Several notable pieces of news have recently emerged from the tech circle: The initially highly-anticipated AI hardware Rabbit R1 has fallen short of expectations, revealing the reality that it is merely an Android app. Additionally, many technology companies and communities are undergoing changes, such as Tianya Community preparing to restore access, and Huawei’s Consumer Business Group’s leader, Yu Chengdong, announcing his resignation as CEO. In the field of artificial intelligence, OpenAI has opened up memory capabilities to all ChatGPT Plus users, and ByteDance is conducting internal anti-corruption efforts.
Among these companies’ activities, some events have sparked heated discussions: An AI hardware product, Rabbit R1, that was highly anticipated has been exposed to be running Android internally, and its interface is entirely supported by Android apps. This device, priced at $199, was once considered the beginning of a new era of wearable AI. Technology blogger Mishaal Rahman revealed that the device is in fact an app that can run on Android phones, making the company’s previous high-profile promotion of the so-called Rabbit OS seem somewhat exaggerated.
In addition, Tesla’s leading figure, Elon Musk, continues to adjust the company’s structure, dissolving the Supercharger team, while Apple is secretly assembling an AI team, aggressively poaching from Google. Kuaishou’s Beijing gaming division faced overall layoffs, and Google also went through a round of layoffs and transferred some functions.
In terms of legal copyright issues, there have been lawsuits: OpenAI and Microsoft have been collectively sued by eight US newspapers over copyright issues. Meanwhile, Apple has put an end to the tradition of continuously upgrading the memory capacity in its Mac products, marking the end of a technological era. Additionally, it’s worth noting that the once influential programming language BASIC is celebrating its 60th anniversary.
In recent cybersecurity events, developments are worrisome: Relevant parties disclosed a severe issue that the team not only stored user sessions on their own computers but also did not implement any additional security measures. This practice neglects the privacy rights of users and reflects very poor engineering practices.
Regrettably, for those familiar with the team, this is not new. Just two years ago, they were promoting NFT transactions. On Wednesday, Lyu Cheng responded to the accusation, claiming that the R1 interface is not an Android app and explained that they were invoking large models in the cloud.
Lyu Cheng also elaborated, “Rabbit OS and LAM run in the cloud. Rabbit OS is tailor-made for the R1 device, and we do not support third-party clients. The use of unofficial APK or web clients may pose severe security risks to users, including the threat of data being stolen by hackers. Therefore, it is strongly recommended that users avoid using those unofficial Rabbit OS applications. The new OTA update has implemented some cloud-based verification enhancement measures to ensure the authenticity of device/client requests.”
Several hours later, Rahman expressed on Twitter that his Pixel 6‘s Rabbit OS was unable to connect, which seemed to corroborate Lv Cheng’s statement about new user verification requirements.
Yu Chengdong, a member of Huawei’s board of directors, plays multiple important roles at Huawei, including Executive Director, CEO of the Consumer Business Group, CEO of the Smart Car Solution Business Unit, and Director of the Smart Terminal & Smart Car Parts Management Committee. Although he has stepped down as CEO, as the Chairman of the Terminal BG, Yu Chengdong still holds a significant position in the company’s strategic direction and decision-making. Analysts generally believe that this personnel adjustment is a strategic layout conducted by Huawei in response to changes in the market environment and for future development.
In the field of new technologies, OpenAI recently announced to all ChatGPT Plus users that it would open the “memory” function of chatbots. This feature allows ChatGPT to store and recall past conversation details, reducing the need for users to ask repeated questions. This move will greatly facilitate users working in novel creation, health record tracking, business planning, social media marketing, and education. The memory function is quite demanding on AI’s computing power and memory resources, and in practical applications, it is limited by various factors related to hardware and network latency, but it is particularly helpful for long-term complex tasks.
At the same time, significant advances in the field of AI-generated video have also attracted attention. In February of this year, OpenAI released the advanced video generation model Sora and showcased preliminary video results, sparking a frenzy of AI video creation. Although Sora is not yet available for public testing, some visual artists and filmmakers are able to use it to create content, sparking widespread discussion both within and outside the industry. Visual effects expert Mike Seymour conducted an in-depth interview with producer Patrick Cederberg of the Canadian company Shy Kids about their video work “Air Head” through an article published on fxguide, revealing Sora’s actual role in the video production process. It was learned that although AI’s role in video creation is highly regarded, the creation of this acclaimed short film actually integrated traditional production methods and the efforts of several experts. The process took two weeks, highlighting the irreplaceable position of manual special effects in the production of high-quality content.
In the intersection of art and technology, despite the rapid development of artificial intelligence (AI) technology, perfect visual effects still require a large amount of manual post-production work. The production process is often not as simple as people imagine. For example, when using AI to generate content, the color of balloons may change with each generation, or flaws may appear in the shots. To obtain the desired final effect, a large amount of editing and improvement must be done. Patrick shared his personal experience, illustrating the challenges of achieving high-quality video segments.
Patrick estimated that during the production of a final 90-second video clip for the film, hundreds of 10 to 20-second video clips might be generated. He noted that the ratio from raw footage to the final product could be as high as 300:1. This revelation allows us to understand that currently, the dream of AI generating ideal movie scenes from text with a single click remains elusive.
Recently, Drew Baglino, who was the Senior Vice President in charge of battery, motor, and energy product engineering and technology development at Tesla, decided to resign. It is understood that Tesla is expanding rapidly worldwide, establishing numerous factories. Behind this rapid expansion, some areas have begun to see overlapping roles and job functions. As the company moves towards higher growth targets, it seeks to face future challenges by cutting costs and increasing productivity.
Importantly, at the 2022 shareholder’s meeting, Tesla CEO Elon Musk announced an ambitious plan—to build 10 to 12 gigafactories worldwide, each with an annual production capacity of 1.5 to 2 million vehicles, and to achieve over 100 million deliveries within 10 years. However, signals of challenges on this path have been revealed: in February of this year, Tesla began to delay performance evaluations for some employees, and there were reports of imminent layoffs, with the number of layoffs at one point potentially estimated to be as high as 20%. The news was confirmed by Musk in a recent internal letter, but the layoff percentage was adjusted to 10%.
Meanwhile, things are also turbulent within ByteDance. Recently, its corporate discipline and ethics committee reported 61 cases of illegal and disciplinary violations involving employees who were all fired, four of whom were subject to criminal investigation. These cases covered actions such as employees taking bribes illegally and introducing service providers for personal gain, which not only violated company rules but also crossed the boundaries of the law. In this anti-corruption action, ByteDance showed strong confidence and determination, seriously dealing with the employees involved, canceling benefits such as stock options and annual bonuses, and further making industry-wide announcements.
Amidst these turbulent times, Tesla’s stock price surged unexpectedly. On April 29, Eastern Time, Tesla’s share price soared 15.39% to $194.2 per share, with an added market value of an astonishing $82.6 billion, marking the largest single-day gain since March 2020. Rumors that Tesla would introduce Baidu Maps to support the release of its Full Self-Driving (FSD) technology in China became widespread. Although Tesla did not attend the Beijing Auto Show, CEO Elon Musk paid a surprise visit to Beijing on April 28, followed by Tesla’s declaration of its continued commitment to the Chinese market, collaborating with other companies in AI, electric vehicles, and storage, to promote practical applications of clean energy and autonomous driving technology, transforming beautiful visions into reality.
According to the latest reports, Tesla CEO Elon Musk made a surprise visit to China not long ago. Market analysts believe this move will speed up the launch of Tesla’s Full Self-Driving (FSD) feature in China. The “Tesla App” update shows that the FSD feature, previously labeled as “coming later,” has been changed to “coming soon,” indicating an accelerated process.
Industry observers further pointed out that the introduction of Tesla’s FSD feature could push the commercial application of automated driving technology in China into a new phase. This will help promote the establishment of relevant norms and the further optimization of the entire autonomous driving ecosystem.
At the same time, there have been new developments in the semiconductor industry. After a period of decline in demand, causing chip manufacturers to cut prices to grab orders, market demand began to pick up since the fourth quarter of last year. With the market recovery, several domestic chip manufacturers have announced price adjustments. Among them, Zhejiang Asia-Core Microelectronics Co., Ltd. issued a price adjustment notice on April 23, stating that the price of its entire product line will increase by 15%-20% and take effect immediately. The reasons for the price increase are mainly the significant rise in the price of international precious metals, with some materials even surging by 50%, and the trend is expected to continue, leading to increased production costs. Shenzhen Chuangxin Microelectronics Co., Ltd. also issued a price adjustment notice on the same day, announcing that prices of some products will be raised from May 1st.
On the other hand, Apple Inc. has recently launched a series of new initiatives that competitors should pay attention to. According to reports, Apple has poached dozens of artificial intelligence experts from Google, forming a new team. Since hiring John Giannandrea as senior vice president for machine learning and artificial intelligence strategy in 2018, Apple has poached at least 36 AI experts from Google. This number is significantly higher than the number of AI talents Apple has poached from other companies, highlighting Apple’s emphasis on Google’s AI team. These AI experts who joined Apple work in artificial intelligence and machine learning teams around the world; their addition will further strengthen Apple’s technical capabilities in the development of new artificial intelligence models and products.
Apple has established a secretive “Zurich Visual Lab” in Switzerland. Although its main AI competitors such as Microsoft, Google, and Amazon have been heavily promoting their massive investments in artificial intelligence technology, Apple has maintained a low profile. However, with Apple’s increased investment in AI technology, the outside world’s expectations for its AI products are growing, and industry professionals expect Apple may showcase their new attempts in the field of artificial intelligence at the next Worldwide Developers Conference (WWDC).
This will be an important moment, and the outside world will pay attention to how Apple will showcase its latest achievements and layout in the AI field.
Recently, the news that Kuaishou’s gaming division in Beijing was completely laid off suddenly dominated the top trending headlines. In fact, as early as March of this year, Kuaishou officially announced the dissolution of the Beijing gaming department, with only a very few staff members being able to transfer to other project groups to continue working. The gaming division had undergone several changes since its establishment. In July 2019, they launched “Journey to the West: Soul Aflame” and tested “Code Name: North 3.” Despite the game “Journey to the West: Soul Aflame” having a development process of over three years and an investment of several hundred million yuan, it ultimately failed to break even. According to information from multiple sources, about 70 staff members of the Beijing gaming division’s middle platform team were laid off just two to three months after the game’s launch. Going back even earlier to the beginning of 2023, the Beijing gaming department had already been significantly reduced to a little over a hundred people in an earlier round of layoffs. After such consecutive rounds of layoffs, the Beijing gaming division has essentially been declared dissolved.
Faced with this round of layoffs, Kuaishou has offered two different compensation options for employees to choose from. The first option is a severance payment of N+1 months’ salary, where “N” represents the average monthly income of the employee over the past 12 months, and employees can opt to leave within two weeks without having to go to the company for work during this period. The second option is to provide severance pay of N months’ salary. In addition to this, employees have the right to extend the original two-week departure period by one more month, during which the company will pay for the social security, and similarly, there is no need to work. However, this choice does not include the additional +1 month salary compensation. It is understood that most employees tend to choose the latter scheme because it provides them more time for career planning, to make detailed job-hunting plans, while also avoiding interruptions in social security.
On the other hand, the latest news shows that Google has cut at least 200 key team members in a corporate restructuring. This round of layoffs included transfers of some positions, involving the relocation of certain job sites to areas such as India and Eastern Europe. According to Google’s official website, these core departments are responsible for building the technical foundations of the company’s flagship products and are committed to ensuring the online safety of users. Pankaj Rohatgi, Vice President of Security Engineering at Google, informed his team members in an email, “To achieve better business development, we are expanding our work to other regions around the globe, which will lead to a reduction in some existing positions.” The teams involved include the Python language team, technical infrastructure custodians, security infrastructure, application platform core developers, and other technical engineering positions. Notably, news of the layoff of the Python language team has attracted widespread attention in the developer community. In late April, there were reports that Google decided to disband the entire Python team to focus on the GenAI project, but due to a lack of specific details in the initial announcement, it did not attract much attention. However, subsequent information confirmed that the disbandment of the team was not a rumor but a fact.
On the highly anticipated earnings day, Google’s parent company Alphabet unveiled its financial results for the first quarter of 2023. The report showed that its revenue reached an astonishing $80.539 billion, which is a 15% increase compared to $69.787 billion in the same period last year, marking the fastest growth rate since the beginning of 2022. At this time, Google has steadily entered the ranks of companies with a market value of $2 trillion. In contrast, Meta’s share price plummeted 15% after the earnings announcement, showing a marked contrast. However, what was unexpected is that while Google was steadfastly investing in AI technology, it surprisingly laid off the entire Python foundation team.
The BASIC language celebrates its 60th anniversary. In the early morning of May 1, 1964, at six o’clock, a quiet revolution in computing history was officially launched at Dartmouth College. Mathematicians John G. Kemeny and Thomas E. Kurtz successfully ran the first program of their newly developed BASIC programming language on the General Electric GE-225 mainframe computer. This easy-to-learn programming language, with its interpretive nature, the execution of lines with numbers, and its simple syntax with simple commands like GOTO and easily understood English keywords, was favored by programming beginners. When BASIC landed on personal computers, it promoted the popularization of computing technology, and Paul Allen and Bill Gates also took this opportunity to establish Microsoft. Steve Wozniak also developed an interpreted BASIC interpreter for the Apple computer.
In the world of Ubuntu, the new generation of the distribution Ubuntu 24.10 has been given the codename Oracular Oriole. It is expected to be released in October and will be a highly free and innovative version. Its name reflects Ubuntu’s view on the diminishing importance of naming complexity, traditionally a combination of an adjective and an animal. Ubuntu 24.10 will be equipped with the Linux 6.11 kernel, GNOME 47 desktop environment, and GCC 14.1 compiler, and will provide support for new hardware such as the upcoming Intel Lunar Lake platform and Battlemage graphics cards, AMD Zen 5, and AMD RDNA4 graphics cards.
Eight US newspaper publishers have filed a lawsuit against OpenAI and Microsoft for copyright infringement in a federal court in New York. The publishers allege that the defendants illegally used their copyrighted articles to train artificial intelligence models, including ChatGPT, without paying any fees. In particular, models such as OpenAI’s GPT-2, GPT-3, and GPT-4 are accused of using these publishers’ corpus for model training. By comparison, Microsoft is accused of copying newspaper content for use in search engine indexing or as an information source for AI-assisted tools.
There is concern that the GPT-4 model can even generate content almost identical to the original work in certain contexts. OpenAI has stated that the company was previously unaware of Alden Global Capital’s concerns and is actively communicating with multiple news organizations to explore possibilities for cooperation in order to resolve the issue. Microsoft has not made any statement on this matter. This copyright lawsuit highlights the news publishing industry’s commitment to its business model and its quest for sustainable development in the digital era. However, generative AI technology based on big data may threaten the economic model of traditional journalism, emphasized MediaNews’ Executive Editor Frank Payne, insisting that the free use of their labor results by OpenAI and Microsoft cannot be tolerated.
Even though the base models of Apple’s Mac computers have always maintained 8GB of memory, the company believes that this configuration is sufficient for web browsing, video playback, light editing, and other uses. David Schaub posted an analysis on the social media platform Mastodon, pointing out that since 2012, after Tim Cook took over Apple, Mac memory upgrades have stalled. In the Steve Jobs era, the base memory capacity of Mac was upgraded every two years, but today’s MacBook Pro and iMac have maintained a configuration of 8GB RAM since 2012, and the MacBook Air reached the same memory capacity in 2017.
Apple is confident that 8GB of memory can meet the daily needs of most users and emphasizes that the company is more focused on optimizing hardware and software integration. Especially with the adoption of the SoC architecture, this integrated system-level chip solution aims to improve the overall communication efficiency and performance of the hardware.
The company announced that it is currently focusing on the communication technology of chips and components, committed to enhancing system performance through dense integration of SoC architectures. The company further claims that its 8GB RAM is as efficient as the performance of competitors’ 16GB RAM.
Related News: Reports emerged that Google has recently made extensive layoffs, triggering collective opposition from senior company personnel. They criticized the company’s leadership for lack of vision and also criticized the incompetence of the management, which is continuously expanding. It is rumored that Google has even laid off the entire Python development team, leading to the founder of PyTorch expressing anger publicly: “This is absolutely absurd, core language teams are indispensable!”
On the other hand, Germany has decided to embrace the Linux operating system once again, migrating tens of thousands of computer systems from Windows to Linux. This decision has raised concerns about whether it will repeat the same problems encountered twenty years ago. Additionally, incidents where system vulnerabilities led to hundreds of people being wrongly incarcerated have made everyone reflect on the severe consequences of technological errors. Despite investing up to 280 million yuan for system upgrades, the enterprise has still not been able to successfully migrate to the cloud platform, which has caused it to continue to suffer from issues with software made in Japan.
Technological Trends: In the field of intelligent technology, The China Large Model Quarterly Monitoring Report for Q1 2024 has been officially released. The report covers aspects ranging from the release of large models like OpenAI Sora and NVIDIA Project GR00T to breakthroughs in humanoid robot technology and the birth of coding intelligent agents, marking the beginning of a new chapter in the era of AI large models. With the coexistence of open-source and closed-source models, AI applications in content creation and visual representations have shown tremendous potential, with examples ranging from AI-assisted writing to PPT production tools. The InfoQ Research Center calls on everyone to pay attention to how this wave of AI will shape our future.
