DEVOURED
Designing with AI without losing your mind
A designer concerned about outsourcing critical thinking to AI built a real-time collaborator that challenges ideas during sketching rather than just generating solutions.
What: Thia is an open-source AI app with vision and voice capabilities that engages in critical dialogue while a designer sketches on a whiteboard, acting more like a questioning colleague than a solution generator. Built using Google AI Studio in three days and released on GitHub.
Why it matters: The article argues that while AI tools accelerate execution, the pressure to ship faster is causing designers to skip crucial thinking stages, leading to weak solutions that get built quickly but lack substance. The author wanted to preserve the "cognitive friction" of early ideation rather than jumping straight to high-fidelity outputs.
Takeaway: Try Thia on GitHub or Google AI Studio to experiment with AI that supports early-stage ideation, or consider how to preserve critical thinking stages in your own AI-augmented workflow.
Deep dive
- The author noticed a behavioral shift where they were reaching for LLMs to solve design problems instead of spending time thinking through and sketching solutions themselves
- Current AI design tools enable organizations to ship fast and test rapidly, but this has devolved into throwing "cheap spaghetti" at the wall rather than filtering out weak ideas through early scrutiny
- The author identified whiteboarding and sketching as a "sacred" design stage where deep thinking shouldn't be outsourced to AI, but wanted support for it when working solo or remotely
- Existing tools like Google Stitch and Gemini either jumped straight to high-fidelity or weren't tuned to be critical, missing the need for challenging dialogue during early ideation
- Thia was built to engage in Socratic dialogue in real-time while watching the designer sketch via webcam, creating a tight feedback loop instead of a waterfall upload-chat-iterate process
- The app was developed using Google AI Studio and refined through iterative prompting over three days, with the author using Thia to design Thia's own features
- Physical sketching and writing activates different brain regions than typing, forging deeper long-term memory connections and supporting internal evaluation
- The real game-changer with AI is execution speed at each stage, not permission to skip stages like user research or robust idea refinement
- AI's ease can create a facade where polished prototypes with features like breathing animations paper over fractured concepts and weak value propositions
- Critical thinking skills—analysis, evaluation, cross-domain synthesis—will become more valuable once the AI hype settles, and maintaining them requires embracing cognitive friction rather than rushing to the finish line
- The roadmap for Thia includes overwriting on screenshots, multiple simultaneous participants, calling multiple models for complex tasks, and orchestrating agents on platforms like OpenClaw
Decoder
- LLM: Large Language Model, AI systems like ChatGPT or Claude trained on massive text datasets to generate human-like responses
- Socratic dialogue: A questioning method that challenges assumptions and encourages critical thinking through systematic questioning rather than direct answers
- Google AI Studio: A development environment from Google providing access to frontier AI models for building applications
- Multi-modal AI: AI systems that can process multiple types of input simultaneously, like text, voice, and images
- Cognitive friction: The mental effort required for deep thinking and evaluation, which the author argues should be preserved rather than eliminated
Original article
Heavy reliance on AI tools is pushing designers to outsource critical thinking, favoring speed and rapid output over deeper evaluation and idea development. To counter this, an AI collaborator (Thia) was created to engage in real-time, critical dialogue during sketching—supporting, rather than replacing, human thinking. The key takeaway is that while AI accelerates execution, strong design still depends on deliberate reasoning and “cognitive friction,” which must be preserved to avoid building fast but flawed solutions.
Explore all 66 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi