AI News — Friday, July 10, 2026
OpenAI has officially launched its new family of models with GPT-5.6, which is immediately being integrated as the preferred model for Microsoft 365 Copilot, signaling a deepening partnership despite recent 'breakup chatter'.
Fidji Simo, a prominent figure, has resigned from her position as OpenAI's second-in-command, marking a significant leadership change within the leading AI research company.
Railway has raised $100 million in funding to develop an AI-native cloud infrastructure platform, aiming to compete directly with established giants like AWS.
Google has unveiled the first redesign of its iconic search box in 25 years, a move that is expected to significantly impact user interaction and likely reflects deeper AI integration.
New research introduces a deep learning approach for accurate, interdisciplinary, and transparent understanding of structure-property relationships, leveraging native structural reasoning.
Researchers propose a novel architecture incorporating dual latent memory into Vision-Language-Action models to enhance robotic manipulation capabilities.
A new model named Vidu S1 has been introduced, capable of generating real-time interactive video, pushing the boundaries of dynamic content creation.
New research explores scaling Mixture-of-Experts (MoE) models for video pretraining to significantly advance embodied intelligence systems.
A new report highlights a significant cost disparity, noting that Claude Code charges up to $200 monthly for services that a competitor, Goose, provides for free.
OpenAI is marketing ChatGPT as an essential partner for users undertaking their most ambitious projects, emphasizing its enhanced capabilities and reliability.
An investigative report reveals a conflict between two AI suppliers where analysis of their logs indicated questionable practices from both parties.
Researchers introduce 'LLM-as-a-Tutor,' a method that uses Large Language Models for policy-aware prompt adaptation to improve non-verifiable Reinforcement Learning.
An opinion piece argues that manual coding, often seen as more 'honest,' is simply less efficient than leveraging AI tools, suggesting a shift in developer mindset.
A developer shares an experience of gaining significant insights by refactoring and deleting AI-generated code, highlighting the learning potential in code review over initial creation.
OpenAI has initiated a bug bounty program specifically for its GPT-5.5 Bio model, inviting researchers to identify and report vulnerabilities related to biological applications.