross-chain purchase 544.5 BTC via THOR Chain today, worth about $42.43 million. As of now, over the past two days, the whale has exchanged 60.37 million USDC for 767.8 BTC through THORChain, with an average purchase price of about $78,628.
Castle Securities calls on the U.S. SEC to regulate event contracts related to U.S. stocks On September 10, Castle Securities (Citadel Securities) submitted comments to U.S. regulators, calling on the U.S. Securities and Exchange Commission (SEC) to regulate event contracts related to publicly traded U.S. companies. The market maker said that trading platforms should not bypass the SEC’s regulatory jurisdiction by self-certifying with the U.S. Commodity Futures Trading Commission (CFTC). The dispute over regulatory authority for event-contract oversight continues to heat up. Currently, some U.S. prediction markets have launched related contracts under the CFTC framework. As event contracts gradually expand to cover financial assets such as stocks, the regulatory boundary between the SEC and the CFTC has become a key focus for the market.
Micron Coin Meme MOO: up more than 80% in 24 hours; market cap rises to $42 million On September 10, according to GMGN market data, Robinhood ecosystem meme coin MOO surged more than 80% in the past 24 hours, with its market cap reaching $42 million. The 24-hour trading volume was $6.4 million. Meme coin MOO leverages the stock trading platform Robinhood, pairing tokenized U.S. stocks of Micron (MU) as the pool, using the MOO/MU trading pair to provide liquidity. Note: Meme-coin trading is highly volatile and relies largely on market sentiment and hype around concepts. It has no real value or use case. Investors should be aware of the risks.
Binance launches US stock recurring investment feature, supports 100 stocks and ETFs On September 10, Binance announced the launch of Stock Recurring Buy (stock recurring investment) functionality, allowing eligible users to automatically purchase specified stocks and ETFs on a fixed schedule without having to place orders manually each time. The feature went live on September 10. Currently, it supports 100 stocks and ETFs, including popular ETFs such as Apple, Tesla, Nvidia, Amazon, Microsoft, Meta, Coinbase, Robinhood, as well as SPY, QQQ, and VOO. Users can select one or multiple stocks within the same recurring investment plan. Binance stated that the feature is available to eligible users, and the exact availability depends on the user’s country or region. The execution price of automatically placed orders may differ significantly from the price at the time the plan is created. Users should ensure that the specified funded account has sufficient assets on the execution date. In June this year, Binance launched its US stock trading service, enabling users to trade US-listed stocks and ETFs through a single account, with a minimum investment amount of $5.
Loracle is currently shorting PONS and is up about $21.8 million; it previously had a floating loss of more than $80 million. On September 10, according to Arkham monitoring, Loracle is currently shorting PONS on Hyperliquid, with a position value of about $16.3 million, accounting for 17.5% of that token’s total open interest. Loracle opened this short position 9 days ago at $0.44 and kept adding to it as the PONS price continued to rise, reaching a high of $0.97. At one point, the floating loss on its short position exceeded $80 million. As the PONS price pulled back, Loracle has returned to profitability, with the short position’s floating profit now about $21.8 million.
Before the US stock market opened, the storage sector saw a slight decline. Micron fell 0.93%, and SanDisk fell 0.83% On September 10, according to market data from BIT (bit.com), the storage sector in the US traded slightly lower before the opening. Specifically: SanDisk (SNDK) fell 0.83%; Seagate Technology (STX) fell 0.67%; Western Digital (WDC) fell 0.98%; Micron Technology (MU) fell 0.93%; and SK Hynix ADR fell 3.50%.
Media reports say the Houthis seized Mokha port; risks for Red Sea shipping are heating up On September 10, Israeli i24 News reported that on Wednesday, the Saudi-backed Yemeni government forces announced that they had made major advances in the northern and central desert regions and appeared to have broken through the Houthis’ defensive lines. In reality, however, the Houthis simply redeployed their forces, voluntarily abandoning open areas in order to take control of more important western coastal regions. By the early hours of Thursday, after a large-scale withdrawal by the Yemeni government forces, Iran-backed Houthis had taken control of the strategically significant coastline, as well as the port of Mokha in Hudaydah Governorate. Thus, neither side appears to have launched an overwhelming rapid offensive; it is more like a land swap: the Houthis gained strategically important coastal frontage along the Red Sea. This will enable the Houthis to pose a more direct threat to shipping through the Bab el-Mandeb Strait, while also further increasing security threats to Saudi Arabia.
Flop Labs Unveils Latest Tokenomics Draft: No VC, No Presale, Total Supply at 18.1B in Year 10 On September 10, Flop Labs officially released an updated FLOP tokenomics draft (Draft), based on community feedback. The project team emphasized that the token includes no venture capital (VC) allocation, no presale (Presale), and that all tokens must be earned through network contributions. Key data and total supply model: Total supply in Year 10: expected to reach 18.1 billion (18.1bn). Long-term inflation rate: maintained at 0.5% per year (Terminal Inflation). Halving mechanism: uses fixed halving cycles (Halvings); after each halving, a permanent tail inflation mechanism is retained to continuously incentivize network participants. Token distribution ratio in Year 10 (Total Supply Breakdown): Miners: 8.8 billion (48.6%), the largest allocation share, reflecting a network-oriented Proof of Useful Inference. Airdrop: 4.4 billion (24.3%). Of which: Miners: 1.2 billion (6.6%); Validators: 1.2 billion (6.6%); Agents: 1.2 billion (6.6%); Reserve / Incentives: 0.8 billion (4.4%); Team + Foundation: 2.0 billion (10.8%). Validators: 1.2 billion (6.5%). Brokers/agents: 1.2 billion (6.5%). Staking rewards: 0.6 billion (3.2%).
OpenAI Faces Fresh Math Results Controversy: Accused of Possibly “Stealing” Major Mathematical Proofs OpenAI Astra’s claimed math breakthrough is once again mired in controversy. Mathematician Andreas Thom recently posted a challenge, questioning whether, during the training of Astra, OpenAI may have absorbed conversation content from his research with mathematician Gábor Kun into the work they conducted around the Gromov soficity conjecture—after which Astra then solved a highly related problem. The Gromov soficity conjecture is an important long-unsolved problem in group theory, and Thom has been studying it for 20 years. When OpenAI released its “Top Ten Mathematical Breakthroughs” for Astra in August this year, the third item was the construction of a non-sofic group. OpenAI previously said the related results were achieved by its internal version Astra, but after facing skepticism from the mathematics community, it modified its wording and acknowledged that the proof relied on recent mathematical work, including research by Kun and Thom. As of now, there is still no evidence showing that OpenAI indeed used the relevant private conversations to train Astra, but Thom’s accusations have once again brought to the forefront questions about AI training data sources, attribution for unpublished research results, and whether “AI discoveries” are built upon human findings that were not publicly disclosed. If the allegations are ultimately confirmed, the incident would no longer be just an academic authorship dispute—it could evolve into a major controversy involving AI company data usage, research ethics, and intellectual property.
Singapore Exchange Receives CFTC Authorization, Opens Bitcoin and Ethereum Perpetual Futures to U.S. Institutions On September 10, according to CoinDesk, the Singapore Exchange (SGX) has received authorization from the U.S. Commodity Futures Trading Commission (CFTC) under Regulation 48.10 of the U.S. Commodity Exchange Act, allowing U.S. institutional investors to directly trade its Bitcoin and Ethereum perpetual futures contracts. KC Lam, head of crypto derivatives at the SGX Group, said that previously U.S. participants were unable to trade the related contracts; this opening will connect traditional U.S. financial trading institutions with Asia’s liquidity pools. SGX’s Bitcoin and Ethereum perpetual futures are scheduled to launch at the end of November 2025. Cumulative trading volume has already reached $5.8 billion, or about 400,000 contracts. As of this August, the two contract types have averaged about 1,300 contracts per day in trading volume, with a nominal value of about $19 million. The Bitcoin contracts account for 66% of open interest since launch and 83% of average daily trading volume. The highest single-day trading reached 115,000 contracts, with a nominal value of $145 million. SGX expects that U.S. clearing members will gradually onboard customers within the next one to two months. The exchange also plans to roll out Bitcoin and Ethereum futures and options products with expiry dates.
Apple’s Pricy Foldable iPhone Faces a Big Test: Can Duo Open a New Growth Cycle? On September 10, Apple officially unveiled its first foldable iPhone, the Duo, pushing the iPhone price ceiling to $1,999—and sparking a debate in the market: will a high-priced foldable screen become Apple’s new growth engine, or will it ultimately end up as a niche high-end product? Gene Munster, Managing Partner at Deepwater Asset Management, significantly raised his Duo forecast, expecting Duo to account for twice the share of iPhone revenue in fiscal 2027—from 5% to 10%—and calling Duo Apple’s most important new product since the AirPods launched in 2016. He believes that if roughly one-third of Pro Max users switch to Duo, Duo could drive iPhone revenue growth of 5% and Apple’s total revenue growth of about 2.5%. However, Wall Street’s stance is clearly cautious. Morningstar expects Duo to lift the average selling price of iPhones, but projects that iPhone unit sales in fiscal 2027 will be largely flat; meanwhile, rising storage costs may put pressure on Apple’s gross margin. Jefferies, meanwhile, keeps a “underperform the broader market” rating, arguing that the 256GB base version of Duo suggests Apple may prioritize unit sales over per-device profit. The bigger challenge comes from market size. Today, foldable phones account for only about 2% of global smartphone sales, and FDM CCS Insight projects that figure will rise to just around 4% over the next decade. Forrester therefore warns that Duo could become a high-priced niche product similar to Vision Pro, or it could, like the Apple Watch, build an independent market—but whether it can reach a meaningful scale remains to be seen. Priced at $1,999, with a 2TB version topping out at $3,199, Duo will open for preorders on October 16 and go on sale on October 23. At the same time, Apple raised the starting price of the iPhone 18 Pro to $1,199. For the newly appointed CEO, John Ternus, whether Duo can become the next iPhone growth engine will also be his first major test after taking office.
Bloomberg AI model predicts the Bank of England may raise rates in February next year On September 10, Bloomberg Economics’ research team, using a machine-learning model built with methodologies similar to those of the European Central Bank, showed that under current assumptions for energy prices and interest-rate expectations, the Bank of England (BOE) is highly likely to raise the benchmark rate from 3.75% to 4% at its first policy meeting next February—earlier than most economists’ forecasts of holding rates unchanged. The model was trained on more than 50 economic and market indicators since the Bank of England became independent in 1997, and it has a prediction accuracy of roughly one-half to two-thirds for actual rate-hike or rate-cut decisions over the past 10 years. The study noted that if elevated oil and gas prices persist and lift inflation to around 4%, upward risks to inflation expectations would force the BOE to tighten monetary policy.
Meme coin ROBIN’s market cap briefly breaks $15 million, up over 92% in 24 hours On September 10, according to GMGN market data, the Robinhood on-chain meme coin ROBIN briefly surpassed a market cap of $15 million, but has since pulled back to $12.62 million. It is up more than 92% over the past 24 hours, with trading volume in the same period reaching $1.8 million. ROBIN (Robin the Frog) is directly drawn from a character that actually exists in The Muppets (Sesame Street): Robin the Frog = Kermit the Frog’s (科米蛙) nephew. The character also appears in Muppet Robin Hood, directly tying into the Robin Hood story, making it highly compatible with the Robinhood Chain narrative. The community calls it “the Pepe of the Robinhood Chain” or “the chain’s native frog.” “The character existed first, the chain came later, and the timing connected them.” It focuses on community-driven, Sherwood-themed Meme culture. A reminder to users: many meme coins have no real use cases, and prices can swing significantly—invest with caution.
BTC falls below $78,000 Market updates show that BTC has fallen below $78,000. It is currently trading at $77,996.97, with a 24-hour decline of 1.61%. Volatility is high—please manage risk accordingly.
DeepSeek Breaks the “Impossible Triangle” of Large Models: Stronger, Faster, and Even Cheaper This time, DeepSeek V4.1 Flash has almost completely redesigned the entire architecture, aiming to enhance capabilities, boost speed, and reduce costs at the same time. Stronger: the model has 552 billion backbone parameters, and it also adds Engram conditional memory with 196 billion parameters. Pretraining used over 45T multimodal tokens, and post-training includes a large number of real Agent tasks, tool environments, and failure cases. DeepSWE v1.1 reaches 74.2%, already surpassing Claude Opus 5 and GPT-5.6 Sol. Faster: the new CED architecture splits a 40-layer model into two halves—front and back. When reading the prompt, each token activates only 8 billion parameters; during generation, it’s also just 16 billion. On top of that, CSA2 cross-layer reuse and DSpark speculative decoding extend the context from 4K to 1M, increasing length by 256x, while the per-token decoding compute cost increases by only about one quarter. Cheaper: DeepSeek continues to push KV Cache compression further. The main cache is changed to FP4, plus cross-layer reuse, so the global KV per token is reduced to only 890 bytes—about 1/4 of V4 Flash; long-term caches stored on SSD or in memory are further reduced to about 1/8. V4.1 Flash doesn’t simply equate “stronger” with “doing more computation.” The model size keeps growing, but each time it only adjusts a small portion of parameters; the context keeps getting longer, while the cache gets smaller. Capabilities improve, and speed and costs aren’t dragged down.
August Crypto Trading Platform Rankings: Spot Trading Volume Jumps Over 18% Month-over-Month; Existing Active Liquidity Drives a Pickup in Trading On September 10, 2026, trading activity in the August crypto market clearly rebounded. Key figures are as follows: Spot-led growth: Mainstream CEX spot trading volume rose 18.64% month-over-month, while perpetual futures trading volume increased 10.89% month-over-month. The spot growth outpaced futures, indicating that the market rebound was mainly driven by spot buying demand. Flows and trading diverge: Mainstream CEX website visits in August edged down 0.26%, and app downloads fell 5.61%. Against a backdrop of slower new traffic growth, trading volume rose against the trend, suggesting that this round of growth came primarily from higher trading frequency and improved conversion efficiency among existing users. Perp DEX also rebounds: Trading volume for mainstream perpetual contract DEXs rose 8.98% month-over-month in August, while overall web traffic remained stable. CEX ranking and performance: Trading volume leader: Binance maintained the industry’s No. 1 position with $233B in spot volume and $1606B in futures volume. OKX (spot $46B / contracts $634B) and Bybit (spot $46B / contracts $365B) ranked second and third, respectively. Position changes: Coinbase’s spot ranking climbed two places to No. 7; Kraken rose one place to No. 9. User traffic rankings: Binance (2.532M downloads, 34.19M web traffic) and OKX (0.55M downloads, 25.62M web traffic) continued to hold the top two spots. Perp DEX highlights: Hyperliquid led the decentralized perpetual futures market with $207B in contract trading volume. edgeX saw a 70.60% month-over-month surge in trading volume; ApeX, Lighter, and Aster all recorded substantial growth in the range of roughly 17%–22%.
Ministry of Commerce: China and the United States are in consultations on a framework for reciprocal tariff reductions totaling US$30 billion On September 10, at a regular press conference held by the Ministry of Commerce on the 10th, Huang Ling, spokesperson of the Ministry of Commerce, said in response to questions that the economic and trade teams of China and the United States have conscientiously implemented the consensus reached between the heads of state of the two countries at their Beijing meeting. They are currently consulting on a framework for reciprocal tariff reductions totaling US$30 billion, aiming to have it implemented as soon as possible. (Xinhua)
Log in to explore more content
Join global crypto users on Binance Square
⚡️ Get latest and useful information about crypto.