Recently, the famous programming competition platform Codeforces made an important decision - to prohibit the use of artificial intelligence (AI) in programming contests. With the rapid enhancement of AI programming capabilities, Codeforces founder Mikhail Mirzayanov stated that this decision is necessary as the development of neural networks has reached 'a new height that cannot be ignored'. Image source note: The image is generated by AI, image authorized by service provider Midjourney. This new regulation mainly targets the participants of the competitions.

Recently, a research team from Shanghai Jiao Tong University and Harvard University introduced a novel model fine-tuning method — LoRA-Dash. This new approach claims to be more efficient than existing LoRA methods, particularly in the fine-tuning of specific tasks, achieving the same results with an 8 to 16 times reduction in the number of parameters. This is undoubtedly a major breakthrough for fine-tuning tasks that require substantial computational resources. In the context of the rapid development of large-scale language models, the demand for fine-tuning specific tasks is steadily increasing. However, fine-tuning often

Recently, reports have emerged that the UAE's sovereign investment company MGX is considering participating in a multi-billion dollar financing round for OpenAI. According to The Wall Street Journal, MGX has not yet determined the specific amount of its investment, while OpenAI is seeking to raise $6.5 billion and increase the company's valuation to $150 billion. This financing plan has the support of venture capital firm Thrive Capital, and Microsoft may also be involved, given its engagement since 2019.

A startup team from Mexico City, 'Somewhere Someone', successfully attracted collaboration from Adidas by designing a football team jersey using artificial intelligence, showcasing the integration of technological innovation and traditional craftsmanship. The team created innovative product concept images through AI image generation technology, particularly incorporating traditional Mexican embroidery elements into the design of the Adidas-styled national team jersey, capturing global attention. This design not only enhances the unique appeal of the product but also provides economic independence opportunities for over 3,000 artisans, potentially impacting 15,000 people. Adidas' involvement not only promotes...

Google DeepMind has released AlphaProteo, an artificial intelligence system focused on generating novel proteins tailored to specific target molecules, aimed at accelerating drug design, disease understanding, and health application research. AlphaProteo successfully designed binders for the protein VEGF-A, which is associated with tumor growth and various disease complications, outperforming traditional methods. Unlike AlphaFold, the proteins designed by this system can interact with biological processes, providing possibilities for developing targeted therapies against harmful proteins.

The United States, the United Kingdom, and the European Union have jointly signed the first legally binding AI treaty, the 'AI Framework Convention', aimed at ensuring that AI applications adhere to principles of human rights, democracy, and the rule of law. The treaty requires signatory countries to follow principles that protect user data, respect the law, and operate transparently, implementing these principles through legislation and administrative measures. Although compliance with the treaty primarily relies on monitoring, and enforcement may be weak, it is seen as an important blueprint for countries to formulate AI laws. Currently, the US, EU, UK, and several other countries have signed, and the treaty will come into force in five...

The first international AI treaty will be opened for signing on September 5, covering 57 countries, including the United States, the United Kingdom, and EU member states. This 'AI Convention' aims to address AI risks, protect human rights, and promote responsible innovation. The UK Justice Minister emphasizes that this treaty ensures the rational use of new technologies without compromising fundamental values. The focus is on protecting human rights affected by AI, in contrast to the EU's AI regulations that emphasize market regulation. The treaty has been in preparation since 2019, aiming to implement it through legislative and administrative measures, with the UK government committed to working with all parties.

Canva has recently made significant adjustments to its pricing strategy, increasing prices for both the Canva Teams plan and individual users, leading to discontent among long-time users. The price of the Canva Teams five-member plan in the U.S. has increased from $119.99 to $500, while Australia's five-user monthly fee has risen from $39.99 to $13.50 per person. The annual fee for individual users has also increased from $119.99 to $100. New customers are already subject to the new pricing, but longtime users face pricing pressure. Canva explained that the price adjustment was necessary.

Google has made a major update to the Chrome browser by integrating an AI chatbot named Gemini into the address bar. Users can simply type '@gemini' followed by their query in the address bar to receive a response from Gemini. This integration makes AI a default part of the browsing experience for Chrome's vast user base, eliminating the need for extra navigation. Gemini uses a lightweight version of Google's advanced language model family, Gemini 1.5 Flash, offering significantly improved AI capabilities over its predecessor.

The Google DeepMind team, in collaboration with academic institutions, has developed an innovative method called Generative Reward Model (GenRM) aimed at improving the accuracy and reliability of generative AI in reasoning tasks. GenRM incorporates a validation process into text generation tasks, allowing the model to simultaneously generate and evaluate potential solutions while supporting Chain of Thought (CoT), enhancing the comprehensiveness of the validation process. Compared to traditional methods, GenRM has shown significant advantages in multiple tests, with accuracy improvements ranging from 16% to 64%, particularly in...

The Stepping Star Open Platform Experience Center has launched the image generation model Step-1X, which debuted at the 2024 World Artificial Intelligence Conference, alongside the Step-2 language model and Step-1.5V multimodal model. Partnering with Shanghai Film, Step-1X developed an H5 interactive experience based on the classic IP 'Havoc in Heaven', and despite scarce data resources, it can still generate high-quality images. It emphasizes deep semantic alignment and detail generation, utilizes a self-developed DiT model architecture, supports long instruction input, and is suitable for advertising and gaming.

ComfyUI version 0.2.0 introduces Flux ControlNets technology, including Canny and Union control nets, significantly enhancing image processing capabilities. The optimized queue management features, new 'Go-to Node' for quick navigation, 'Contain' image display, and task output folder view, greatly improve work efficiency. Enhanced node library functionalities such as bookmarks, recursive expand/collapse, and search filters make resource management easier. The user interface and experience have been completely revamped.

The newly launched 1.5 flash model update of Gemini has significantly improved response speed by 50%, especially noticeable in the Pro Plan, benefiting all user tiers. This optimization is evident on both web and mobile platforms, enhancing user interaction experience, particularly when handling complex tasks. The competitiveness of Gemini in the AI assistant domain has thus increased, with a new round of updates scheduled for release on September 4, introducing more enhanced features and continuously optimizing services. This update reflects the rapid iteration of AI technology and fierce competition, allowing users to benefit and enjoy an improved experience.

Baidu Intelligent Cloud showcased its latest advancements in the traffic big model sector at the 2024 China International Big Data Industry Expo. By integrating with Baidu's traffic big data, it significantly enhanced the efficiency of omni-directional signal control, improving traffic flow by 20-30%. In the pilot project in Guiyang's Guanshanhu District, the application of AI signal control technology led to a 15% reduction in average weekday vehicle delay. This system integrates internet and local traffic police data to predict traffic flow distribution within 5 minutes and dynamically adjusts traffic light durations to optimize green light utilization.

The new version of the Playground app significantly enhances the usability and convenience of image editing. Users can easily add, delete, or modify text on posters and adjust their position and style; after uploading images, they can effortlessly remove or add elements through text commands, such as deleting trees or adding new objects. The updated app introduces design requirement categories, including posters, logos, T-shirts, etc., allowing users to choose from templates or create their own designs, greatly increasing design flexibility to meet diverse needs. This update greatly facilitates users in performing detailed image editing tasks.

This article outlines the latest developments in the AI inference chip field, particularly the results of the MLPerf Inference v4.1 competition, which gathered the latest chips from companies such as AMD, Google, UntetherAI, Cerebras, FuriosaAI, and Nvidia. Nvidia's H200 and GH200 chips maintain a lead in the 'Data Center Closed' category, but chips from other companies excelled in specific tasks, especially in terms of energy efficiency. AMD's Instinct Accelerator.

Meitu Inc.'s financial report for the first half of 2024 shows a total revenue of 1.62 billion yuan, a year-over-year increase of 28.6%, with an adjusted net profit of 270 million yuan, representing an 80.3% growth. The increase in revenue and net profit is mainly attributed to the advancement of generative AI technology, particularly in the application of imaging and design products. Subscription revenue reached 930 million yuan, reflecting a year-over-year growth of 54.5%, accounting for 57.4% of total revenue. As of June 30, the number of paid subscribers exceeded 10.81 million, with a penetration rate of approximately 4.2%. Meitu's Design Studio focuses on e-commerce design.

Google has released three experimental AI models from the Gemini 1.5 series, including the smaller variant Gemini 1.5 Flash-8B, which, despite its size, delivers outstanding performance and is suitable for lightweight AI applications. The more powerful Gemini 1.5 Pro model shows significant improvements in coding capabilities and the handling of complex prompts, making it ideal for users who require advanced AI assistance. The Gemini 1.5 Flash model has also been optimized for performance, demonstrating impressive results. These new models are now available for user testing on the AI Studio platform, with Google providing a smooth transition.

As a domestically produced game that has attracted global attention, Black Myth: Wukong achieved sales of 390 million yuan in its first month and surpassed 1.2 million copies sold, setting a new pre-sale record for Chinese games. This game has not only become a huge hit in the Chinese market but has also sold well worldwide, creating a new record for Steam download bandwidth. Its aesthetic effects, rich storyline, and unique charm of Chinese culture have earned high praise, including from IGN China. The game utilizes NVIDIA AI technology, particularly its ray tracing and DLSS 3 technology, to achieve stunning graphics and ultra-high performance.