In today's rapid development of Large Language Models (LLM), the training and inference costs of models have increasingly become a focus of research and application. Recently, the Tencent Hunyuan team released an important study that delves into the 'Scaling Laws' of low-bit floating point quantization training, which refers to the principles governing the scale of floating point quantization training. The core of this research lies in exploring how to significantly reduce computational and storage costs without sacrificing performance by lowering the precision of the model.

At the recent CES conference, NVIDIA CEO Jensen Huang stated that the performance improvement speed of the company's AI chips has surpassed the historical standards set by Moore's Law. Moore's Law, proposed by Intel co-founder Gordon Moore in 1965, predicted that the number of transistors on a computer chip would approximately double every year, consequently doubling chip performance as well. However, in recent years, the pace of progress of Moore's Law has significantly slowed. Image source note: The image was generated by AI, image licensed by Midjourney.

With the rapid development of artificial intelligence technology, deepfake content has gradually attracted widespread social attention. To address this emerging issue, California has introduced a series of new laws and regulations at the beginning of the new year, aimed at strengthening the regulation of deepfake content generated by artificial intelligence. These laws will affect social media platforms, content creators, and related industries, marking the entry of AI technology application regulation into a new phase. Note on image source: Image generated by AI, image licensed by Midjourney. One of the important laws expands the scope.

For me, experimenting with AI has become a hobby and a part of my work. I had previously used AI to quickly find recipes, but this time, I decided to take on a challenge to see if it could customize a week's worth of dinners for me, considering whether I was cooking for myself and my wife or dining alone. I consider myself somewhat knowledgeable in cooking, although my skills are nowhere near my wife's (whose family has appeared on a cooking show). Nevertheless, I was amazed by the new culinary possibilities that AI (this time ChatGPT) opened up for me, even though there were some 'traps' to be aware of.

The Italian Data Protection Authority (GPDP) recently issued a warning to local media group GEDI, indicating that its collaboration agreement with OpenAI, the parent company of ChatGPT, may violate EU privacy protection laws. According to this agreement, GEDI plans to provide its Italian language news content to ChatGPT subscribers, while OpenAI may use GEDI's news archives to enhance the accuracy of its products. The GPDP stressed that GEDI's digital archive contains a significant amount of personal information, including numerous details and even sensitive data.

Recently, a lawyer in Texas was penalized for using AI-generated fake cases and citations in court documents. This incident has reignited concerns over the use of AI tools by lawyers in legal work. The case involved a wrongful termination lawsuit against Goodyear Tire & Rubber Company. Image source note: Image generated by AI, image licensed by service provider Midjourney.

The New York Times and the Daily News encountered an unexpected twist in their copyright lawsuit: an OpenAI engineer inadvertently deleted virtual machine search data that could have been key evidence, adding a dramatic turn to this high-profile legal dispute. According to a letter submitted to the U.S. District Court for the Southern District of New York on Wednesday night, lawyers and technical experts for the two media companies had previously invested over 150 hours searching OpenAI's AI training dataset. However, on November 14, an OpenAI engineer accidentally deleted data stored on the virtual machine.

Recently, researchers from ByteDance Research Institute and Tsinghua University jointly released a new study, pointing out that current AI video generation models, such as OpenAI's Sora, while capable of creating stunning visual effects, have significant flaws in understanding basic physical laws. This study has sparked widespread discussion on the capabilities of AI in simulating reality. The research team tested AI video generation models under three different scenarios: predictions under known patterns and predictions under unknown patterns.