Search AI Products and News

On May 19th, ByteDance's DouBao AI smart-body headphones Ola Friend received a significant upgrade. After this upgrade, users can directly activate the independent English learning agent Owen via voice for real-time conversation, bringing a brand-new experience to English learning. Prior to this upgrade, Ola Friend only supported activation of the proactive agent DouBao. The newly integrated Owen agent is a dedicated independent English tutor tailored for English speaking and learning scenarios. Its question-and-answer logic and interaction mode are closely ...

ByteDance's Jimeng AI has officially launched Jimeng AI 3.0 globally, marking a significant expansion of its text-to-image and video generation technology into international markets. According to AIbase, the new version boasts cinematic image quality, 2K resolution output, hyperrealistic textures, and precise English typography, particularly excelling in English text generation and font control, surpassing the performance of the previous Chinese version. The launch announcement has generated significant buzz on social media platforms. Features can be experienced via the Jimeng AI website and mobile application.

Zhihu AI has officially released CogView4, its latest open-source text-to-image model. Boasting 600 million parameters, CogView4 notably supports both Chinese and English prompts and text-to-image generation, making it the first open-source model capable of generating Chinese characters within images. Its key feature is its support for bilingual prompts, excelling at understanding and following complex Chinese instructions, a boon for Chinese content creators. As the first open-source text-to-image model to generate Chinese characters in images, it fills a significant gap in the open-source landscape.

Dell Technologies is forecasting lower profit margins for its fiscal year 2026, citing increased costs associated with the rising demand for AI servers.

ElevenLabs has unveiled its new Scribe speech-to-text model, boasting unprecedented accuracy levels, particularly in English where it achieves a remarkable 96.7% accuracy rate.

On the first day of DeepSeek Open Source Week, we officially released our latest technological achievement, FlashMLA, an efficient Multi-Layer Attention decoding kernel specifically designed for NVIDIA Hopper architecture GPUs. This technology is optimized for variable-length sequence scenarios and can significantly enhance the inference performance of large models. The core technical features of FlashMLA include comprehensive support for BF16 precision and a paged key-value cache with a block size of 64.

In today's digital world, the use of short text has become central to online communication. However, these texts often lack common vocabulary or context, posing numerous challenges for Artificial Intelligence (AI) during analysis. In response, Justin Miller, an English Literature graduate student and data scientist from the University of Sydney, proposed a novel approach that utilizes Large Language Models (LLMs) to gain deeper understanding and analysis of short texts. Miller's research focuses on how to analyze a vast array of short texts, such as social media profiles,

NVIDIA recently introduced a generative artificial intelligence (AI) supercomputer called Jetson Orin Nano Super, designed for developers to offer a significant performance leap at a more accessible price. Priced at $249, approximately 1814 RMB, it features a 1.7x improvement in generative AI performance. Compared to previous models, Jetson Orin Nano Super achieves a 70% performance increase, capable of performing 67 million INT8 operations per second, or 67 INT8 TOPS.

Teuken-7B, a language model with 7 billion parameters, has now been launched on Hugging Face, supporting all 24 official EU languages. This model was developed by the EU OpenGPT-X research project and is available as an open-source project for users. Unlike most AI language models that are English-centric, Teuken-7B was built from the ground up, with approximately half of its training data sourced from non-English European languages. Image source note: The image is AI-generated.

Entrepreneur Ivan Crewkov from Siberia discovered the immense potential of AI in education after a failed attempt in the smart speaker market, inspired by his daughter's experience learning English. His children's English education AI company Buddy.ai recently raised $11 million in seed funding, led by BITKRAFT Ventures, as its AI technology and business model focused on addressing children's English learning challenges gained recognition from investors. Crewkov's entrepreneurial journey has been quite dramatic.

Recently, a new study has drawn attention, revealing that ChatGPT's responses reflect the cultural values of people from English-speaking and Protestant countries. The core of this study is that large language models like ChatGPT were trained on a vast amount of data from certain countries and cultures, which may lead to outputs that carry cultural biases. The research team, led by René F Kizilcec and his colleagues, examined five different versions of OpenAI's GPT.

The Bilibili UP master 'Geek Bay' has created the 'Daydream' robot, which has quickly gone viral, surpassing 1.13 million views in just four days. This robot can automatically complete assignments, including flipping pages, framing questions, solving problems, and writing, covering multiple subjects. During the development process, the team progressed from modifying a writing machine to independent research and development, employing a CoreXY structure and an integrated AI model, significantly improving the accuracy and efficiency of homework completion. Test results show that the English homework has a high accuracy rate but formatting issues; the Chinese homework features unique handwriting, but some questions are incorrect; the math homework has formatting and logic problems.

Recently, LMSYS Organization's Large Model Arena (Chatbot Arena) released its latest rankings update. LMSYS Chatbot Arena's top-ranked model has surpassed all open-source models, including Llama3-70B, Qwen2-72B, Nemotron-4-340B, Gemma2-27B, and others, earning the prestigious title of the global leader among open-source models. Compared to the 0507 open-source Chat version, DeepSeek-V2-0628 has shown comprehensive improvements in areas such as code and mathematical reasoning, instruction followi