hotai.newsAn archive of 20 AI news briefs collected on 2026-06-25. Check the original sources for details and updates.
This content still needs to be cross-checked against the original publication page.
Amazon’s latest India investment comes as global tech companies race to expand AI infrastructure in the country.
Source links can be verified on the original siteRead original ↗The current environment is abnormal. After completing verification, you can continue to access. Go to verification
Source links can be verified on the original siteRead original ↗The current environment is abnormal. After completing verification, you can continue to access. Go to verification
Source links can be verified on the original siteRead original ↗The entire solution supports low-code development and rapid deployment, and can be flexibly adapted according to the needs of different industries.
Source links can be verified on the original siteRead original ↗The Authors Guild tested five AI detectors on human-written texts. Pangram and Grammarly correctly identified all of them, while Sidekicker and ZeroGPT flagged human-written articles as AI-generated. But the Guild also warns of a paradox: professionally written texts look statistically similar to AI output because language models were trained on exactly that kind of writing.
Source links can be verified on the original siteRead original ↗Jensen Huang stated that the construction cycle for such infrastructure will span decades, mirroring the development of other critical systems like power grids, transportation networks, and the internet.
Source links can be verified on the original siteRead original ↗On June 24, Beijing Shengshu Technology Co., Ltd. and Shanghai Jidou Technology Co., Ltd. officially reached a strategic cooperation agreement.
Source links can be verified on the original siteRead original ↗Gartner's latest report points out that as Token prices keep rising, by 2028 AI programming cost will exceed average salary of ordinary developers. Billing model shifts from subscription to usage-based billing, making enterprise spending hard to predict, while developers are unlikely to voluntarily reduce usage due to pursuit of efficiency. #AI programming cost# #Gartner forecast#
Source links can be verified on the original siteRead original ↗By 2025, Meta will have already replaced about half of all human moderation requests with large language models and aims to increase that percentage to over 90 percent for certain types of content by the end of the year.
Source links can be verified on the original siteRead original ↗Google appears to be hemorrhaging key AI researchers to competitors.
Source links can be verified on the original siteRead original ↗Chen Zhenkuan, Vice President of Lenovo Group and General Manager of China Infrastructure Business Group, announced a performance growth target: in 2027, Lenovo China Infrastructure Group will anchor a target of 1000 billion yuan.
Source links can be verified on the original siteRead original ↗Meta says the page management platform shows you ‘exactly what to do next to succeed on Facebook.’
Source links can be verified on the original siteRead original ↗Huang Zhenxin stated that the ultimate goal is to explore the limits of intelligence and hopes to compete with those three overseas model companies.
Source links can be verified on the original siteRead original ↗Google has integrated "Computer Use" directly into Gemini 3.5 Flash, letting the model operate computers, browsers, and mobile devices on its own. On the OSWorld benchmark, it scores 78.4, putting it on par with GPT-5.5. Developers can use the Gemini API to build agents for software testing or office automation.
Source links can be verified on the original siteRead original ↗Currently, the model is internally codenamed "U1 Pro", advanced by SenseTime Research Institute, a member of SenseTime Rixin model family. It is expected that this July the model will start internal invite testing and provide services to customers.
Source links can be verified on the original siteRead original ↗Fang Rong revealed that ZTE's new generation AI phone will be launched soon. This is an important step for ZTE to push intelligent capability from cloud to device, will provide an interaction experience of 'understanding and being able to work', thereby driving the growth potential of the entire terminal ecosystem.
Source links can be verified on the original siteRead original ↗Using Microsoft Excel as an example, he pointed out that new technologies often reshape work methods rather than directly eliminating jobs, 'disappearing and changing are two different things. The key is not to only look at the current static picture of the world and then conclude that a certain position will no longer exist and the people doing these jobs will be unemployed. New jobs will appear.'
Source links can be verified on the original siteRead original ↗Schmidt thinks Chinese people are very smart, have already built top AI models with weaker AI hardware, and specifically mentioned Huawei Ascend chips, as well as using various software methods to bypass latency and architecture issues. #Former Google CEO criticizes Chinese AI open source#
Source links can be verified on the original siteRead original ↗Baidu open-sourced and launched the Unlimited OCR model on June 22, with a total parameter count of 3 billion, activating only 500 million parameters during inference, aiming to solve the problem that end-to-end OCR models slow down the more they generate when parsing long documents.
Source links can be verified on the original siteRead original ↗Corning unveiled its glass-based optical interconnect technology 'Glass Bridge' for next-generation AI data centers at a Seoul tech conference, aiming to directly couple photonic chips with optical fibers, simplifying assembly and increasing density. The technology offers a new solution for frontier architectures such as co-packaged optics (CPO), and is expected to address the interconnect challenges brought by AI compute growth. #Corning# #AI Data Center#
Source links can be verified on the original siteRead original ↗