Lookonchain APP

App Store

U.S. pre-market trading saw broad declines in the semiconductor, storage, and optical communication sectors, with NOK falling more than 1%.

59 minutes ago

According to BIT (bit.com) market data, U.S. pre-market trading on Tuesday saw broad declines in the semiconductor, storage, and optical communications sectors. Semiconductor stocks: Lam Research (LRCX) fell 0.36%, Applied Materials (AMAT) dropped 0.41%, Arm (ARM) declined 0.1%. Storage stocks: SK Hynix (SKHY) fell 0.32%, Micron Technology (MU) dropped 0.4%, Western Digital (WDC) declined 0.93%, SanDisk (SNDK) rose 0.45%, Seagate Technology (STX) fell 1.06%. Optical communications-related stocks: Applied Optoelectronics (AAOI) gained 0.28%, Coherent (COHR) dropped 0.47%, Credo (CRDO) declined 0.79%, Lumentum (LITE) fell 0.63%, Nokia (NOK) dropped 1.1%.

Relevant content

Market news: Binance invests $100 million to take an equity stake in Circle.

According to market sources, Binance has invested $100 million to take a stake in Circle.

7 minutes ago

Hunan Provincial Public Security in China has cracked a virtual currency "pig butchering" scam, and two suspects have been placed under criminal detention in accordance with the law.

According to a notice from the Public Security Bureau of Linwu County, Chenzhou City, Hunan Province, China, on August 19, 2026, the bureau uncovered fraud clues during routine patrols and successfully cracked a virtual currency investment fraud case, arresting two suspects. Upon investigation, on May 1, 2026, suspect Li contacted victim Ms. Yue via private message on a social media platform, falsely claiming he could teach virtual currency investment skills and guarantee stable returns, inducing her to download an app to purchase Tether (USDT) and convert it into a newly launched virtual currency operated by his upstream accomplice. Li and his accomplice artificially inflated the coin’s price to create an illusion of growth, repeatedly persuading the victim to invest more funds. Once the money was received, they dumped the coins rapidly, causing the virtual currency’s value to plummet to zero. Ms. Yue suffered losses of approximately 30,000 yuan, while Li illegally profited tens of thousands of yuan. The illicit funds were collected by his friend Lei and then squandered entirely. Li was arrested by investigators on August 19; Lei, who was suspected of collecting the illicit funds, was taken into custody on September 17. Both suspects have been placed under criminal detention in accordance with the law. The case is currently under further investigation. The police issued a reminder: Virtual currencies are not legal tender, and there are no "guaranteed profit" investment schemes. Any private message recommendation online that promises high returns is a fraud trap. Do not trust investment recommendations from strangers or transfer funds to unfamiliar platforms.

7 minutes ago

Binance invests $100 million in Circle, with the latter rising more than 1% in pre-market trading.

According to BIT (bit.com) market data, Circle’s US stock pre-market trading rose over 1%, currently at $95 per share. The Wall Street Journal reported that Binance has invested $100 million in Circle to expand their stablecoin partnership. Separately, an 8-K filing submitted by Circle (CRCL) on September 22 reveals its subsidiary entered a new five-year cooperation agreement with Binance on September 17, expanding USDC promotion collaboration and replacing prior agreements signed by the two sides in November 2024 and August 2025.

7 minutes ago

A crypto whale withdrew approximately 14,000 ZEC from Binance, and may now hold a total of $42.5 million in spot holdings.

According to TradingBeats' monitoring, a new address withdrew a total of 14,373.05 ZEC in three transactions from Binance today, at an average price of roughly $1,510.24. It then transferred 14,373 ZEC to another aggregation wallet over the past two hours. This aggregation wallet had previously received withdrawal funds from the same exchange. Since September 17, two withdrawal addresses have cumulatively withdrawn 30,983.15 ZEC from Binance, at an average price of approximately $1,459.42. Including historical transit and downstream wallets, six associated addresses hold a total of 28,286.73 ZEC, worth around $42.562 million. Address: t1g5GALLmKnSLB1YjZWprjAaFUUHaHPYsLJ

7 minutes ago

ChatGPT can now access U.S. consumers' credit scores: integrated with Experian, it tells users which credit card is dragging down their score.

Dongcha Beating AI News Flash: OpenAI adds credit score tracking to ChatGPT Finances. U.S. Plus and Pro users can now directly link Experian credit reports and VantageScore 3.0, without needing to first connect bank accounts. Rollout is underway for web, iOS, and Android platforms. Once connected, ChatGPT can access reported accounts, balances, credit limits, credit utilization, payment history, and inquiry records, explaining which factors are impacting the credit score. Credit reports and scores are updated monthly, and users receive alerts when new inquiries, new accounts, or address changes occur. Credit checks conducted here count as soft inquiries, which do not lower credit scores. It can also analyze credit reports alongside account data integrated into Finances via Plaid. For example, it can identify which credit card uses the most available limit, recommend which card to pay off first, and advise on considerations before applying for a rental, auto loan, or mortgage. Previously, ChatGPT Finances could already pull data on spending, bills, investments, and liabilities; this update adds personal credit records to its capabilities. However, the credit score displayed by ChatGPT is VantageScore 3.0 calculated based on Experian reports, and may not match the credit scores used by banks or mortgage lenders for actual approval. After users disconnect Experian, OpenAI states that underlying credit report data will be deleted within 30 days, but information already included in chat history or Financial Memories will not automatically be removed.

7 minutes ago

Jev optimizes the attention mechanism, reducing MiniMax H3's generation time by 41.7%.

Beating AI News reports that Japanese developer Kamimoto has integrated Jev into MiniMax H3’s video generation workflow. Based on the operational status of each layer, Jev selects the proportion of Attention to retain from four preset options (1%, 3%, 5%, 10%), while Sparse Attention determines which specific computations can be omitted. When generating a roughly 5-second video on an RTX 4070 (12GB), the original 4-step approach took 6 minutes and 7 seconds; adding Sparse Attention plus Jev’s dynamic control cut the total runtime to 3 minutes and 34 seconds, a 41.7% reduction. However, this 41.7% speedup is not solely attributable to Jev. Sparse Attention itself can boost speed by skipping some Attention computations, and fixed retention rates of 5% or 10% also reduce runtime. Jev’s function is dynamic adjustment: allocating more computations to important layers and less to less critical ones. To date, there is no conclusive evidence that this dynamic adjustment delivers better video quality than fixed ratios, and the developer has explicitly stated that the quality has not been fully validated. The experimental code has been open-sourced.

7 minutes ago

Popular tokens

BitcoinEthereumHyperliquidSolanaTRONBNBTetherAaveXRPPepeFartcoinOndoJupiterUniswapBonkPendleEthenaArbitrumAvalancheLidoChainlinkPolygonDogecoinCardano