hotai.newsAn archive of 20 AI news briefs collected on 2026-07-26. Check the original sources for details and updates.
This content still needs to be cross-checked against the original publication page.
For years, the Silicon Valley tech community has had ideological divisions over how to develop AI software. This week, the debate reached a boiling point.
Source links can be verified on the original siteRead original ↗SK Group and Nvidia reach AI cooperation exceeding $500 billion, plan to build a 2 gigawatt-capacity AI factory in South Korea using Nvidia Vera Rubin DSX platform and SK Hynix HBM4 memory, with the first factory planned to go online in 2027. The two sides will cooperate to develop next-generation AI memory technology to address surging AI demand in the Asia-Pacific region.
Source links can be verified on the original siteRead original ↗To counter social platforms training AI on user images, a developer launched a new tool that converts photos into noisy animations to interfere with AI recognition. Unexpectedly, this protective technique, due to its unique jittery visual style, was instead used by netizens as a creative tool to make various fun or retro-style image works.
Source links can be verified on the original siteRead original ↗Publication AI’s effects on overall employment is likely small, though a tough job market for new graduates may be partly due to AI. AI’s impact on worker productivity is mixed but generally positive. Firm adoption has accelerated but unevenly across the economy. Early evidence is hardly the last word on the future of work in an AI world.
Source links can be verified on the original siteRead original ↗The KwaiKAT Team at Kuaishou has introduced the KAT-Coder-V2.5. It is a coding model trained to operate inside real, executable repositories rather than emit single-turn code. The served model is available through StreamLake. An open-weight variant, KAT-Coder-V2.5-Dev, was released separately on Hugging Face under Apache-2.0.
Source links can be verified on the original siteRead original ↗Anthropic's Claude Opus 5 scored 30.2 percent on ARC-AGI-3, nearly quadrupling GPT-5.6 Sol's previous record of 7.8 percent. The benchmark's developers say the model independently formulated reflection equations, a behavior they had never seen from another model, and attribute to stronger logical reasoning.
Source links can be verified on the original siteRead original ↗In this tutorial, we exploreFAIRChem v2 and the UMA universal machine-learning interatomic potential as a unified framework for atomistic simulation across molecular chemistry, catalysis, and inorganic materials. We configure an environment, authenticate with Hugging Face to access the gated UMA model weights, and initialize task-specific calculators for the omol, oc20, and omat domains. We then apply the same pretrained potential to a broad set of computational chemistry workflows, including single-point energy and force prediction, molecular
Source links can be verified on the original siteRead original ↗In summer 2025, OpenAI internally flagged GPT-5 as high-risk because it helped users create biological hazards, but downgraded the model's risk rating that fall. According to the Wall Street Journal, some users got step-by-step instructions for making poisons and biological weapons. Hundreds asked for that kind of information.
Source links can be verified on the original siteRead original ↗Article 50 of the EU AI Act, the AI transparency guideline, is formally implemented on August 2. Chatbots and voice assistants must clearly inform users that they are AI systems, and deepfake content must be labeled. The new rule has a 4-month transition period, and violating companies face fines of up to 15 million euros.
Source links can be verified on the original siteRead original ↗For our second Content Independence Day, we’re giving website owners finer options to manage AI traffic. Instead of a one-size-fits-all block, all customers can now easily distinguish and manage Search, Agent, and Training bots, alongside the new ability to protect ad-monetized pages.
Source links can be verified on the original siteRead original ↗MiHoYo's AI chat software AnuNeko announced it will permanently shut down on July 30. This AI chat tool, known for being 'aggressive', once provided emotional companionship to users, and now will redirect resources to other directions. After the server shutdown, all user data will be deleted.
Source links can be verified on the original siteRead original ↗The two sides met in San Francisco, focusing on discussion of cooperation on semiconductor products such as HBM and DRAM and Samsung's AI transformation strategy. Since the strategic cooperation last October, the two sides have committed to expanding advanced memory chip production and building next-generation AI infrastructure.
Source links can be verified on the original siteRead original ↗An ACM survey of 763 computer science educators from 49 countries shows that 68 percent have already changed their exams because of AI, shifting toward oral exams, proctored tests, and project-based work. Teaching is moving from writing code to understanding it. But nearly half of respondents say they lack proven examples for integrating AI into their courses.
Source links can be verified on the original siteRead original ↗Seoul National University announced a partnership with NVIDIA to establish a comprehensive AI laboratory (NVAITC) on campus. The lab will integrate talent exchange, education and training, and AI infrastructure, focusing on robotics and physical AI, semiconductors, and AI scientific computing, while promoting startup incubation and industrial applications, aiming to become an "Asia-leading" AI research center.
Source links can be verified on the original siteRead original ↗Don't panic! You will be automatically returned to the ITHome homepage in 5 seconds, please wait...
Source links can be verified on the original siteRead original ↗After the Kimi K3 large model drew outside attention, Moonshot AI is rumored to have held a celebration at a bar in Beijing. In photos circulating from the scene, celebration slogans read 'K3 scale up and upgrade!' 'K4 go all out to the extreme!' 'Charge to the moon!', showing the team's ambition for the future. Moonshot AI co-founder and president Zhang Yutong was reportedly present.
Source links can be verified on the original siteRead original ↗On the day Claude Opus 5 launched, its complete system prompt of up to 34,000 tokens was made public. This document is not only a tool manual, but also a detailed compliance handbook and business guide, strictly specifying the AI's memory, citation, privacy, and commercial recommendation behaviors, revealing how Anthropic sets behavioral boundaries for top-tier AI models.
Source links can be verified on the original siteRead original ↗Time Line Proposal A Proposer Proposal A Seconds Proposal A Proposal B Proposer Proposal B Seconds Proposal B Proposal C Proposer Proposal C Seconds Proposal C Proposal D Proposer Proposal D Seconds Proposal D Proposal E Proposer Proposal E Seconds Proposal E Proposal F Proposer Proposal F Seconds Proposal F Proposal G Proposer Proposal G Seconds Proposal G Proposal H Proposer Proposal H Seconds Proposal H Quorum Data and Statistics Majority Requirement Outcome
Source links can be verified on the original siteRead original ↗[](https://github.com/slvDev/esp32-ai#running-a-289m-parameter-llm-on-a-microcontroller) 𝕏 slvDev·LinkedIn 
Source links can be verified on the original siteRead original ↗Nvidia CEO Jensen Huang, in an interview with Bloomberg, emphasized that talent is the core driving force of AI development. He pointed out that the number of AI talents cultivated by China each year is astonishing, and it is destined to achieve excellence in this field. Although China and the US have different conditions, outstanding researchers can always find answers, and should continue to learn from and promote development with each other.
Source links can be verified on the original siteRead original ↗