ToolsDesign Mode: user interface can be modified by pointing, drawing, or speechThe feature called Design Mode allows users to directly update the user interface by pointing, drawing, or using voice; this simplifies UI editing and speeds up prototype creation.1 min read
ToolsCursor reduces gap between what users see and what the agent understands with visual promptsCursor introduces visual prompts designed to bring the information seen by users closer to the context interpreted by the artificial agent.1 min read
ToolsMeta launches AI assistant to help creators with performance insights and content ideasMeta has rolled out an AI-powered creative assistant that gives personalized recommendations to social media creators based on their content style, performance metrics, community and goals.2 min read
ToolsCursor displays agents' context usage on an interactive canvasCursor's new feature now presents agents' context usage as an interactive report on a canvas surface; the Context Explorer breaks down the token distribution among system prompt, tool definitions, rules and skills.1 min read
ToolsWith the Cursor canvases feature, dashboards, reports and internal tools can be created and shared via URLCursor has introduced the canvases feature, which allows dashboards, reports and internal tools to be created.1 min read
ToolsOpenAI's new ChatGPT memory system switches to automatic detailing and double storage for Plus and Pro usersOpenAI is rolling out a new ChatGPT memory system that automatically tracks important details; Plus and Pro users in the United States will receive it today with twice the memory.1 min read
ToolsOpenAI Introduces New Memory System for ChatGPTOpenAI has introduced a new memory system for ChatGPT that, via a memory editor summary, allows users to review and control what the system remembers.1 min read
ToolsOpenAI introduces enhanced ChatGPT memory systemOpenAI today introduced ChatGPT's enhanced memory system, which selectively carries context across conversations and preserves its usefulness over time.1 min read
ToolsOpenAI integrates Codex agent into ChatGPT and makes agents the defaultAt its "Intelligence at Work" event, OpenAI announced it will fold Codex, its coding agent, into ChatGPT in the coming weeks and introduce three upgrades: role-based plugins, inline annotations, and Sites for one‑prompt web apps.2 min read
ToolsGrok Imagine Video available on Vercel AI Gateway, image‑video model for developersThe Grok Imagine Video image‑video AI model has been made available on the Vercel AI Gateway, allowing developers and operators to more easily integrate the service that generates videos from images.1 min read
ToolsThree architecture patterns for generative UIs: Controlled, Declarative, Open-endedGenerative user interfaces are changing frontends: agents now produce UI in real time instead of merely describing it.11 min read
ToolsGoogle makes Gemini's 'Thinking' modes freely available on web and mobileGoogle has begun a broad rollout of two 'Thinking' modes for its Gemini chatbot—Normal and Expanded—after first announcing them at its I/O developer conference.2 min read
