7*24

Hot topic

Market

Activities

According to monitoring by Beating, ByteDance's AI application Doubao has widely launched a task mode, changing the mode options at the top of the app interface to Quick, Expert, and Task, while upgrading the thinking mode to Expert mode. The task mode focuses on the capabilities of intelligent agents, enabling autonomous execution of the entire process from task breakdown, step planning, tool invocation to result delivery. After users input clear objectives, Doubao can automatically perform online searches, content generation, web page creation, or PPT generation in the background. Currently, the core functionalities supported by the task mode include generating a personal portfolio zero-code web page in minutes, one-click complete PPT generation, and automatic analysis and generation of bar or line charts for data visualization after uploading Excel files. The task mode also supports setting scheduled tasks to automatically batch generate reports, organize data, or create images in the background at preset times. Unlike the task mode, which emphasizes execution delivery, the upgraded expert mode focuses more on the quality of thinking. The expert mode utilizes Doubao's large model 2.0 Pro version, employing a native Agent architecture to provide deep reasoning capabilities in complex scenarios such as mathematical reasoning, long text processing, and legal document verification. In terms of pricing, Doubao's basic functions like daily Q&A and ordinary conversations will remain free, while paid benefits are mainly targeted at high-computation professional scenarios such as PPT generation, data analysis, software development, and financial analysis. The pricing for Doubao's professional version is divided into three tiers: the standard version is priced at 68 yuan per month or 688 yuan per year, the enhanced version at 200 yuan per month or 2048 yuan per year, and the professional version at 500 yuan per month or 5088 yuan per year.

[Byte's Doubao Launches Task Mode: Supports Web and PPT Generation, Upgrades Thinking Mode to Expert Mode]According to monitoring by Beating, ByteDance's AI application Doubao has widely launched a task mode, changing the mode options at the top of the app interface to Quick, Expert, and Task, while upgrading the thinking mode to Expert mode. The task mode focuses on the capabilities of intelligent agents, enabling autonomous execution of the entire process from task breakdown, step planning, tool invocation to result delivery. After users input clear objectives, Doubao can automatically perform online searches, content generation, web page creation, or PPT generation in the background. Currently, the core functionalities supported by the task mode include generating a personal portfolio zero-code web page in minutes, one-click complete PPT generation, and automatic analysis and generation of bar or line charts for data visualization after uploading Excel files. The task mode also supports setting scheduled tasks to automatically batch generate reports, organize data, or create images in the background at preset times. Unlike the task mode, which emphasizes execution delivery, the upgraded expert mode focuses more on the quality of thinking. The expert mode utilizes Doubao's large model 2.0 Pro version, employing a native Agent architecture to provide deep reasoning capabilities in complex scenarios such as mathematical reasoning, long text processing, and legal document verification. In terms of pricing, Doubao's basic functions like daily Q&A and ordinary conversations will remain free, while paid benefits are mainly targeted at high-computation professional scenarios such as PPT generation, data analysis, software development, and financial analysis. The pricing for Doubao's professional version is divided into three tiers: the standard version is priced at 68 yuan per month or 688 yuan per year, the enhanced version at 200 yuan per month or 2048 yuan per year, and the professional version at 500 yuan per month or 5088 yuan per year.

Byte's Doubao Launches Task Mode: Supports Web and PPT Generation, Upgrades Thinking Mode to Expert Mode

According to monitoring by Beating, during an internal live-streamed meeting at Meta attended by thousands, an engineer suddenly erupted in profanity, accusing the AI department of making them 'the company's bitch' and demanding to convey to executives that 'they are a piece of shit.' This unexpected outburst left the speaker embarrassed and the live chat flooded with comments. In response to the accumulated grievances from the restructuring, Meta CEO Mark Zuckerberg issued an apology memo on June 12, promising to make changes. The AI department, established in March this year, consists of about 6,500 engineers who were forcibly reassigned with only the options to accept or resign, humorously referred to as 'conscription.' Originally responsible for social applications, these engineers are now compelled to create two unsolvable model problems each week under strict monitoring, with no traces online, and conduct edge testing. The monotonous and mechanical nature of the labeling work has led engineers to feel underutilized, describing their roles as akin to a 'Gulag' concentration camp. The practice of using high-paid engineers for labeling originated from Meta's Chief AI Officer Alexandr Wang. In an April meeting, Zuckerberg mentioned that Wang believes Meta employees are far more intelligent than outsourced workers and that they are more efficient at data labeling. Ironically, after Meta's acquisition of Scale AI last year, the new head was shocked by the forced data labeling for R&D and immediately halted it. With Alexandr Wang taking charge of Meta's lab, the abandoned model has been revived on a larger scale, even leading to the paralysis of some security teams due to personnel being forcibly reassigned. In addition to forced reassignment, Meta has also implemented key monitoring within the company to generate AI data, prompting over 1,600 employees to sign a protest. Meta's Chief Product Officer Chris Cox admitted in an internal meeting that the recent environment is extremely harsh, describing the employee situation as 'running a marathon in hail, suddenly your teammate is swapped out, and the company is also recording you. What the fuck is this?' In the face of the crisis, Zuckerberg promised in the memo to limit the number of direct reports for managers and reiterated that there would be no large layoffs this year. He stated that the AI department is merely a temporary transfer station, and opportunities will be provided for affected employees to be reassigned to more valuable positions.

[Meta Forces 6,500 Engineers into AI Data Labor Amidst Internal Outcry]According to monitoring by Beating, during an internal live-streamed meeting at Meta attended by thousands, an engineer suddenly erupted in profanity, accusing the AI department of making them 'the company's bitch' and demanding to convey to executives that 'they are a piece of shit.' This unexpected outburst left the speaker embarrassed and the live chat flooded with comments. In response to the accumulated grievances from the restructuring, Meta CEO Mark Zuckerberg issued an apology memo on June 12, promising to make changes. The AI department, established in March this year, consists of about 6,500 engineers who were forcibly reassigned with only the options to accept or resign, humorously referred to as 'conscription.' Originally responsible for social applications, these engineers are now compelled to create two unsolvable model problems each week under strict monitoring, with no traces online, and conduct edge testing. The monotonous and mechanical nature of the labeling work has led engineers to feel underutilized, describing their roles as akin to a 'Gulag' concentration camp. The practice of using high-paid engineers for labeling originated from Meta's Chief AI Officer Alexandr Wang. In an April meeting, Zuckerberg mentioned that Wang believes Meta employees are far more intelligent than outsourced workers and that they are more efficient at data labeling. Ironically, after Meta's acquisition of Scale AI last year, the new head was shocked by the forced data labeling for R&D and immediately halted it. With Alexandr Wang taking charge of Meta's lab, the abandoned model has been revived on a larger scale, even leading to the paralysis of some security teams due to personnel being forcibly reassigned. In addition to forced reassignment, Meta has also implemented key monitoring within the company to generate AI data, prompting over 1,600 employees to sign a protest. Meta's Chief Product Officer Chris Cox admitted in an internal meeting that the recent environment is extremely harsh, describing the employee situation as 'running a marathon in hail, suddenly your teammate is swapped out, and the company is also recording you. What the fuck is this?' In the face of the crisis, Zuckerberg promised in the memo to limit the number of direct reports for managers and reiterated that there would be no large layoffs this year. He stated that the AI department is merely a temporary transfer station, and opportunities will be provided for affected employees to be reassigned to more valuable positions.

Meta Forces 6,500 Engineers into AI Data Labor Amidst Internal Outcry

On June 13, Japanese media reported that the Bank of Japan may raise its policy rate from the current 0.75% to 1.0% during its monetary policy meeting scheduled for June 15-16. Kato, president of the Tokyo Research Institute, believes that the rate hike will put pressure on small and medium-sized enterprises and the lives of ordinary Japanese families, and its impact on the yen's exchange rate remains to be observed. For a long time, the core reason for the yen's depreciation has been the significant interest rate differential between Japan and the U.S., leading to the prevalence of 'yen carry trades' where low-interest yen is used to buy high-yield assets. However, following the announcement of the impending rate hike by the Bank of Japan, there has been no significant appreciation of the yen. Kato believes that the market is concerned that the Bank of Japan's future rate hikes will be slow and lag behind.

[BOJ May Raise Policy Rate Next Week]On June 13, Japanese media reported that the Bank of Japan may raise its policy rate from the current 0.75% to 1.0% during its monetary policy meeting scheduled for June 15-16. Kato, president of the Tokyo Research Institute, believes that the rate hike will put pressure on small and medium-sized enterprises and the lives of ordinary Japanese families, and its impact on the yen's exchange rate remains to be observed. For a long time, the core reason for the yen's depreciation has been the significant interest rate differential between Japan and the U.S., leading to the prevalence of 'yen carry trades' where low-interest yen is used to buy high-yield assets. However, following the announcement of the impending rate hike by the Bank of Japan, there has been no significant appreciation of the yen. Kato believes that the market is concerned that the Bank of Japan's future rate hikes will be slow and lag behind.

BOJ May Raise Policy Rate Next Week

On June 13, Zcash founder Zooko Wilcox expressed gratitude to Anthropic for helping protect Zcash users. At the request of Shielded Labs, they conducted a security audit of Zcash using Mythos, and no additional serious vulnerabilities were found in the Zcash protocol. Shielded Labs and other teams continue to advance security enhancement efforts, with updates to be released in the future. Previously, on June 5, Zooko Wilcox disclosed that there was a serious forgery vulnerability in the Zcash Orchard pool, which could be exploited to generate an unlimited amount of undetectable forged ZEC within the Orchard pool. This vulnerability was discovered by security researcher Taylor Hornby during a targeted audit using the Anthropic Opus 4.8 model on May 29 and was reported to the Zcash Open Development Lab (ZODL). ZODL subsequently coordinated an emergency response within the Zcash ecosystem, and relevant fixes were completed by June 2.

[Zcash Founder: Security Audit with Mythos Found No Additional Serious Vulnerabilities]On June 13, Zcash founder Zooko Wilcox expressed gratitude to Anthropic for helping protect Zcash users. At the request of Shielded Labs, they conducted a security audit of Zcash using Mythos, and no additional serious vulnerabilities were found in the Zcash protocol. Shielded Labs and other teams continue to advance security enhancement efforts, with updates to be released in the future. Previously, on June 5, Zooko Wilcox disclosed that there was a serious forgery vulnerability in the Zcash Orchard pool, which could be exploited to generate an unlimited amount of undetectable forged ZEC within the Orchard pool. This vulnerability was discovered by security researcher Taylor Hornby during a targeted audit using the Anthropic Opus 4.8 model on May 29 and was reported to the Zcash Open Development Lab (ZODL). ZODL subsequently coordinated an emergency response within the Zcash ecosystem, and relevant fixes were completed by June 2.

Zcash Founder: Security Audit with Mythos Found No Additional Serious Vulnerabilities

According to monitoring by Beating, European AI startup Mistral AI is in discussions with investors for a new round of financing. Insiders revealed that Mistral AI plans to raise approximately €3 billion (about $3.5 billion), with a projected valuation reaching €20 billion, which may even increase further due to investor demand. This represents a significant doubling from the €11.7 billion valuation during the financing round in September last year. Currently, negotiations are still in the early stages, and specific terms may change. Semiconductor equipment giant ASML is the largest shareholder of Mistral AI, having invested €1.3 billion last year for an 11% stake. Mistral AI was founded in 2023 by former researchers from Google DeepMind and Meta. The company primarily provides cloud infrastructure services to European governments and enterprises, managing local data centers in France and Sweden, and has recently signed cooperation agreements with Airbus and BMW. To expand its business, Mistral AI is negotiating with institutions such as European banks, planning to offer cybersecurity models to replace the Mythos model from Anthropic. Previously, Mistral AI CEO Arthur Mensch publicly pointed out that over-reliance on external sources for vulnerability detection poses national security risks, emphasizing that Europe must independently control related technologies.

[European AI Startup Mistral Negotiates €3 Billion Funding: Valuation May Reach €20 Billion, ASML as Major Shareholder]According to monitoring by Beating, European AI startup Mistral AI is in discussions with investors for a new round of financing. Insiders revealed that Mistral AI plans to raise approximately €3 billion (about $3.5 billion), with a projected valuation reaching €20 billion, which may even increase further due to investor demand. This represents a significant doubling from the €11.7 billion valuation during the financing round in September last year. Currently, negotiations are still in the early stages, and specific terms may change. Semiconductor equipment giant ASML is the largest shareholder of Mistral AI, having invested €1.3 billion last year for an 11% stake. Mistral AI was founded in 2023 by former researchers from Google DeepMind and Meta. The company primarily provides cloud infrastructure services to European governments and enterprises, managing local data centers in France and Sweden, and has recently signed cooperation agreements with Airbus and BMW. To expand its business, Mistral AI is negotiating with institutions such as European banks, planning to offer cybersecurity models to replace the Mythos model from Anthropic. Previously, Mistral AI CEO Arthur Mensch publicly pointed out that over-reliance on external sources for vulnerability detection poses national security risks, emphasizing that Europe must independently control related technologies.

European AI Startup Mistral Negotiates €3 Billion Funding: Valuation May Reach €20 Billion, ASML as Major Shareholder

On June 13, according to Alternative data, the cryptocurrency Fear and Greed Index today stands at 13 (up from 12 yesterday), indicating a slight easing of the market's 'extreme fear' sentiment. Note: The fear index ranges from 0 to 100 and includes the following indicators: volatility (25%), market trading volume (25%), social media activity (15%), market surveys (15%), Bitcoin's dominance in the market (10%), and Google Trends analysis (10%).

[Crypto Fear Index Rises to 13, Slight Easing of 'Extreme Fear' Sentiment in Market]On June 13, according to Alternative data, the cryptocurrency Fear and Greed Index today stands at 13 (up from 12 yesterday), indicating a slight easing of the market's 'extreme fear' sentiment. Note: The fear index ranges from 0 to 100 and includes the following indicators: volatility (25%), market trading volume (25%), social media activity (15%), market surveys (15%), Bitcoin's dominance in the market (10%), and Google Trends analysis (10%).

Crypto Fear Index Rises to 13, Slight Easing of 'Extreme Fear' Sentiment in Market

On June 13, Zcash founder Zooko Wilcox expressed gratitude to Anthropic for helping protect Zcash users. At the request of Shielded Labs, they conducted a security audit of Zcash using Mythos, and the results showed no other serious vulnerabilities in the Zcash protocol. Shielded Labs and other teams are continuing to advance security enhancements, with updates to be released in the future.

[Zcash Founder: Security Audit Using Mythos Found No Serious Vulnerabilities]On June 13, Zcash founder Zooko Wilcox expressed gratitude to Anthropic for helping protect Zcash users. At the request of Shielded Labs, they conducted a security audit of Zcash using Mythos, and the results showed no other serious vulnerabilities in the Zcash protocol. Shielded Labs and other teams are continuing to advance security enhancements, with updates to be released in the future.

Zcash Founder: Security Audit Using Mythos Found No Serious Vulnerabilities

On June 13, leading RWA trading platform MSX officially launched the SpaceX stock token (SPCX.M) for spot trading. Market data shows that SPCX.M reached a peak of $176.52 during trading, with the current price at $166.85, reflecting an increase of approximately 40% from the Pre-IPO subscription price of $119.138. With the launch of SPCX.M spot trading, MSX has completed the distribution of subscription shares from the previous Pre-IPO phase. It is reported that the project was valued at $1.38 trillion during the Pre-IPO phase, with an initial locked sale price of $119.138. This follows the successful launch and share distribution of Cerebras (CBRS.M), marking another core Pre-IPO asset that has completed the full process of 'early subscription, asset launch, and share distribution.'

[MSX Launches SpaceX Stock Token SPCX.M, Hits $176.52 on First Day, Pre-IPO Subscription Shares Distributed]On June 13, leading RWA trading platform MSX officially launched the SpaceX stock token (SPCX.M) for spot trading. Market data shows that SPCX.M reached a peak of $176.52 during trading, with the current price at $166.85, reflecting an increase of approximately 40% from the Pre-IPO subscription price of $119.138. With the launch of SPCX.M spot trading, MSX has completed the distribution of subscription shares from the previous Pre-IPO phase. It is reported that the project was valued at $1.38 trillion during the Pre-IPO phase, with an initial locked sale price of $119.138. This follows the successful launch and share distribution of Cerebras (CBRS.M), marking another core Pre-IPO asset that has completed the full process of 'early subscription, asset launch, and share distribution.'

MSX Launches SpaceX Stock Token SPCX.M, Hits $176.52 on First Day, Pre-IPO Subscription Shares Distributed

On June 13, according to monitoring by Trader T, the US spot Bitcoin ETF recorded a net inflow of $85.82 million yesterday.

[US Spot Bitcoin ETF Sees Net Inflow of $85.82 Million Yesterday]On June 13, according to monitoring by Trader T, the US spot Bitcoin ETF recorded a net inflow of $85.82 million yesterday.

US Spot Bitcoin ETF Sees Net Inflow of $85.82 Million Yesterday

On June 12, analyst Singh stated that with the acquisition of Cursor, SpaceX's positioning as an artificial intelligence infrastructure provider and programming agent is expected to be further solidified through an initial public offering (IPO). He noted, 'We believe that by the end of 2026, xAI's contribution to SpaceX's revenue is expected to increase from the current approximately 17% to over 50%.' Singh pointed out that by 2027, xAI's sales are expected to reach $30 billion, which could support a valuation of $300 billion to $400 billion for its business segment.

[Analyst: xAI's Revenue Contribution to SpaceX Expected to Increase]On June 12, analyst Singh stated that with the acquisition of Cursor, SpaceX's positioning as an artificial intelligence infrastructure provider and programming agent is expected to be further solidified through an initial public offering (IPO). He noted, 'We believe that by the end of 2026, xAI's contribution to SpaceX's revenue is expected to increase from the current approximately 17% to over 50%.' Singh pointed out that by 2027, xAI's sales are expected to reach $30 billion, which could support a valuation of $300 billion to $400 billion for its business segment.

Analyst: xAI's Revenue Contribution to SpaceX Expected to Increase

Claude Opus 4.8 Finds a $4.5 Billion Bug: The AI Era is Mass-Producing Hackers

Worldcoin Falls Terribly as Arthur Hayes Dumps WLD

Want In On SpaceX? Kraken Unveils Early IPO Access Via xStocks

Weekly Token Unlock: HOME Unlocks Over 20% of Circulating Supply

South Korea's KOSPI Plunges 8.37% at Opening, Triggers Circuit Breaker: The 'Two Stocks' That Powered the Bull Market Recoil in a Single Day

Humanity Loses $31 Million in Attack, Token Price Plummets 90% Due to a Single Private Key

Over $31 Million Stolen from Humanity, Is the Team Behind It Paving the Way for a New Project?

Claude's Latest Model to be Released Tomorrow, Anthropic Turns Model Launches into an IPO Language

Just Now, Claude Mythos 5 Released, Handles 50 Million Lines of Code in One Day

SemiAnalysis Report Claims Delay in Two Key Technologies, Triggers Sharp Decline in 'Optoelectronics', Sparking Online Debate Over CPO

Article

Video | Michael Saylor Responds to Controversy Over Selling Bitcoin: 'Never Sell Bitcoin' Refers to Individuals, Not the Company

Analyst: xAI's Revenue Contribution to SpaceX Expected to Increase

Isenção de responsabilidade：O conteúdo acima não representa as posições da HTX.，A HTX não faz quaisquer recomendações de negociação.。

Artigos relacionados

Todos os comentários0Mais recentePopular

Tópicos em destaque

Artigos relacionados