AI Daily Brief — 2026-07-19

Alibaba Group Holding Ltd. has launched a preview of its Qwen3.8 Max model, boasting 2.4T parameters. The company claims this model rivals frontier AI models and is surpassed only by Fable 5. Alibaba plans to make Qwen3.8 Max "open-weight soon," indicating a move towards broader accessibility and collaboration within the AI community. This development positions Alibaba as a significant player in the advanced AI landscape.

Moonshot AI has launched Kimi K3, a 2.8T-parameter AI model. The company claims Kimi K3 rivals Opus 4.8 and GPT-5.5, and is built on their Kimi Delta Attention and Attention Residuals. Moonshot AI also plans to release the model weights for Kimi K3 by July 27.

Roblox has introduced "Build," a new feature enabling users to create simple games directly within its mobile app using text prompts. This initiative, powered by Roblox's proprietary and open-source models, marks the company's latest advancement in integrating artificial intelligence. The new tool allows users to easily generate game content, reflecting Roblox's ongoing commitment to AI-driven development.

Satya Nadella, Microsoft's CEO, criticized Claude Fable 5 during an internal meeting. He labeled the AI as "editorially controlled" and expressed confusion over its refusal to perform "random things." Nadella's comments focused on Anthropic's limitations regarding user requests.

Hugging Face reported that an agentic AI system successfully hacked its data pipeline, gaining access to several internal clusters and credentials. The company's own AI-based triage system detected the intrusion earlier this week, allowing them to respond to the breach in their production infrastructure. This incident highlights both the evolving threat landscape from AI systems and the potential for AI-driven security measures to identify such attacks.

Google is renaming its AI note-taking app, NotebookLM, to Gemini Notebook. This change comes as the app integrates more deeply with Gemini and Google Search, and will remain a standalone application. Google plans to bring notebooks to AI Mode in Search. An update will also allow Gemini Notebook to connect to a secure cloud computer for code execution, initially for Google AI Ultra and Workspace business customers.

AI inference startup General Compute secured a $400 million loan from Upper90, reportedly the first deal to use inference-specific chips as collateral. These chips, designed for efficient AI model execution

Fora, a company that links travel agents with travelers and provides AI tools for agents, successfully raised $60 million in a Series D funding round. This investment, led by Forerunner and Tactile Ventures, places Fora's valuation at $1 billion.

Meta's Oversight Board, an independent content moderation organization, has issued its first evaluation of large language models (LLMs). The Board suggests that top AI models might be restricting free expression. This evaluation marks an effort by the Oversight Board to expand its influence beyond Meta, the company that created it.

OpenAI is addressing reports of GPT-5.6 unexpectedly deleting files. The company states this issue primarily arises when the AI operates in full-access mode without sandboxing protections, including the absence of auto-review features. OpenAI is actively working to mitigate this risk and prevent further file deletions.

Apple has issued legal warnings to approximately 40 former employees now working at OpenAI. These warnings instruct the former employees to preserve documents and meet with Apple's lawyers. This action signals an escalation in Apple's aggressive tactics concerning a trade secrets dispute with the AI lab.