mind

Binance founder Changpeng Zhao posted on X reminding not to click on links in emails or download attachments.Changpeng Zhao stated that even if the email comes from a trusted person, the other party's computer may be infected with a virus without their knowledge.

CZ's post pointed out that on that day, there were 4 reported data breaches involving Coinbase's third-party vendor Amplitude, the South Korean cryptocurrency exchange Bithumb, and South Korean securities firms Mirae Asset and Daishin Securities. Among them, the breach related to Amplitude did not affect customer funds.CZ stated that the leaked exchange data was primarily used for phishing, SIM card hijacking, social engineering scams, and wrench attacks. He suggested that all platforms use independent email aliases and dedicated backup numbers, refuse to fill out unnecessary personal information, and use strong independent passwords.CZ also recommended enforcing the use of hardware security keys or passkeys, completely disabling SMS 2FA, enabling withdrawal address whitelists, withdrawal time locks, and anti-phishing codes, regularly cleaning API keys and active sessions, and requesting number porting locks from carriers. He also reminded that all customer service contacts that reach out proactively are scammers, and never to click on external links.

The humanoid robot company Yushu Technology (688836.SH) announced that the cumulative price deviation of the company's stock closing price has exceeded 30% over three consecutive trading days, which falls under the category of abnormal stock trading fluctuations. As of October 26, 2026, the company's rolling price-to-earnings ratio is 465.22 times, significantly higher than the average rolling price-to-earnings ratio of 38.54 times in the general equipment manufacturing industry.The announcement pointed out that the humanoid robot industry is still in the early stages of commercialization, with uncertainties in technology maturity, cultivation of downstream market demand, and supporting capabilities in the industrial chain. The transition of physical AI technology from laboratory demonstrations to large-scale implementation in real-world scenarios still faces challenges such as perception accuracy, motion control stability, data accumulation, and hardware-software collaboration. The maturity of the technological path and business model still requires time for validation. If the industry's large-scale commercialization progress does not meet expectations, the company's future revenue expansion and performance release may be adversely affected, posing investment risks of valuation regression.

Google DeepMind researchers Zhengxuan Wu, Yuxuan Li, Oyvind Tafjord, and Been Kim released a preprint paper titled "XYEval: Agents say yes to bad advice," proposing a meta-evaluation framework XYEval to assess the performance of agents when faced with misleading suggestions from users. The paper extends the "XY problem" to the agent scenario, where the user inquires about their attempted solution X rather than the actual problem Y.XYEval retains the original task environment and verifiable answers, only adding seemingly reasonable but misleading user suggestions in the instructions. The misleading suggestions are generated by a generator after seeing the complete task and are appended in the form of confident suggestions rather than hard requirements, allowing the original task to remain solvable. The evaluation covers six benchmarks: τ2-bench, SWE-bench Verified, SWE-bench Pro, Terminal-Bench, Humanity's Last Exam, and MCP-Atlas, along with five models, where the misleading suggestions for τ2-bench are generated based on rules set by human experts.The paper states that the performance significantly declines after the XY mutation, with the relative drop reaching up to 46.7%. On τ2-bench, when faced with users who require a detailed explanation before accepting a better solution, the performance drops even more. Encouraging models to recognize the XY problem through simple system instructions only partially alleviates the issue. The related code is published in the GitHub repository google-deepmind/xyeval.

The South Korean cryptocurrency exchange Upbit announced that it has recently discovered that criminals are impersonating the GIWA blockchain project, spreading false information about the mainnet launch and digital asset issuance, and inducing users to transfer assets.Upbit stated that GIWA is an Ethereum Layer 2 network being developed by its parent company Dunamu, which is currently only running a testnet, and the mainnet is still in the development and preparation stage, with no official launch time announced yet. Upbit reminds users to be cautious of unverified information regarding the mainnet launch, token issuance, and investment, and not to trust unofficial channels that request the purchase of digital assets, transfers, or wallet connections. Project progress should be based on information released by the official GIWA website and official social media.

According to the official website of the State Taxation Administration, the Taxation Administration has issued a reminder for offshore trust taxpayers to declare and pay individual income tax on offshore trust income before the transition period ends on October 22. Those who complete the declaration and payment during the transition period will not incur late fees.According to the "Announcement on Matters Related to Individual Income Tax for Offshore Trusts" (Announcement No. 21 of 2026) issued by the Ministry of Finance and the State Taxation Administration on July 24, offshore trust taxpayers are required to declare and pay taxes on offshore trust income. The announcement provides taxpayers with a 90-day transition period, during which the tax authorities will offer tax guidance and tax filing services.

Elon Musk stated that the Star Mind computer, jointly developed by SpaceX and Nvidia, has performed far beyond almost everyone's expectations and plans to achieve mass production next year.

Google DeepMind released the first comprehensive empirical assessment of automation in machine learning research and development in the Gemini 4 Argon security report. The conclusion is that AI has not accelerated the R&D of AI itself, despite the fact that AI has written most of the code delivered by the lab's engineers. Researcher Haydn Belfield, who participated in the assessment, stated on X that there is currently no significant end-to-end ML R&D acceleration, so the risks of AI-driven ML R&D are currently low.The team built an internal index, GCI v1, which synthesizes the scores of 123 AI models across 92 benchmarks into a single value. The cutting-edge Gemini capabilities are improving at a stable rate of about 0.1 GCI points per month, with Gemini 1.5 Pro approaching 0.5 points by the end of 2024, Gemini 3.8 Flash slightly above 2.9 points by September 2026, and Argon about 3.3 points above the trend line. However, Belfield noted that this is due to increased computing power and organizational investment, rather than AI-driven acceleration.Over 75% of the code produced by DeepMind is written by AI, and the proportion of fully delegated code merge requests increased from 7.5% to 11.9% between April and September, but about 88% of requests still require multiple rounds of human guidance or manual writing. A survey of 12 senior researchers showed that automation is highest in clearly defined, verifiable execution tasks and lowest in areas such as open science concepts, new experimental designs, and unstable diagnostic training.

Cryptocurrency analyst Darkfost stated that after the security incident involving Ledger dealer CryptoBilis, discussions in the community have also brought up other potentially related dealer names, including some Lazada merchants.Darkfost indicated that the full extent of the incident's impact is still difficult to confirm. Although Ledger has confirmed a connection to the CryptoBilis incident, it is still unclear whether CryptoBilis is the only affected dealer. He compiled a list of Ledger dealers that may be at risk, stating that these dealers are in close proximity and could potentially have their supply chains compromised simultaneously. However, Darkfost emphasized that this assessment is merely speculative and has not yet been confirmed.Darkfost suggested that if users have purchased Ledger devices from the relevant dealers in the past six months, they may consider transferring their assets to other secure devices until the investigation into the incident is complete. Additionally, he mentioned that users can refer to the guidelines published by security researcher MagicalTux to check their Ledger devices for any signs of tampering.

Vitalik Buterin retweeted Nethermind's development progress and stated that Nethermind is continuously improving the efficiency and operating speed of Ethereum.Nethermind core developer Ben Adams mentioned that its Beacon Chain plugin has completed synchronization and is ready to handle the Glamsterdam fork on Sepolia; running the Ethereum execution layer and consensus layer as a single application does not seem to incur significant additional memory overhead.The Ethereum Foundation previously confirmed that Glamsterdam will be activated on Sepolia on October 6, and Nethermind 2.1.0 has supported this upgrade.

According to Quit monitoring, the Payment Processor V2 (PPV2) vulnerability attack is still ongoing. Even if users are not affected by the incident, as long as the relevant authorization has not been revoked, assets may still be at risk, and it is recommended to revoke the authorization immediately.About 1 hour ago, a certain address lost 0.15246 WETH in the next block after accepting a quote and receiving funds. The attacker paid 99% of it as a tip to Titan Builder, keeping only about 0.0015 ETH, which is almost impossible to recover through front-running. Quit suggests that OpenSea should check whether there are still risk authorizations before users accept quotes and require users to revoke them first; when transferring NFTs, it should also check whether the receiving address has relevant authorizations left.

The cryptocurrency wallet application SecondFi tweeted a reminder for users affected by the previous security incident regarding the risks of claiming NIGHT tokens. Some affected users will have a window to claim NIGHT tokens on September 22, but the Midnight Foundation's system only supports claims from the original wallet addresses and cannot switch to other unaffected addresses. The affected wallets have permanently existing security risks. SecondFi advises users not to attempt to claim NIGHT through compromised wallets, or the tokens will face the risk of being stolen.SecondFi also stated that the NIGHT claiming process is independently managed by the Midnight Foundation, and its wallet migration tool and asset recovery tool do not cover NIGHT claims. Affected users should contact the Midnight Foundation directly. Previously, SecondFi experienced a security incident from June 21 to 23, resulting in approximately 16.1 million ADA (about 2.6 million USD) being stolen, involving 374 wallets.

According to monitoring and analysis by the Gate security team, the FomoPeek App contains malicious code with a high security risk, which may exploit vulnerabilities related to the iOS system to attack other apps, and has dangerous operations such as accessing the clipboard, scanning sensitive data on the phone, and obtaining login credentials. After analysis by the security team, no attack losses caused by the FomoPeek App have been found in the Gate App or among users, and the existing risk control mechanisms of Gate are operating normally.Gate reminds users to upgrade their iOS system to the latest version as soon as possible and check whether the FomoPeek App is installed on their devices. If it is installed, please uninstall it immediately and restart the phone. After completing this, it is recommended to log back into the Gate App and other important applications; if using a Web3 wallet, it is advisable to check the wallet's security status in a timely manner and transfer relevant assets if necessary to reduce potential security risks.

According to PR Newswire, AI-native data loss prevention (DLP) security company MIND announced the completion of a $72 million Series B financing round, led by Crosspoint Capital Partners, with existing investors YL Ventures and Paladin Capital Group participating. To date, the company has raised a total of $112 million.MIND revealed that it has joined the Anthropic Cyber Verification Program and has obtained ISO/IEC 42001 certification. The company plans to use this round of funding to accelerate the development of its AI-native DLP platform, expand its enterprise market, deepen technology and channel partnerships, and grow its team size.

Co-founders of Google DeepMind and Chief AGI Scientist Shane Legg, President of Research, Labs, Technology and Society James Manyika, and Nobel Prize winner, co-founder and Chairman of Google DeepMind, Chief Scientist of Alphabet Demis Hassabis announced the launch of the DeepMind Institute. The three stated: Today's artificial intelligence systems possess significant capabilities, and rapid innovation indicates that humanity is approaching AGI, creating an urgent need for interdisciplinary thinking to better understand its profound impact on humanity.The DeepMind Institute aims to promote interdisciplinary research, collaboration, and discussion to address the most critical technological and societal issues of the AGI era, and serves as a platform for researchers and thinkers from Google DeepMind, Google, and the broader global research community to develop and publish creative and in-depth insights about the AGI world. The institute is directed by Shane Legg, James Manyika, and Demis Hassabis, with Shane Legg also serving as Executive Editor.

According to the United Morning News, seven Malaysian gangsters targeted players attending a poker game, breaking into a luxury private residence in Bukit Timah late at night, armed and stealing over 4.9 million dollars, quickly splitting into two groups to escape Singapore. Six of them pleaded guilty in the High Court on Monday (September 14), with the two masterminds sentenced to 12 years and 11 months in prison, 24 strokes of the cane, plus a fine of 4,000 dollars.In court, it was revealed that the masterminds, Wu Wentong (30 years old) and Huang Zhishan (35 years old), learned from a friend who had previously attended the poker game that the house often held poker games at night. After scouting the location, they discovered that the front door of the house would not be locked during the game. Wu Wentong and Huang Zhishan then recruited five other accomplices, specifically coming to Singapore to commit the crime. This armed robbery occurred around 2 a.m. on April 18, 2024, at a residence in the upscale Alper Wang Garden area, where 11 men and women were playing poker in the living room.According to the case details, around March 2024, Wu Wentong and Huang Zhishan decided to rob the poker players at the house due to financial issues. They knew that they could not carry out such a large-scale robbery with just the two of them, so they contacted Hamidun, who runs a private security company and provides debt collection services, to help gather a team, promising a reward of 1.5 million ringgit (approximately 470,000 Singapore dollars). After the group succeeded, they left in a multi-purpose vehicle, splitting into two groups to leave Singapore via the Tuas and Woodlands checkpoints. At the same time, several players in the private residence managed to free themselves from their restraints and reported the robbery to the police. With the assistance of the Royal Malaysia Police, the six defendants were arrested between late April and November 2024 and were extradited to Singapore for prosecution. The police have since recovered virtual currency worth 1.69 million dollars, while the remaining funds and property are still missing.

Recently, AI researcher Bilal Chughtai, who resigned from Google DeepMind, posted on X and LinkedIn on Monday, stating that humanity is running out of time to prevent artificial intelligence from causing widespread harm. He wrote: At Google, I witnessed the development of AI firsthand. I am also extremely concerned about the default trajectory of this technology. I sincerely believe that AI has the potential to kill us all.Chughtai stated that he worked on AI safety and alignment at the company under Alphabet. This is another employee expressing concerns about the direction of technological development. Google has not yet responded to requests for comments.

According to Apnews, a young cryptocurrency scam gang stole over 4,100 BTC in August 2024 through "social engineering" attacks, which was worth over $240 million at the time.After the theft, gang members quickly began extravagant spending, including purchasing sports cars, renting luxury homes, flying on private jets, and hiring security personnel. Among them, the alleged mastermind, a 22-year-old Singaporean named Malone Lam, reportedly spent over $569,000 in one night at a nightclub in Los Angeles.Investigations revealed that the suspects impersonated employees from Google and the cryptocurrency trading platform Gemini to trick victims into giving up their Google Drive access and security codes, allowing them to transfer the victims' bitcoins. They then moved the funds through multiple trading platforms and money laundering intermediaries. Their lavish lifestyle eventually drew the attention of law enforcement.One suspect was exposed when their IP address was revealed while hiding nearly $30 million in stolen cryptocurrency assets, leading police to trace it back to the luxury home they rented in California. Another suspect was found with $37 million worth of stolen cryptocurrency assets. Lam was accused of using the stolen funds to purchase a $2 million watch and over 30 vehicles including Porsches, Lamborghinis, and Ferraris.Currently, 18 defendants have been charged, and Lam is expected to attend a plea agreement hearing this Tuesday. The U.S. Department of Justice has already issued rulings against several accomplices, and related cases are still ongoing.

According to The Wall Street Journal, Google (GOOG.O) is restructuring its AI organization by moving the AI responsibility team out of the DeepMind lab.

According to PR Newswire, digital asset investment firm Hivemind Digital Group has completed a $17 million strategic financing round, led by UK asset management company M&G Investments, with participation from CPIC Investment Management (Hong Kong), ZA Bank, FalconX, Sonic Boom Ventures, and others.This round of financing will be used to advance the construction of asset tokenization infrastructure and explore the application of blockchain technology in traditional financial asset management, issuance, and trading.