On December 19, 2024, at a press conference, ZhiYuan Research Institute and Tencent announced the launch of LongBench v2, a benchmark designed to evaluate the deep understanding and reasoning capabilities of large language models (LLMs) in real-world long text multi-task scenarios. The platform aims to advance long text models in understanding and reasoning, addressing the current challenges faced by large language models in practical applications.

The WeChat platform has introduced a new feature - 'Author Reading Voice', allowing public account authors to provide voiceovers for their articles using personalized voices. This feature means that readers can click '🎧 Listen to Full Text' to hear the article read by the author in their own voice, enhancing the interactivity and personalization of the reading experience.

Waymo, the autonomous driving division of Alphabet, has announced plans to launch tests of its self-driving car technology in Tokyo in early 2025. This will mark the first time the company's autonomous taxis operate on public roads outside of the United States. The expansion into Japan is part of Waymo's 'Road Trip' initiative, aimed at deploying its technology across multiple cities and conducting field tests, each facing unique challenges. In Tokyo, Waymo’s autonomous taxis will encounter different traffic conditions compared to the U.S., including left-side driving.

Alibaba DAMO Academy, in collaboration with the ModelScope community, recently announced the open sourcing of a new multilingual benchmark testing set, P-MMEval, aimed at comprehensively evaluating the multilingual capabilities of Large Language Models (LLMs) and conducting comparative analysis of cross-language transfer abilities. This testing set covers efficient datasets for both basic and specialized capabilities, ensuring consistency in multilingual coverage across all selected datasets, and provides parallel samples across multiple languages, supporting up to 10 languages from 8 different language families, including English, Chinese, and Arabic.

The video generation feature for Doubao on PC, a product of ByteDance, has entered its beta testing phase, with some users already receiving access. After applying for the beta test, users can visit doubao.com, select the 'Video Generation' feature, upload images, and input prompt words while also adding camera movements and storyboard information to generate video content. During the beta phase, users can generate up to ten videos for free each day.

Leap Star announces that the video generation model Step-Video is now in internal testing, and users can apply for testing qualifications on the Yuewen Video official website. The Step-Video model currently supports generating videos from text (text-to-video), capable of producing high-definition videos up to 1080P resolution and lasting up to 10 seconds. It takes an average of 2.5 minutes to complete a video, but it demonstrates significant advantages in automatic optimization of prompts, aesthetic style performance, and motion control.

At today's Zhipu Agent OpenDay event, Zhipu AI's CEO Zhang Peng announced the latest upgrade of the company's smart agent AutoGLM. This upgrade enables AutoGLM to perform complex operations exceeding 54 steps and to complete tasks across applications, marking an important advancement in agent technology for autonomous task completion.

Microsoft recently released the open-source library TinyTroupe, a groundbreaking AI system that simulates human behavior in virtual environments, providing businesses with innovative product testing and market research solutions. TinyTroupe utilizes large language models such as GPT-4o to create AI agents called "TinyPersons" with unique personalities, interests, and goals. These virtual agents can interact within "TinyWorlds" environments, simulating real user scenarios. Unlike traditional game simulations, Tin

Kunlun Wanwei Technology Co., Ltd. announced the official launch of invitation testing for its 'Tiangong Large Model 4.0' o1 version (Skywork o1). This news marks a significant step forward for the company in the field of general artificial intelligence, aiming to achieve higher-level AI applications through technological innovation.

ESPN is developing an AI-generated virtual character named FACTS for its Saturday college football show SEC Nation. This virtual character will use information from ESPN Analytics, including the Football Power Index (FPI), player and team statistics, and game schedules, to enhance education and enjoyment in sports analysis. Image: ESPN FACTS is described as ESPN's first statistician Howie Schwab.

Kunlun Wanwei Technology Co., Ltd. announced that its latest developed TianGong Large Model 4.0 O1 version (English name: Skywork O1) will launch invitation testing on November 27, 2024. The TianGong Large Model 4.0 O1 version is the first domestic O1 model with Chinese logical reasoning capabilities. It not only incorporates abilities such as thinking, planning, and reflection in its model output but also shows a significant improvement in reasoning capabilities compared to the baseline model in standard evaluation sets, achieving an essential enhancement in the model's reasoning abilities.

Microsoft has recently open-sourced an innovative business validation tool called TinyTroupe. This project, based on multi-agent virtual character simulations, offers enterprises a new low-cost market validation solution. By building a virtual user population, companies can quickly gather user feedback before the official product launch, significantly reducing market trial and error costs. The core advantage of TinyTroupe lies in its use of large language models like GPT-4 as the driving engine. By setting the personality traits, interests, and behavioral patterns of virtual users, the system can simulate...

At the Baidu World Conference held on November 12, 2024, Baidu announced an important innovation in the field of AI creation — 'Free Canvas', a new feature launched in collaboration with Baidu Wenku and Baidu Netdisk. Baidu's founder, chairman, and CEO Robin Li emphasized that 'Free Canvas' is a versatile whiteboard powered by the Wenxin multimodal large model, designed to inspire limitless inspiration and creativity in everyone, allowing anyone to become the protagonist of their own story, a cartoonist, or a short video director. He pointed out that 'Free Canvas' is not a futures product but is available immediately.

ByteDance launched its latest image editing model, SeedEdit, on November 11, becoming the first productized universal image editing tool in China. Users can simply input a straightforward natural language instruction to change the background, transform styles, and add or remove elements in images, without the need for complex tracing and painting. SeedEdit excels at understanding commands, accurately interpreting both Chinese and English prompts, including idioms and proper nouns. For example, a user can just input 'replace the watermelon with grapes,' and SeedEdit...

Recently, scientists have successfully developed a new device called the 'Electronic Tongue'. This device utilizes artificial intelligence technology to accurately differentiate between various coffee blends and has the ability to detect whether food and beverages are about to spoil. With the continuous advancement of technology, this new invention provides a novel solution for food testing and quality control. The 'Electronic Tongue' mimics human taste, using sensors to perceive the chemical components of different substances to determine their flavor. Compared to traditional taste testing methods, the Electronic Tongue not only...

In today's rapidly developing artificial intelligence landscape, the phenomenon of AI 'hallucinations' has become increasingly frequent, causing significant inconvenience for many companies. Customer service chatbots confidently describe non-existent products, financial AIs fabricate market data, and medical bots provide dangerous medical advice. These issues are no longer mere curiosities; they pose serious risks that are affecting company reputation and profitability. To tackle this challenge, San Francisco-based startup Patronus AI has announced the launch of the world's first self-service platform aimed at real-time detection and prevention.

Recently, some media reports indicated that WeChat is conducting gray testing of the AI Q&A feature, allowing some users to input questions in the WeChat search box to trigger the AI Q&A function. The AI Q&A can generate an approximately 400-word response in about 20 seconds and provide 5 reference materials from WeChat official accounts. The test results show that the AI Q&A feature currently mainly cites WeChat official accounts as reference materials.