生成式人工智能 · 121

[Comparative Daily News Picks] Manus: It will resume operations as an independent company, and some user data will be deleted; Google Gemini app has more than 1 billion monthly active users; Bloomberg: Despite the “Clarity Act” impasse, the US SEC is preparing to announce a major crypto regulation plan; 10x Research: Bitcoin may break its correlation with the S&P 500, and Bitcoin and gold are expected to benefit if employment weakens in summer

Daily AI · Cryptography · Macro · Market News, Bitpush helps you set priorities ↓ AI · News [Manus: Will resume operations as an independent company, some user data will be deleted]. On August 11, Manus published “A Letter to Manus Users”. Manus said it will soon resume operating as an independent company as part of resuming independent operations. At the same time, in order to comply with the regulatory requirements of specific jurisdictions, data generated by some users on or after December 29, 2025 will be deleted from 08:00 on August 23 to August 24 (SGT). Affected users can back up data from now until 08:59 (SGT) on August 23, 2026, and restore data from 08:00 (SGT) on August 25, 2026. During this time, unaffected users can use Manus as usual without taking any action. [Google Gemini app has more than 1 billion monthly active users] Compared to Twitter news, the Google Gemini app has more than 1 billion monthly active users, making it the fastest growing product in Google history. [CoreWeave's second-quarter revenue doubled, and the stock price surged 12% after the market] Compared to news, Coreweave (CRWV.O) rose 12% in post-market trading on Tuesday. It previously announced second-quarter revenue of US$2.58 billion, up 112% year over year, exceeding Wall Street expectations, indicating that the market's demand for AI computing power is still growing rapidly; the net loss is US$626 million, up from US$290 million in the same period last year; currently, order reserves have reached US$104 billion, and the project capacity under construction is 1.5 gigawatts. CoreWeave is speeding up the expansion of its data center business, competing with cloud computing giants such as Amazon, Google, and Microsoft to compete for the data center market where chips are deployed and can run generative artificial intelligence models. However, CoreWeave is currently unprofitable. By the end of the quarter, the company's balance sheet debt reached $35 billion to cover Nvidia GPU and other equipment procurement costs. This quarter, Meta said it would invest an additional $21 billion into CoreWeave. Additionally, CoreWeave also announced a multi-year cooperation agreement with Anthropic and received $6 billion in committed funding from quantitative trading company Jane Street. Crypto · Market [Bloomberg: Despite the “Clarity Act” at an impasse, the US SEC is preparing to announce a major crypto regulation plan]. According to Bloomberg, despite the impasse in the “Clarity Act”, the US SEC is preparing to announce a major crypto regulation plan. [CME and Silicon Data plan to launch computing power futures in October] According to comparison news, the US CME (CME) and Silicon Data have confirmed that they will launch ComputeFutures (ComputeFutures) on October 5 this year. These innovative trading tools will provide hedging and investment tools for businesses that want to control computational costs. The two contracts, Silicon Data H100 Rental Index Futures and Silicon Data B200 Rental Index Futures, will track an index that measures the hourly rental cost of image processors (GPUs). The two new contracts will be listed on CME and are subject to New York Mercantile Exchange (NYMEX) rules, and will not take effect until approved by the Supervisory Authority. [Polymarket reorganizes the team and introduces multiple executives to accelerate compliance and US expansion, and welcome the fall forecasting market boom] Comparing news, the prediction market platform Polymarket is undergoing organizational upgrades for the upcoming fall trading peak, preparing for the expansion of the US market by introducing multiple executives, restructuring the marketing system, and strengthening compliance teams. Polymarket recently hired Travis VanderZanden, founder of electric scooter sharing company Bird and a former Uber and Lyft executive, as Chief Growth Officer, responsible for building the company's growth strategy and marketing system. VanderZanden said the forecasting market is at a critical stage of rapid development, and the company needs to further improve its management team to support long-term growth. Macro · Agency [Trump: Iran likes to play tricks, face one set behind the other] To compare the news, US President Trump recently stated, “If we don't attack Iran, it will acquire nuclear weapons. So,...

10d agoWendy#Compare Daily Picks

CoreWeave's revenue doubled in the second quarter, and the stock price surged 12% after the market

According to news, Coreweave (CRWV.O) rose 12% in post-market trading on Tuesday, after announcing second-quarter revenue of US$2.58 billion, up 112% year on year, exceeding Wall Street expectations, showing that the market's demand for AI computing power is still growing rapidly; net loss is US$626 million, up from US$290 million in the same period last year; order reserves have reached US$104 billion, and the project capacity is 1.5 gigawatts under construction. CoreWeave is speeding up the expansion of its data center business, competing with cloud computing giants such as Amazon, Google, and Microsoft to compete for the data center market where chips are deployed and can run generative artificial intelligence models. However, CoreWeave is currently unprofitable. By the end of the quarter, the company's balance sheet debt reached $35 billion to cover Nvidia GPU and other equipment procurement costs. This quarter, Meta said it would invest an additional $21 billion into CoreWeave. Additionally, CoreWeave also announced a multi-year cooperation agreement with Anthropic and received $6 billion in committed funding from quantitative trading company Jane Street.

10d agoWendy

India: The world's first country to be shorted by artificial intelligence?

Source: Qin Shuo's Circle of Friends In the first half of 2026, an impactful new label appeared in the Indian stock market — “the first country in the world to be shorted by artificial intelligence.” This assertion is not without foundation. The Nifty IT Index, which has long been regarded as a weather vane for India's technology industry, fell sharply in the first half of the year. Software service leaders Tata Consulting Services, Infosys, and Wepro are generally under valuation pressure. Meanwhile, international capital continues to withdraw from the Indian market. According to Reuters data, in the first half of 2026, foreign investors sold approximately $29 billion of Indian stocks on a net basis. After entering July, although the Nifty IT Index rebounded 16.7%, and the net inflow of foreign capital surpassed 1.6 billion US dollars, this round of market was largely due to sector rotation after global capital withdrew from crowded AI hardware transactions, which is not enough proof that India's software industry has escaped trouble. A large country with a population of 1.4 billion and many industries will of course not be easily “shorted” by a technology. India also has huge industries such as banking, pharmaceuticals, energy, electricity, communications, and consumption. Software outsourcing did not lose orders or lose value overnight. What has really been repriced by the market is India's most successful and internationally competitive growth model over the past 30 years. India has built itself as a “world office” with English-speaking talent, the number of engineers, and wages significantly lower than those in Europe and the US. Now that artificial intelligence has begun to enter the fields of programming, testing, operation and maintenance, customer service, and data processing on a large scale, India has suddenly discovered that what was once its proudest cost advantage may also become the part most easily replaced by technology. “AI shorting India” is inevitably exaggerated, but it accurately captures an even more important issue. When a country places too much hope for growth, employment, and the middle class on the same industrial circuit, a technological paradigm shift may evolve from industry shocks to development anxiety at the national level. Human arbitrage has been a huge software industry in India for a long time. The National Association of Software and Service Enterprises of India predicts that in the 2025-2026 fiscal year, India's IT industry revenue will reach US$315 billion, an increase of 6.1% over the previous year, and the number of employees will increase to 5.95 million. According to data released by the Indian government, IT and related services revenue for the 2024-2025 fiscal year was US$283 billion, and there are also more than 1,700 global competency centers across the country, employing about 1.9 million people. As a result, it is inaccurate to describe India's software industry as completely collapsing. It is still growing, has a large number of international customers, and has decades of project management capabilities, customer relationships, and industry experience. The transformation of core systems of financial institutions, databases of multinational enterprises, government information platforms, and highly complex legacy systems cannot all be completed with just a few AI agents. However, the capital market is more concerned about future growth prospects. The real problem with software outsourcing in India is that “revenue growth” and “manpower growth” are being decoupled. Their pattern in the past was very clear. European and American companies hand over standardized development, testing, operation and maintenance, data entry, and customer service to India. Indian companies charge according to the number of engineers invested and working hours. The more people a project requires, the bigger the bill the service provider can pay. The most direct way for an enterprise to increase revenue is to recruit more engineers and then undertake more projects. Although “selling people” doesn't sound decent, it is an underlying mechanism for the expansion of India's software services industry. Generative artificial intelligence breaks this cycle. Coding, debugging, and documentation work that used to require dozens of junior programmers can now be completed by a small number of senior engineers using AI tools; software testing, data collation, and customer Q&A, which originally relied on a large number of manual tasks, are also increasingly being taken over by automated systems. Customers are beginning to shift from buying hours to buying results, and are no longer willing to pay for a huge offshore team for a long time. The impact of this change is very special. Even if the order amount does not drop immediately, the number of people required for the same order may be drastically reduced. AI has improved delivery efficiency while simultaneously reducing billable labor hours. For product-based companies such as Microsoft and Google, increased efficiency usually means increased profits; for Indian outsourcers that charge per hour, increased efficiency may first mean shrinking bills. Technological advances have created a conflict of interest within the business model here. Of course, software companies in India can also use AI, but the more effectively they use AI, the faster traditional human outsourcing business shrinks. If they refuse to use it, they will also be defeated by European and American consulting firms and new service providers that use AI. Businesses must choose between weakening their old business and losing their future competitiveness. Therefore, the target of market shorting is mainly the old valuation logic of software outsourcing in India. This...

16d agoWendy#AI #India

Samsung Lee Jae-rong and OpenAI are reportedly discussing cooperation plans in the field of AI and semiconductors

Comparative news, according to Kim Joo's report, Samsung Electronics Chairman Lee Jae-rong and OpenAI founder Sam Altman met at OpenAI's headquarters in San Francisco to discuss cooperation plans in the field of artificial intelligence and semiconductors. OpenAI announced on the 26th. On the morning of the 25th local time, Lee Jae-rong and Altman held talks at their San Francisco headquarters. OpenAI did not disclose the details and topics of the negotiations, but industry observers judged that the two sides are likely to deepen cooperation on AI infrastructure such as high-bandwidth memory, dynamic random access memory (DRAM), and advanced foundry. The two sides may also have discussed Samsung's digital transformation plan for implementing generative artificial intelligence across all business lines.

27d ago

Norway plans to almost completely ban elementary school students from using generative AI tools to prevent impact on learning

According to Reuters, Norwegian Prime Minister Jonas Gahr Stoere said on Friday that Norway will implement an almost complete ban on the use of generative artificial intelligence (AI) tools by elementary school students while limiting the use of AI by older students in education to avoid its negative impact on learning. Faced with the general decline in educational testing results, the Norwegian government banned the use of smartphones by students in schools in 2024 and re-gave teachers more power to maintain classroom discipline. Stoere said at a news conference on Friday that using AI could cause younger students to skip some critical steps in the learning process. At the same time, he said that the new regulations will be officially implemented from the new school year beginning at the end of August. According to guidelines published by the Norwegian government: in principle, AI tools are not allowed to be used at the elementary school level (ages 6 to 13); in middle school (ages 14 to 16), AI tools can be used carefully under the supervision of teachers; and high school students (17 to 19 years old) should learn how to properly use AI to prepare for further education and entry into the workplace in the future. The Norwegian government believes that although AI has potential value, for young students, relying on AI too early may weaken basic abilities such as reading, writing, calculation, and independent thinking, so it hopes to ensure the quality of learning while cultivating the AI application capabilities needed in the future through age-specific management.

63d ago

Like psychotherapy, Bloomberg reveals Anthropic's cultural interview details

Comparative news, according to monitoring, Bloomberg reports that generative artificial intelligence giant Anthropic is implementing an extremely controversial and extremely difficult interview process, strictly prohibiting job seekers from using artificial intelligence tools to cheat, and introducing cultural interviews known as psychological diagnosis and treatment. With the rapid expansion of the workforce in nearly half a year, CEO Dario Amodei revealed that he would spend nearly 1/3 to 40% of his time building culture. This cultural interview evaluation has a single veto power. The core purpose is to select job applicants who are highly suited to safety missions and have the ability to think independently. Candidates are required to pass up to 5 rounds of rigorous screening, including signing a Non-Disclosure Agreement (NDA). In interviews, it is strictly prohibited to use large language models to assist, unless explicitly authorized. In cultural interviews where the elimination rate is extremely high, the interviewer will use the continuous pressure method of quick question and answer to obtain evaluation signals, or even directly interrupt the response when there is no more incremental information. Some candidates revealed that the interviewers did not have any facial expressions or feedback when asking questions, putting tremendous psychological pressure on job seekers. The core level of the interview is to assess the candidate's values and perception of the cutting edge of the destructive threat of artificial intelligence. The interviewer will throw out classic professional ethics dilemmas and interrogate the candidate's emotional feelings at the moment of conflict, coping measures, and afterthoughts, with the aim of evaluating the real moral struggles of the job seeker. Additionally, candidates also need to answer what unusual beliefs they have held to and how to defend their beliefs in difficult times. President Daniela Amodei pointed out that the assessment does not pursue specific beliefs, but rather assesses whether candidates can rationally debate and focuses on finding job seekers who have health doubts about the company.

86d ago

Powell: There will be no situation where interest rate cuts are needed in the short term

Comparative news, according to Kim's report, Federal Reserve Chairman Powell said that generative artificial intelligence tools will have a positive impact on productivity growth in the next few years, but their impact on inflation needs to be carefully assessed. He pointed out that the current boom in artificial intelligence data center construction in the US is driving up the prices of many goods and services, which may push up the level of inflation to a certain extent. Powell stressed that there will be no immediate interest rate cuts in the short term.

157d ago

Hong Kong: Concerned about the potential risks of OpenCLAW, it is recommended that relevant authorities take adequate security measures

Comparative news, according to the Securities Times, the download and usage of the OpenClaw (“crayfish”) application has been booming recently. In response, People's Financial News learned from the Hong Kong Digital Policy Office, which is responsible for AI policy, that the Hong Kong Digital Policy Office has been continuously monitoring the latest trends in artificial intelligence, and has recently been concerned about potential risks related to OpenClaw, including excessive authority, data leakage, and system security. It is recommended that relevant units and individual users take adequate security measures when deploying and applying OpenClaw, including: strengthening network control and strictly isolating the operating environment to reduce the risk of excessive permissions; strengthening credential management to avoid storing keys in clear text format in environment variables; strictly managing plug-in sources to ensure the trustworthiness and security of plug-ins; continuing to pay attention to officially released patches and security updates, and promptly carry out version updates and install security patches. According to reports, the Hong Kong Government attaches great importance to the governance and risk prevention of artificial intelligence applications, and has formulated documents such as the “Ethical Framework for Artificial Intelligence” and “Guidelines for Generative Artificial Intelligence Technology and Application in Hong Kong”. At the same time, the Government has formulated a comprehensive set of “Government IT Security Policies and Guidelines” for all departments to abide by and use. Each department must conduct a risk assessment before installing any type of software.

163d ago

AWS and Ripple are exploring Amazon Bedrock AI for XRP Ledger

Comparative news, according to market news, people familiar with the matter revealed that Amazon Web Services (AWS) and Ripple are studying the use of Amazon Bedrock's generative artificial intelligence (Gen-AI) capabilities to improve monitoring and analysis methods for XRP Ledger (XRPL). The two companies plan to apply artificial intelligence analysis to XRPL's system logs to reduce the time required to investigate network issues. According to some internal assessments by AWS engineers, processing processes that used to take days can now be completed within 2 to 3 minutes. This work aims to address the long-term operational challenges caused by massive C++ logs in XRPL's global network of nodes.

226d ago

2025 Asset Recovery: Why will Bitcoin significantly outperform gold and US stocks?

When watching Bitcoin's performance in 2025, many people fall into simple price comparisons and don't understand why it outperformed US stocks led by Nvidia, or even the traditional safe-haven asset gold. If you look at it from a high-dimensional perspective, this is actually a question of physics and information theory. The price is only an indication; the energy flow and information density behind it are the essence. 1. The crowding effect of energy arbitrage: The transfer of hegemony in computing power In Musk's logic, value is often linked to energy conversion efficiency. Over the past decade, Bitcoin was the only machine that could transform energy into a digital scarce asset on a large scale, a thermodynamics-based value anchor. But from 2024 to 2025, an extremely strong competitor has emerged: generative artificial intelligence. The core driver of US stocks now is not fiat inflation, but the exponential explosion of total factor productivity (TFP) brought about by AI. When tech giants invest hundreds of billions of dollars to build data centers, they are essentially snatching up global electricity quotas. At this stage, the economic added value generated by training the next generation of large models or driving high-performance computing power chips temporarily exceeds the revenue generated by hash collisions to generate Bitcoin. The difference in marginal revenue forms price and capital choices. If you don't believe me, check how many Bitcoin mines have been transformed into AI computing power centers. Capital is profitable and sensitive. When the growth curve of silicon-based intelligence is steeper than the scarcity curve of “digital reserves,” excess global liquidity will prioritize productivity assets with nonlinear growth potential over simple digital assets. 2. Gold's “atomic properties” and Bitcoin's “code consensus” The strong performance of gold this year is essentially the result of increased global geopolitical entropy. In the face of de-globalization and systemic uncertainty, sovereign players need an asset that does not require an internet connection and does not rely on any clearing system. Under this extreme logic of system failure prevention, ancient gold provides atomic-level certainty. Although Bitcoin is known as digital gold, it still relies heavily on internet infrastructure and centralized liquidity channels. When the system faced the risk of being cut off at the physical level, atomic-level certainty overcame Bits' consensus in the short term, and physical gold could at least be held in hands or put into a cave. Gold hedges against the collapse of the system, and Bitcoin is currently viewed by the market more as an overflow of system liquidity. 3. The “volatility damping” tool brought by ETFs determines behavior. The popularity of Bitcoin spot ETFs marks the official domestication of this beast. After Bitcoin entered the traditional asset allocation portfolio, it began to follow the risk control model of traditional finance. Although this brought long-term financial support, it also greatly smoothed out its volatility and stifled its explosive power. Today, Bitcoin is becoming more and more like a high beta technology index. Since the Federal Reserve has maintained high interest rates for longer than market expectations, this “long-tail asset” that is extremely sensitive to liquidity will naturally be suppressed. 4. Productivity singularities siphon the Bitcoin narrative Charlie Munger emphasized opportunity costs. If AI leaders with monopoly positions can achieve highly deterministic nonlinear growth, then the opportunity cost of holding Bitcoin, which does not generate cash flow, becomes extremely high. 2025 is the eve of a rare productivity singularity in human history, and all the money is chasing the node that could generate superintelligence. As a “challenger to the monetary system,” Bitcoin's appeal was diluted in the short term in the face of a productivity revolution narrative. 5. The phase transition adjustment period in a fractal structure is in the AI-driven parabolic acceleration phase from the perspective of a complex system. In fractal geometry, tiny structures continuously replicate and amplify themselves through simple iterative formulas. AI is playing the role of this iterative operator. From Nvidia's computing power at the bottom, to cloud services at the middle level, to software applications at the top, every layer replicates the logic of a “productivity explosion.” This structure is extremely grand, but it also means that the system is approaching the physical limits of this local dimension. The performance of gold in the collapse of the old order can be understood by the construction process of the Cantor Set (Cantor Set), which continuously removes one-third of the middle. Excluded from the current global financial fractal are “credit expansion,” “unfulfilled promises,” and “high entropy debt.” As the old order continues to be torn apart by debt crises and geopolitical turmoil, the last group of unconnected yet indestructible points is gold. This is a kind of value density generated by “subtraction,” and it is the most stable physical bottom layer in a fractal structure. Bitcoin's current state is essentially a hedging result of forces on different scales: profit pressure from early participants, with sovereign countries and long-term...

242d agoLuxurytracy#2026 topics #Bitcoin #US stocks #gold