Search AI Products and News

Frustrated by the slow processing speed of large language models on long documents? Researchers from Tsinghua University have unveiled a groundbreaking technology – the APB parallel inference framework – that dramatically accelerates processing. Benchmark tests show this technology achieves a 10x speed improvement over Flash Attention when handling ultra-long texts. With the rise of models like ChatGPT, AI's ability to process vast amounts of text (hundreds of thousands of words) has increased significantly. However, this often comes at the cost of processing speed...

DeepSeek launched an official account on Zhihu and published a technical article, "DeepSeek-V3/R1 Inference System Overview," detailing its model inference system optimization and cost-profit margin information for the first time, marking the end of the highly anticipated DeepSeek Open Source Week. The article highlights two key optimization goals for the DeepSeek-V3/R1 inference system: higher throughput and lower latency. To achieve these goals, DeepSeek employs large-scale cross-node expert parallelism (EP) technology, although...

Former Twitter CEO Parag Agrawal has recently launched his artificial intelligence startup Parallel Web Systems. The company's official website shows that it is developing a system that enables advanced AI to collaborate with web pages. According to The Information, Agrawal has raised around $30 million for his startup, with investors including venture capital firm Khosla Ventures.

Amazon Web Services (AWS) has launched a new High Performance Computing (HPC) management service - AWS Parallel Computing Service, aimed at simplifying the process for businesses to access and utilize supercomputing resources, lowering the barrier to entry. This service utilizes the open-source tool Slurm to manage clusters, allowing users to easily set up and manage clusters of Amazon Elastic Compute Cloud (EC2) instances on AWS without complex system administrator support. Businesses only need one AWS account to experiment and test the advantages of their workloads in large-scale computing.

In a copyright lawsuit against The New York Times, OpenAI has requested access to the newspaper's internal research materials and documents in order to evaluate the validity of claims that ChatGPT copied content. The New York Times has opposed this request, arguing that it could pose a threat to press freedom. The core focus of the lawsuit is the scope of copyright assessment, with The New York Times insisting it only pertains to published articles, while OpenAI's lawyers are interested in internal documents to verify the accusations. The lawsuit also hints at future trends in AI training methods, which may increasingly rely on high-quality data rather than sheer volume.

Latest News! Version 0.2 of Ollama has been released! It has been reported that this update enables concurrency by default, allowing Ollama to handle multiple requests simultaneously, delivering a faster user experience. This update not only unlocks the feature of concurrent requests but also supports loading different models simultaneously, making Ollama more efficient in handling various tasks.According to the official news released by Ollama, this update enables Ollama to handle multiple chat

YouTube, the world's largest video platform, has introduced a new mechanism that allows people to request the removal of content generated through AI that mimics their appearance or voice. This mechanism is an extension of the relatively lenient regulation of the technology so far.Although this mechanism was quietly added to YouTube's privacy guide updated last month, it was only noticed by TechCrunch this week. YouTube views the use of AI technology to "change or create synthetic content that l

American AI researcher Stephen Thaler claims that his AI DABUS has sentience and that its creations should be granted copyright. The court dismissed Thaler's claims, reaffirming that copyright requires a human creator and that AI does not possess copyright. Thaler's representative, Professor Ryan Abbott, argued that to encourage AI to create things like vaccines, its creations should be eligible for patent protection.