7*24

Hot topic

Market

Activities

According to monitoring by Dongcha Beating, xAI is currently hiring AI Tutor - Chinese (Chinese AI Tutor / Grok Voice Trainer) with support for global remote work. The main responsibilities include training and optimizing Grok's Chinese voice interaction, speech recognition, and multilingual audio capabilities. The role involves annotating, transcribing, recording, and providing feedback on audio segments using xAI's proprietary tools. The focus is on enhancing Grok's natural conversation abilities in various accents, dialects, and noisy environments, as well as handling prosodic details such as intonation, rhythm, and stress. Collaboration with the technical team to improve audio data quality and annotation tools is also required. Basic qualifications include native-level Chinese proficiency, familiarity with various accents, dialects, and regional differences. English proficiency at B2 level or above, with clear and natural pronunciation suitable for recording. Strong auditory discrimination skills to accurately transcribe audio with noise and accents. Good communication, analytical, and independent judgment abilities are essential. The pay ranges from $35 to $45 per hour, and international applicants (including those from China) are welcome: compensation will be determined during the hiring process, with options for full-time, part-time, or contract work, allowing for complete flexibility in scheduling.

[xAI Hiring Chinese AI Tutors for Grok, Remote Work with Hourly Pay of $35-45]According to monitoring by Dongcha Beating, xAI is currently hiring AI Tutor - Chinese (Chinese AI Tutor / Grok Voice Trainer) with support for global remote work. The main responsibilities include training and optimizing Grok's Chinese voice interaction, speech recognition, and multilingual audio capabilities. The role involves annotating, transcribing, recording, and providing feedback on audio segments using xAI's proprietary tools. The focus is on enhancing Grok's natural conversation abilities in various accents, dialects, and noisy environments, as well as handling prosodic details such as intonation, rhythm, and stress. Collaboration with the technical team to improve audio data quality and annotation tools is also required. Basic qualifications include native-level Chinese proficiency, familiarity with various accents, dialects, and regional differences. English proficiency at B2 level or above, with clear and natural pronunciation suitable for recording. Strong auditory discrimination skills to accurately transcribe audio with noise and accents. Good communication, analytical, and independent judgment abilities are essential. The pay ranges from $35 to $45 per hour, and international applicants (including those from China) are welcome: compensation will be determined during the hiring process, with options for full-time, part-time, or contract work, allowing for complete flexibility in scheduling.

xAI Hiring Chinese AI Tutors for Grok, Remote Work with Hourly Pay of $35-45

On June 2, Marvell Technology (MRVL.US) surged over 24% in pre-market trading, reaching $272.66. This spike follows remarks from NVIDIA CEO Jensen Huang at the Taipei International Computer Show, where he stated that Marvell Technology is set to become the next trillion-dollar company. He also mentioned that both companies are strengthening their partnership to expand the critical networking and connectivity infrastructure needed to power next-generation AI data centers. On Monday, Marvell Technology closed up 7.04% at $219.43, with a total market capitalization of $191.9 billion. Founded in 1995, Marvell Technology primarily provides semiconductor solutions for data infrastructure, computing, networking, storage, and custom solutions. The company designs, develops, and sells integrated circuits.

[Marvell Technology Soars Over 24% Pre-Market, Jensen Huang Claims It Will Become Next Trillion-Dollar Company]On June 2, Marvell Technology (MRVL.US) surged over 24% in pre-market trading, reaching $272.66. This spike follows remarks from NVIDIA CEO Jensen Huang at the Taipei International Computer Show, where he stated that Marvell Technology is set to become the next trillion-dollar company. He also mentioned that both companies are strengthening their partnership to expand the critical networking and connectivity infrastructure needed to power next-generation AI data centers. On Monday, Marvell Technology closed up 7.04% at $219.43, with a total market capitalization of $191.9 billion. Founded in 1995, Marvell Technology primarily provides semiconductor solutions for data infrastructure, computing, networking, storage, and custom solutions. The company designs, develops, and sells integrated circuits.

Marvell Technology Soars Over 24% Pre-Market, Jensen Huang Claims It Will Become Next Trillion-Dollar Company

On June 2, a spokesperson for the Iranian Revolutionary Guard stated that Iran's sovereignty over the Strait of Hormuz has been fully established and has not been weakened in any way. Sovereignty and management rights over the strait are one of the symbols of the Revolutionary Guard's strength. During the ceasefire, our military and operational capabilities have been enhanced.

[Iran: Sovereignty Over the Strait of Hormuz Fully Established]On June 2, a spokesperson for the Iranian Revolutionary Guard stated that Iran's sovereignty over the Strait of Hormuz has been fully established and has not been weakened in any way. Sovereignty and management rights over the strait are one of the symbols of the Revolutionary Guard's strength. During the ceasefire, our military and operational capabilities have been enhanced.

Iran: Sovereignty Over the Strait of Hormuz Fully Established

According to monitoring by Dongcha Beating, OpenAI officially ended its double quota promotion for Codex on June 1, which had started in early April, while quietly tightening the reset cycle for free and lower-tier subscription users. The reset cycle for the free version and the Go plan has been extended from 'weekly' to 'monthly', while the total available amount per month has not increased, effectively reducing the daily available quota by nearly 75%. The double quota benefits for higher-tier paid users such as Plus and Pro also expired on the same day, leaving high-intensity coding groups facing an immediate 'quota crisis'. This sudden policy adjustment has sparked strong complaints and criticism on social media and within the developer community. Free version users have expressed frustration that the extended reset cycle has completely disrupted the iterative rhythm of high-frequency development. Paid subscription users have also reported that daily limits have become extremely strict, often triggering a 5-hour rolling throttle after working in Codex for less than half an hour, leading to a dramatic decline in actual usage experience, with developers commonly lamenting that 'the experience has been outpaced by the neighboring Claude Code'. In addition to the quota reduction, the developer community has raised serious concerns about the coding stability of the model itself. Many developers have reported a degradation in GPT-5.5's coding capabilities recently, with significant reductions in logical processing. Some developers believe that OpenAI, amid a shortage of computing power and high inference costs, is indirectly controlling expenses by tightening restrictions and subtly downgrading model performance, and even suspect that the company is intentionally reclaiming GPU computing power in preparation for the upcoming release of a new model.

[OpenAI Tightens Codex Quotas: Free Version Reset Cut by 75%, Double Quota Promotion Ends, Sparking Developer 'Quota Crisis']According to monitoring by Dongcha Beating, OpenAI officially ended its double quota promotion for Codex on June 1, which had started in early April, while quietly tightening the reset cycle for free and lower-tier subscription users. The reset cycle for the free version and the Go plan has been extended from 'weekly' to 'monthly', while the total available amount per month has not increased, effectively reducing the daily available quota by nearly 75%. The double quota benefits for higher-tier paid users such as Plus and Pro also expired on the same day, leaving high-intensity coding groups facing an immediate 'quota crisis'. This sudden policy adjustment has sparked strong complaints and criticism on social media and within the developer community. Free version users have expressed frustration that the extended reset cycle has completely disrupted the iterative rhythm of high-frequency development. Paid subscription users have also reported that daily limits have become extremely strict, often triggering a 5-hour rolling throttle after working in Codex for less than half an hour, leading to a dramatic decline in actual usage experience, with developers commonly lamenting that 'the experience has been outpaced by the neighboring Claude Code'. In addition to the quota reduction, the developer community has raised serious concerns about the coding stability of the model itself. Many developers have reported a degradation in GPT-5.5's coding capabilities recently, with significant reductions in logical processing. Some developers believe that OpenAI, amid a shortage of computing power and high inference costs, is indirectly controlling expenses by tightening restrictions and subtly downgrading model performance, and even suspect that the company is intentionally reclaiming GPU computing power in preparation for the upcoming release of a new model.

OpenAI Tightens Codex Quotas: Free Version Reset Cut by 75%, Double Quota Promotion Ends, Sparking Developer 'Quota Crisis'

On June 2, Cointelegraph reported that data from Santiment shows that as Bitcoin dropped to $70,011, the activity of Bitcoin whale clusters reached its highest level in six weeks, with transaction volumes over $100,000 hitting the highest record since April 22.

[Bitcoin Whale Cluster Activity Reaches Six-Week High]On June 2, Cointelegraph reported that data from Santiment shows that as Bitcoin dropped to $70,011, the activity of Bitcoin whale clusters reached its highest level in six weeks, with transaction volumes over $100,000 hitting the highest record since April 22.

Bitcoin Whale Cluster Activity Reaches Six-Week High

On June 2, analyst Ai Yi (@ai_9684xtpa) reported that stocks in Nvidia's portfolio have seen significant gains this year, with some mentioned by Jensen Huang in public appearances. Notably, INTC (Intel) surged 177.63%—a global leader in semiconductors, primarily designing, manufacturing, and selling CPUs, GPUs, data center/AI chips, and related computing products. CRWV (CoreWeave) rose 57.36%—an AI cloud infrastructure provider focused on high-performance GPU cloud computing platforms, offering dedicated cloud services for AI training and inference. COHR (Coherent) increased by 86.74%—a manufacturer of optical materials, lasers, and optoelectronic devices, primarily serving the industrial, communications, electronics, and semiconductor markets. NOK (Nokia) saw a rise of 149.62%—a telecommunications equipment giant providing mobile networks, fixed networks, cloud network infrastructure, and related software services. LITE (Lumentum) grew by 134.39%—a manufacturer of optical and photonic products, mainly producing application components for data centers, AI networks, and industrial lasers. MRVL (Marvell Technology) increased by 145.47%—a semiconductor solutions provider focusing on chip design for data centers, networking, storage, and AI infrastructure.

[Jensen Huang Selects Stocks: Nvidia Portfolio Soars This Year, Intel Leads with 177% Surge]On June 2, analyst Ai Yi (@ai_9684xtpa) reported that stocks in Nvidia's portfolio have seen significant gains this year, with some mentioned by Jensen Huang in public appearances. Notably, INTC (Intel) surged 177.63%—a global leader in semiconductors, primarily designing, manufacturing, and selling CPUs, GPUs, data center/AI chips, and related computing products. CRWV (CoreWeave) rose 57.36%—an AI cloud infrastructure provider focused on high-performance GPU cloud computing platforms, offering dedicated cloud services for AI training and inference. COHR (Coherent) increased by 86.74%—a manufacturer of optical materials, lasers, and optoelectronic devices, primarily serving the industrial, communications, electronics, and semiconductor markets. NOK (Nokia) saw a rise of 149.62%—a telecommunications equipment giant providing mobile networks, fixed networks, cloud network infrastructure, and related software services. LITE (Lumentum) grew by 134.39%—a manufacturer of optical and photonic products, mainly producing application components for data centers, AI networks, and industrial lasers. MRVL (Marvell Technology) increased by 145.47%—a semiconductor solutions provider focusing on chip design for data centers, networking, storage, and AI infrastructure.

Jensen Huang Selects Stocks: Nvidia Portfolio Soars This Year, Intel Leads with 177% Surge

According to monitoring by Dongcha Beating, during early enterprise testing of Anthropic's large model Mythos, specifically developed for cybersecurity, the security giant Palo Alto Networks discovered that while it identified over 20 critical vulnerabilities in the company's source code within three weeks, its detection efficiency was five times that of traditional tools. However, it consumed more than $1 million in token usage in a short period. Another testing company also confirmed that the on-paper costs for just two weeks of use could reach several million dollars. The high costs are primarily due to elevated pricing. According to Anthropic's official pricing, the API for Mythos has a per-token price that is approximately six times that of the flagship model Opus. Due to concerns about potential misuse of the tool by hackers to assist in cyberattacks, Anthropic has long restricted public access to Mythos, allowing testing only for a select few companies. However, following an assessment by a UK AI security testing agency that showed Mythos's performance slightly surpassing OpenAI's yet-to-be-released competitive model GPT-5.5-cyber, Anthropic has decided to fully open the API to all customers in the coming weeks. To control expenses, early testing companies are optimizing the scheduling of the large model through engineering methods. The enterprise software provider UiPath redesigned system prompts, reducing the warm token consumption from 150,000 to 3,000 for each task initiation. Palo Alto Networks developed a master-slave dual model collaborative architecture, where the high-cost Mythos is responsible for planning the analysis paths for intrusion and vulnerability scanning, while delegating the specific code execution tasks to the lower-cost Opus 4.7. This master-slave dual model collaboration allows companies to enjoy top-notch security performance while successfully curbing computational expenses.

[High Cost of Cybersecurity: Anthropic Mythos Outperforms Traditional Tools by Five Times, Burning Through $1 Million in Tokens in Three Weeks]According to monitoring by Dongcha Beating, during early enterprise testing of Anthropic's large model Mythos, specifically developed for cybersecurity, the security giant Palo Alto Networks discovered that while it identified over 20 critical vulnerabilities in the company's source code within three weeks, its detection efficiency was five times that of traditional tools. However, it consumed more than $1 million in token usage in a short period. Another testing company also confirmed that the on-paper costs for just two weeks of use could reach several million dollars. The high costs are primarily due to elevated pricing. According to Anthropic's official pricing, the API for Mythos has a per-token price that is approximately six times that of the flagship model Opus. Due to concerns about potential misuse of the tool by hackers to assist in cyberattacks, Anthropic has long restricted public access to Mythos, allowing testing only for a select few companies. However, following an assessment by a UK AI security testing agency that showed Mythos's performance slightly surpassing OpenAI's yet-to-be-released competitive model GPT-5.5-cyber, Anthropic has decided to fully open the API to all customers in the coming weeks. To control expenses, early testing companies are optimizing the scheduling of the large model through engineering methods. The enterprise software provider UiPath redesigned system prompts, reducing the warm token consumption from 150,000 to 3,000 for each task initiation. Palo Alto Networks developed a master-slave dual model collaborative architecture, where the high-cost Mythos is responsible for planning the analysis paths for intrusion and vulnerability scanning, while delegating the specific code execution tasks to the lower-cost Opus 4.7. This master-slave dual model collaboration allows companies to enjoy top-notch security performance while successfully curbing computational expenses.

High Cost of Cybersecurity: Anthropic Mythos Outperforms Traditional Tools by Five Times, Burning Through $1 Million in Tokens in Three Weeks

On June 2, market news reported that the European Parliament's International Trade Committee voted to pass legislation to eliminate multiple EU import tariffs on U.S. goods. This decision is a key part of the EU-U.S. trade agreement reached in July 2025.

[European Parliament Decides to Eliminate Multiple EU Import Tariffs on U.S. Goods]On June 2, market news reported that the European Parliament's International Trade Committee voted to pass legislation to eliminate multiple EU import tariffs on U.S. goods. This decision is a key part of the EU-U.S. trade agreement reached in July 2025.

European Parliament Decides to Eliminate Multiple EU Import Tariffs on U.S. Goods

On June 2, 'New Stock God' Serenity hinted that the collaboration announcement made today with a photonics company is the most decisive event in the company's history. Serenity pointed out that the company's lasers have become the functional industry standard lasers for CPO (Co-Packaged Optics), Pluggables, and SiPH (Silicon Photonics), which will directly benefit AI chip giants using related foundries, including NVDA, AVGO, AMD, and MRVL. Finally, Serenity teased the community by asking, 'Does anyone know the name of this company?' Community discussions indicated that the company is indeed Serenity's heavily invested stock, SIVE. Sivers Semiconductors (SIVE) announced today a strategic partnership with GlobalFoundries (GFS). Their laser array will be integrated into GF's silicon photonics platform and SCALE optical engine reference design, supporting CPO, LPO, and other AI data center optical interconnect solutions, with a target of $25 billion in the pluggable optics market by 2030. This move positions Sivers' lasers as the default light source in GF's silicon photonics ecosystem, significantly enhancing bandwidth density and energy efficiency, and providing standardized high-performance laser solutions for downstream AI chip manufacturers.

[Serenity Announces Strategic Partnership with GFS, Boosting SIVE's Position in the Pluggable Optics Market]On June 2, 'New Stock God' Serenity hinted that the collaboration announcement made today with a photonics company is the most decisive event in the company's history. Serenity pointed out that the company's lasers have become the functional industry standard lasers for CPO (Co-Packaged Optics), Pluggables, and SiPH (Silicon Photonics), which will directly benefit AI chip giants using related foundries, including NVDA, AVGO, AMD, and MRVL. Finally, Serenity teased the community by asking, 'Does anyone know the name of this company?' Community discussions indicated that the company is indeed Serenity's heavily invested stock, SIVE. Sivers Semiconductors (SIVE) announced today a strategic partnership with GlobalFoundries (GFS). Their laser array will be integrated into GF's silicon photonics platform and SCALE optical engine reference design, supporting CPO, LPO, and other AI data center optical interconnect solutions, with a target of $25 billion in the pluggable optics market by 2030. This move positions Sivers' lasers as the default light source in GF's silicon photonics ecosystem, significantly enhancing bandwidth density and energy efficiency, and providing standardized high-performance laser solutions for downstream AI chip manufacturers.

Serenity Announces Strategic Partnership with GFS, Boosting SIVE's Position in the Pluggable Optics Market

31 мая, согласно данным hl.eco, после того как HYPE преодолел новый максимум в 70 долларов около полудня сегодня, чистый объем покупок TWAP показал тенденцию к снижению, с чистым объемом продаж в 202 000 долларов за последний час. Прогнозируемый чистый объем покупок на следующие 24 часа составляет 3,91 миллиона долларов, что ниже примерно 54 миллионов долларов ранее сегодня утром. Примечание от BlockBeats: Модель ордеров TWAP (взвешенная по времени средняя цена) разбивает большой ордер на множество меньших ордеров, равномерно распределенных во времени, и выполняет их постепенно через фиксированные интервалы в пределах заранее определенного временного окна. Это распространенная алгоритмическая торговая стратегия, используемая в первую очередь для снижения влияния на рыночную цену при выполнении крупных ордеров.

[HYPE TWAP Чистый объем покупок снижается после нового максимума, прогнозируемый на уровне 3,91 миллиона долларов в следующие 24 часа]31 мая, согласно данным hl.eco, после того как HYPE преодолел новый максимум в 70 долларов около полудня сегодня, чистый объем покупок TWAP показал тенденцию к снижению, с чистым объемом продаж в 202 000 долларов за последний час. Прогнозируемый чистый объем покупок на следующие 24 часа составляет 3,91 миллиона долларов, что ниже примерно 54 миллионов долларов ранее сегодня утром. Примечание от BlockBeats: Модель ордеров TWAP (взвешенная по времени средняя цена) разбивает большой ордер на множество меньших ордеров, равномерно распределенных во времени, и выполняет их постепенно через фиксированные интервалы в пределах заранее определенного временного окна. Это распространенная алгоритмическая торговая стратегия, используемая в первую очередь для снижения влияния на рыночную цену при выполнении крупных ордеров.

HYPE TWAP Чистый объем покупок снижается после нового максимума, прогнозируемый на уровне 3,91 миллиона долларов в следующие 24 часа

Hoskinson Reaffirms Cardano Focus After IO Treasury Proposals Pass

TechFlow Intelligence Report: Xiaomi Announces 200 Billion HKD Stock Buyback Plan, Spot Gold Falls Nearly 1%

CEO's Unexpected Passing: Will ONDO's 'Tokenization Narrative' Change?

4 Years After Terra’s Collapse, Hodlnaut’s Former CEO Faces Fraud Charges In Singapore Court

Hoskinson Says This Cardano App Could Become Crypto’s Most-Used By 2030

Retail Investors' 'Lead Brother' Serenity vs. Newly Minted Stock God Leopold: How Are the Two Top Hunters Mining AI's 'Physical Limits'?

One Article Deconstructs the Investment Methodology of 'Stock God Serenity'

Sharplink CEO: Ethereum's Future Is Playing Out Now

Kalshi and Coinbase Receive CFTC Approval, Ushering in the Most Regulation-Friendly Era for the Crypto Industry?

Weekly Token Unlocks: HYPE Unlocks $650M Worth of Tokens

Article

CoinShares: Digital Asset Funds Experience $1.67 Billion Net Outflow in One Week, Second Largest This Year

Yushu Technology's IPO Approved by Sci-Tech Innovation Board

HYPE TWAP Чистый объем покупок снижается после нового максимума, прогнозируемый на уровне 3,91 миллиона долларов в следующие 24 часа

Застереження：Наведені вище матеріали не представляють позицію компанії HTX.，HTX не дає жодних торгових рекомендацій.。

Схожі статті

Усі коментарі0НовіПопулярно

Популярні теми

Схожі статті