Lisa Shalett of Morgan Stanley expressed concern over the sustainability of the current stock market, which she argues relies on a single narrative supported by massive investments in generative AI.
Google announced that it has applied 72 security fixes to open-source projects using its AI agent, CodeMender. Leveraging Gemini Deep Think models, the system enables automated vulnerability detection and patching.
Researchers analyzed the embedding matrix of GPT-OSS, indicating that GPT-5's training data included phrases from adult websites. They also suggested that coding-related reinforcement learning may have been the final stage of the training process.
OpenAI announced that it will soon implement enhanced management authority for rights holders and monetization measures within its video generation AI, Sora.
Based on feedback from Sora users and rights holders, OpenAI has revealed plans for fine-grained control over character generation and the early launch of revenue sharing.
OpenAI revealed in disclosure documents to shareholders that its revenue for the first half of 2025 was $4.3 billion, with a net loss of $13.5 billion.
edge264, a cross-platform open-source decoder for H.264/AVC video, has been released. Designed with a focus on speed and ease of use, it utilizes C vector extensions instead of hand-written assembly code.
A minimal implementation featuring a manifest.json, Service Worker, and a single HTML file to trigger the PWA installation flow in Chrome has been introduced.
Google has reportedly adopted a specification that prevents AI Overviews from appearing for queries regarding President Trump's dementia or Alzheimer's disease, while similar searches for former Presidents Biden and Obama still trigger AI summaries.
In ranking data released by OpenRouter, Grok has become the most utilized model based on the number of processed tokens.
Google has announced Gemini Robotics 1.5 and Gemini Robotics-ER 1.5, enabling robots to perceive environments and plan actions. The former handles vision-language-action, while the latter manages embodied reasoning to support complex multi-step tasks.
The macOS app Dayflow uses a local VLM to read screen content and automatically generate semantic timelines, such as recording PR reviews for 45 minutes instead of simply Chrome for 3 hours.
Security firm Veria Labs has reported a vulnerability where flaws in the authentication flow allow remote code execution when connecting to Claude Code or Gemini CLI via MCP servers.
OpenAI and NVIDIA have signed a letter of intent for a strategic partnership to deploy at least 10GW of NVIDIA systems as next-generation AI infrastructure.
On September 20, 2025, xAI announced Grok 4 Fast, a new model specialized for cost efficiency. It reduces average inference tokens by 40% and achieves benchmark performance equivalent to Grok 4 at a 98% lower price.
xAI has announced Grok 4 Fast, a cost-efficient reasoning model. The company states that it achieves performance equivalent to Grok 4 using 40% fewer thought tokens, reducing the cost for the same performance level by 98%.
Google has announced the early release of a preview version of Gemini in Chrome for Google AI Pro and Ultra subscribers in the US, enabling summaries and Q&A based on the context of open tabs.
The Cactus team has released Cactus, an AI inference engine designed for smartphones. Using a Qwen3-600m INT8 model, it achieves speeds of 50 to 70 tokens per second on the latest devices.
Demis Hassabis of Google DeepMind stated that the ability to "learn how to learn" will be the most essential skill for the next generation to adapt to rapid AI-driven changes.