Author: Advanced AI Editor

A survey by Robin AI of 4,152 people across the US and UK found that only 30% at present would trust a legal AI tool ‘to represent them on its own’. Meanwhile, only 10% said legal services are ‘truly accessible to everyone’, and just 23% believe good quality legal services are available to the average person. Also, respondents said they would need a 57% discount to choose a legal AI system over a human lawyer. I.e. the general public doesn’t yet really value the output – or expect to value the output – of legal AI systems. Yet, some people…

C3.ai experienced a 14% increase in its stock price over the past week, likely spurred by its recent collaboration with Arcfield to enhance Enterprise AI solutions for defense and intelligence. This partnership underscores the company’s dedication to advancing AI technology, potentially strengthening its service offerings. Meanwhile, broader market trends showed mixed movements with the Dow Jones down, S&P 500 steady, and Nasdaq gaining, amid tariff uncertainty. As the market saw a 4% rise over the same period, C3.ai’s price movement aligns with the general upward market trend, albeit boosted by the positive sentiment around its recent announcement. We’ve identified 2…

China’s leading tech firms, including ByteDance, Alibaba (HK:9988), and Tencent (HK:0700), have reportedly stockpiled close to one million units of Nvidia’s (NASDAQ:NVDA) H20 AI chips in anticipation of tightened U.S. export restrictions, according to Nikkei Asia. The bulk purchases—valued at over $12 billion—aimed to secure nearly a year’s supply before the April export ban took effect. The H20 chip, a scaled-down version of Nvidia’s high-performance H100, is tailored to meet U.S. export controls and is widely used for AI inference tasks. Demand for AI computing in China is soaring, driven by the rapid growth of AI-powered applications like DeepSeek, now…

The integration of long-context capabilities with visual understanding unlocks unprecedented potential for Vision Language Models (VLMs). However, the quadratic attention complexity during the pre-filling phase remains a significant obstacle to real-world deployment. To overcome this limitation, we introduce MMInference (Multimodality Million tokens Inference), a dynamic sparse attention method that accelerates the prefilling stage for long-context multi-modal inputs. First, our analysis reveals that the temporal and spatial locality of video input leads to a unique sparse pattern, the Grid pattern. Simultaneously, VLMs exhibit markedly different sparse distributions across different modalities. We introduce a permutation-based method to leverage the unique Grid pattern…

Emerging technology, when first being presented to the public-at-large, is bound to get people fearful and protective of the status quo. Such is the current case with AI and its practical application in multiple industries, particularly in the space of visual media arts, but rather than turn away, renowned filmmaker and technological innovator James Cameron is choosing to lean in. Last year, he joined the board of the generative AI company Stability AI and during a recent interview on the “Boz to the Future” podcast, he explained how the tech’s ability to shorten VFX workflow could truly be a game-changer…

One pressing question at the artificial-intelligence (AI) summit in Paris this week was this: is Mistral AI’s assistant a cat, or a chat? Called Le Chat and developed by a French startup as a competitor to ChatGPT, it launched as a smartphone app on February 6th. To the English speaker, Le Chat looks like a French twist on AI chat, which it conducts in English (and other languages). Yet at the jamboree President Emmanuel Macron plugged it using a soft “sh”, rendering Le Chat distinctly feline. Arthur Mensch, Mistral’s 32-year-old boss, says his baby is indeed four-legged. Look carefully at…

Alibaba Group Holding on Tuesday unveiled the highly anticipated third generation of its open-source artificial intelligence (AI) model series, which promises faster processing and enhanced multilingual capabilities, intensifying competition in an already crowded Chinese market.The Qwen3 family consists of eight models, ranging from 600 million parameters to 235 billion, with enhancements across all models, according to the Qwen team at Alibaba’s cloud computing unit. Alibaba owns the South China Morning Post.In AI, parameters are a measurement of the variables present during model training. They serve as an indicator of sophistication: larger parameter sizes typically suggest greater capacity.Benchmark tests cited by…

“In initial benchmarks, both of these models have been measured as being roughly on par with OpenAI and DeepSeek models and slightly behind Grok 3 beta and Google Gemini 2.5 Pro,” said Hyoun Park, CEO and chief analyst at Amalgam Insights. “Baidu’s Ernie 4.5 Turbo Measures up on par with OpenAI’s newest GPT models while being priced much more competitively than OpenAI.” Analysts also point out that Chinese AI models are reaching performance levels comparable to Western counterparts at a fraction of the cost, between 20 and 40 times lower, putting pressure on US firms to speed up innovation and…

At its Create 2025 conference in Wuhan last Friday, Baidu doubled down on its vision that the future of AI lies in creating smarter applications. Speaking to a packed room of developers, Baidu founder and CEO Robin Li said that as foundational models improve, they become a double-edged sword: while they unlock more possibilities for application development, they also risk making apps obsolete if the model itself renders them redundant. He pointed out that the key to building apps that stand the test of time is to “find the right scenario, pick the right model, and learn a little about…

China’s tech champion aims to fill gaps left by US sanctions Huawei is preparing to test its latest AI chip, the Ascend 910D, hoping it can challenge higher-end products from Nvidia as US restrictions batter the Chinese semiconductor industry. According to the Wall Street Journal Huawei has begun approaching Chinese tech firms to test the new chip’s technical feasibility. The Shenzhen giant expects the first sample batches by late May. While the development is still early, it signals a fresh effort to counter Washington’s moves to choke China’s access to advanced chip-making kit. Huawei hopes the Ascend 910D will outperform Nvidia’s…