Tanjic Technology Launches Sales Large Model SaaS Product 'SalesGPT'


The Alibaba Tongyi Qianwen team has launched two lightweight models in the Qwen3-VL series, with parameter scales of 4B and 8B. This series is the strongest family of vision-language models to date, adding small-parameter versions to lower deployment barriers while maintaining strong performance. Each scale offers two versions: instruction following and chain-of-thought reasoning, providing developers with more flexible options.
Alibaba releases Qwen3-VL compact vision-language models (400M/800M params), enhancing STEM reasoning, visual QA, OCR & video understanding, matching large models' performance.....
Coco Robotics partners with UCLA Prof. Bolei Zhou to establish a physical AI lab, appointing him as Chief AI Scientist to enhance robotics automation research. The 2020 startup focuses on last-mile delivery, initially relying on manual remote operation.....
Google Meet has introduced an AI makeup filter feature, offering 12 virtual makeup options to help users enhance their appearance during video conferences. This move aims to compete with similar features already launched by competitors such as Microsoft Teams and Zoom, enhancing market competitiveness. Users can find this feature under the "Appearance" settings in "Portrait Editing." The feature has been available since 2023 and continues to optimize virtual beauty effects.
AI pioneer Hinton presents a controversial view in an interview: current AI systems may already possess some form of subjective experience, but have not yet developed self-awareness. He emphasizes that the key lies in human misunderstanding of the nature of consciousness, rather than whether AI has awareness. At the same time, he reviews the development of AI from simple keyword matching to the present.