Pinterest reports Q3 revenue up 17% YoY to $1.05B, MAUs up 12% to 600M, vs. 590M est., but forecasts Q4 revenue below est.; PINS drops 17%+ after hours (Jonathan Vanian/CNBC)

Shares of Asian AI-linked companies have dropped significantly, following a similar trend in the US market. Investors are expressing concerns over high valuations, leading to a notable decline in stocks like SoftBank, which fell over 13%, and Advantest, which dropped more than 8%.

Match Group reported a slight increase in Q3 revenue, reaching $914 million, but fell short of estimates. Net income rose by 18% to $161 million, yet the number of paying users decreased by 5% to 14.5 million. Looking ahead, the company forecasts Q4 revenue to be below expectations.

Coupang Inc., the South Korean online retailer, has reported third-quarter profits that significantly missed Wall Street's expectations, even though it managed to outperform competitors like Alibaba in terms of revenue.

Pinterest Inc. has reported disappointing revenue projections for the holiday quarter, falling short of analysts' expectations. This raises concerns about the growth of its advertising business, especially with the holiday shopping season approaching. As a result, the company's stock plummeted over 19% in after-hours trading.

Perplexity AI has accused Amazon of bullying after receiving a strong legal threat that demands the Comet browser stop making purchases on behalf of users. This situation highlights ongoing tensions in the tech industry regarding user autonomy and corporate control.

Shopify has reported a remarkable 32% year-over-year increase in Q3 revenue, reaching $2.84 billion, surpassing estimates of $2.76 billion. The company's gross merchandise volume (GMV) also rose by 32% to $92 billion. Looking ahead, Shopify anticipates Q4 revenue growth in the mid-to-high 20% range, exceeding the expected 23.4%. This strong performance highlights Shopify's resilience and growth potential in the competitive e-commerce landscape, making it a key player to watch as the holiday season approaches.

Palantir has reported a remarkable 63% year-over-year increase in Q3 revenue, reaching $1.18 billion, surpassing estimates of $1.09 billion. The company's revenue from the US government also saw a significant rise of 52%, totaling $486 million. This strong performance not only highlights Palantir's growing influence in the tech sector but also sets a positive tone for future quarters, as the company forecasts Q4 revenue to exceed expectations. Such growth is crucial as it reflects the increasing demand for data analytics and software solutions in both government and commercial sectors.

OpenAI has announced a significant partnership with AWS, marking a major step in cloud computing. By leveraging AWS infrastructure and utilizing hundreds of thousands of Nvidia GPUs, OpenAI aims to enhance its capabilities and efficiency in processing workloads. This collaboration not only boosts OpenAI's operational power but also highlights the growing importance of cloud services in AI development, making it a noteworthy development for both companies and the tech industry.

Recent advancements in LLM Multi-Agent Systems are making it easier to manage numerous tools and sub-agents effectively. The introduction of Tool-to-Agent Retrieval aims to enhance agent selection by providing a clearer understanding of tool functionalities, leading to better orchestration and improved performance.

VidEmo introduces a new approach to understanding emotions in videos, leveraging advancements in video large language models. This innovative method aims to tackle the complexities of emotional analysis, addressing the dynamic nature of emotions and their dependence on various cues.

The iFlyBot-VLA is an innovative Vision-Language-Action model that enhances robotic manipulation through a unique training framework. It features a dual-level action representation and a mixed training strategy, making it a significant advancement in the field.

The MVAFormer introduces an innovative approach to multi-view action recognition, leveraging RGB data and transformer technology to enhance performance. By effectively combining multiple camera views, it addresses challenges like occlusion from obstacles and crowds, paving the way for more accurate human action recognition.

IllumFlow is an innovative framework that enhances low-light images by combining conditional Rectified Flow with Retinex theory. It effectively separates and optimizes illumination and reflectance components, improving image quality even in challenging lighting conditions.

FlowRL introduces a novel approach to reinforcement learning for large language models by matching reward distributions through flow balancing. This method addresses the limitations of traditional reward-maximizing techniques, which often overlook less frequent but valid reasoning paths, ultimately enhancing diversity in model responses.