Tanya Donska Publishes Looks Good to Me: On AI Sycophancy, Context Loss, and Inverted Baselines Book

LONDON - Feb. 27, 2026 - PRLog -- Product designer Tanya Donska has published "Looks Good to Me: On AI Sycophancy, Context Loss, and Inverted Baselines," a collection of nine essays examining the structural problems that emerge when AI tools enter the design process. The book is available now on Google Play Books.

"Looks Good to Me: On AI Sycophancy, Context Loss, and Inverted Baselines" addresses a gap in current AI discourse: while most writing on AI in design focuses on capability and productivity gains, Donska documents what is quietly breaking underneath. The essays cover a range of structural failures — from AI sycophancy and context decay to the extinction of contextual design solutions and the inversion of quality baselines — drawn from direct observation of how AI tools are being used in product design and software development today.

The nine essays in "Looks Good to Me" are: Yes Men, Flattening, The Waiting Room, Prompt, Telephone, Recovery, Estimate, Collapse, and Inversion. Each essay identifies a specific structural problem, traces its causes, and examines its consequences for designers, developers, and product managers working with AI tools.

"AI agreed with everything I wrote. That felt like a problem worth investigating," said Tanya Donska. "These aren't predictions about what AI might do to design. They're observations about what it's already doing. The productivity pitch is loud. The structural problems are much quieter."

"Looks Good to Me: On AI Sycophancy, Context Loss, and Inverted Baselines" is published under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 license and is available free on Google Play Books. The book will be available on Apple Books, Kobo, and Barnes & Noble shortly.

About the Author

Tanya Donska is a product designer focused on the intersection of design process and artificial intelligence. She has been working with AI tools since the early days of large language models, observing their integration into design and development workflows. "Looks Good to Me: On AI Sycophancy, Context Loss, and Inverted Baselines" is her first book.