Portraits by Google Labs lets you talk to AI versions of real experts. Starting with 'Radical Candor' author Kim Scott, it offers AI coaching based on her knowledge & voice. An early experiment.
It's interesting to see how AI is evolving. While the raw capabilities and voice quality of AI are advancing incredibly fast, real-world value often comes down to something more fundamental: trust. We might believe an AI is capable, but we tend to place our trust in the experience and authority of proven human experts.
A new experiment from Google Labs called Portraits builds on this exact concept. Instead of a generic AI, it offers conversational AI representations of trusted experts, built in partnership with them. It’s launching with a Portrait of Kim Scott, the author of "Radical Candor," to provide AI coaching based directly on her work and in her voice.
So, you're not just getting an answer from a capable model, you're getting guidance grounded in the specific expertise and content of a person you already trust. It's a fascinating approach to bridging AI capability with human authority, and a very important direction for making AI assistants genuinely more useful and trusted.
Tracking progress in this area...we also have audio enablement in our MarketFit product roadmap ...but haven't seen that maturity yet.
Fascinating to see how this evolves. On one hand, feels like a way for someone to scale their impact by giving advice to everyone who can pay a small free. But, on the other hand some of the value of 'experts' is their (lack of) accessibility. If I can talk with every 'expert' in a particular field, wonder what will make me reach out to a particular one.
This is an intriguing approach! What do you think about the idea of blending AI capabilities with trusted human expertise to create more personalized and authoritative AI assistants?
Portraits offers a thoughtful way to access expert insights through AI — bringing voices like Kim Scott’s directly into daily guidance. A promising step toward more personalized, high-impact learning.
Love the concept of building trust through expert-backed AI — really smart direction. But asking for login before showing anything makes it hard to evaluate
Read about it today and I am pretty curious how things will gonna play out because this attempt was initiated by Meta first, but then they shut down the project with AI celebrities.
