Co-constructing intent with AI agents | by TenoLiu | Aug, 2025

Published on:

Press enter or click to view image in full size

Claude webpage with headline “Meet Claude, your thinking partner.”
Anthropic’s official introduction: Meet Claude, your thinking partner — screenshot via

Reading between the lines with multimodality

Press enter or click to view image in full size

A screenshot of Hume AI, showing a podcast on the left and emotion and behavior analysis on the right, displaying three people with their emotions labeled as joy, awe, and curiosity based on the analysis.
Hume AI can analyze the emotion in a speaker’s voice and respond with empathetic intelligence.

Information as a flowing process

Press enter or click to view image in full size

A screenshot of the Apple Lisa 2’s black-and-white graphical user interface, showing the cursor selecting the “Copy” command from the “Edit” menu.
Apple Lisa 2 (1984): Features like desktop icons, the menu bar, and graphical windows significantly lowered the barrier to entry for personal computers
Press enter or click to view image in full size

A mind map in Chinese that visually lays out an AI entire thinking process for analyzing and solving a complex problem.
Metaso: Visualizes its entire thinking process on a canvas as it works on a problem.
Press enter or click to view image in full size

A screenshot of the Google NotebookLM interface demonstrating how source materials on art history are automatically transformed into a narrated video overview titled “The World of Surrealism”.
Google NotebookLM can transform source materials into various easy-to-digest formats, such as narrated video overviews, conversational podcasts, and interactive mind maps. This shifts learning from a process of passive consumption to a dynamic, co-creative experience.

Progressive construction through dialogue and memory

Press enter or click to view image in full size

A composite image explaining the “Dia Memory” feature, showing a person recording their environment with a phone on the left, and a corresponding personalized AI chat suggestion on the right.
Google’s Project Astra remembers what it sees and hears in real time, allowing it to answer contextual questions like, “Where did I leave my glasses?” The Dia browser’s memory feature continuously learns from your browsing history to develop a genuine understanding of your tastes
Press enter or click to view image in full size

A screenshot of a conversation with an AI tutor, where the AI asks clarifying questions about the user’s math background and goals before explaining Bayes’ theorem to personalize the lesson.
ChatGPT Study Mode. When given a task, its first instinct isn’t to jump straight to an answer. Instead, it begins by asking the user clarifying questions to better define the problem

Source link

Related