Users can now access Claude in the sidebar of Google Workspace to read and adjust content, or open Google documents side by side with Claude conversations. This new feature is currently in beta testing for all paid plans.
According to the announcement from Anthropic
Để đủ điều kiện cho bậc Grow cao nhất, giờ đây người dùng chỉ cần tổng chi tiêu API là 500 USD, giảm từ mức 1.000 USD trước đó.
Theo công bố của OpenAI Devs.
OpenAI nới lỏng điều kiện đạt giới hạn API cao hơn, giảm ngưỡng chi tiêu cho bậc cao nhất. Cụ thể, năm bậc sử dụng trả phí nay được sắp xếp lại thành ba: Build, Launch và Grow.
#CAL#OpenAI
OpenAI launched new guidelines for using 'Custom Instructions' in API models, enabling developers to define behavior and maintain consistent formatting. This feature allows instructions to be applied across interactions, eliminating the need for repeated reminders.
#CAL#OpenAI
Manus AI Lab unveiled Manus 2.0, featuring the new H3 Max AI video model. This is a deeply trained video model that promises significant improvements in video processing and creation.
#CAL#AI
Cosmos Digest A04: the US 10-year yield hits 5.29%, the highest since May 2002.
The Fed raised rates on 16 Sep for the first time since 2023 (+0.25 pt, to 3.75–4.00%, vote 12–0). By 30 Sep the 2-year yield was 4.88% and the 10-year 5.29%.
#CAL#Bitcoin
Also in A04: foreign investors net sold VND 2,868 billion of Vietnamese stocks in the first week after the FTSE Russell upgrade, and four scenarios to Q3 2027 are updated (base case "higher for longer" now 40%, from 35%).
Information, not investment advice.
New AI assistant from OpenAI Devs promises to revolutionize work efficiency. The feature can learn user habits, maintain context across applications, coordinate Codex tasks, and automatically flag important information.
#CAL#OpenAI
The new model can generate images in 2K and 4K resolution, allowing up to 10 reference images to be combined and multiple targeted edits processed in a single run.
NVIDIA unveils new capabilities to build visual AI agents from text prompts using NVIDIA Cosmos and VSS 3.3. This marks a significant step forward in simplifying AI development, allowing users to create powerful AI applications with just text input.
#CAL#NVIDIA
This lightweight version supports videos lasting 1–15 seconds with natural sound, provides multiple resolution options from 480p to 1080p, and includes 7 different aspect ratios.