Skip to content
ToTheMoon host

Tatyana Tsvetkova — interior designer, CEO of PT Interiors and ToTheMoon host

Tatyana Tsvetkova is the CEO of PT Interiors and an interior designer whose work has been published in AD, Elle Decor and Vogue. On ToTheMoon she looks at artificial intelligence not from the lab's side but from inside a profession where taste, the client's trust and the story behind an object decide everything. Tatyana tries models at work and on her travels, but reminds listeners that they make mistakes too, that their answer is just one opinion among others and that their output can be monotonous. Here — who she is, what she talks about in the episodes and where to start listening.

Updated November 16, 2026 · ToTheMoon editors

Who is Tatyana Tsvetkova

The project's page introduces Tatyana Tsvetkova briefly: CEO of PT Interiors, interior designer, published in AD, Elle Decor and Vogue. That experience comes through clearly in what she says: she reasons about AI as someone to whom, in her own words, a client comes for confidence in a decision, and not at all for yet another hundred options.

On the podcast Tatyana is one of the regular hosts: more than a hundred episodes have come out with her. You can hear her together with Alexander Volchek and Ilnar Shafigullin — both in themed conversations and in the Sunday reviews of the week. While her colleagues break down models, limits and deals, she adds to the conversation what she sees with her clients, on a trip to China and in her own attempts to work with Claude Design and Codex.

What Tatyana Tsvetkova talks about on the podcast

The latest episodes with Tatyana, from May to September 2026, add up to a chronicle of the AI race: the release of Fable 5 and Fable 5.1, Claude's memory opened up to the user, the pause in Astra training at OpenAI, arguments over how the OpenAI and Anthropic IPOs will affect access to their models, and AI stepping out of the chat window — from Claude Design to an autonomous Codex. Alongside them run conversations about people: how models blur the boundaries between professions, why entry-level jobs are melting away and what will happen when agents start buying and haggling for us.

China comes up in her remarks more than once. In one of the Sunday reviews Tatyana said that everyone there who can afford it uses American models. And when Ilnar, getting by in China with Google Translate, asked why not use ChatGPT as a near-simultaneous interpreter, it turned out that Tatyana had done exactly that there three months earlier, and it had helped a lot. In the same conversation she remarked that models' chats, reports and research sometimes come out very strange.

Tatyana Tsvetkova's key ideas

Her position comes across most fully in the episode on where AI helps and where it does harm. In interior design, Tatyana believes, it currently does more harm than good: a client already has a thousand pieces of advice — from a husband, a mother, friends — and when a model that has ground through a mountain of data piles on top, the person is horrified. In medicine or mathematics, where everything is more clear-cut, it works better, she says. To Ilnar's idea of checking work with a model before sending it off, she objected: the model also makes mistakes and hallucinates, and its answer is one more opinion, not the truth.

Tatyana tried Claude Design when she was urgently preparing two presentations for clients, and admitted that viewers had been right: the slides come out looking alike, and explaining the task to the model takes longer than doing it all herself in her usual program. She notices the same effect with emails — the recipient senses that a person did not write it, even if the text has no em dash. She never did start using Codex: it suggested basic steps she had arrived at on her own — admittedly in five to ten minutes instead of her several months — but things went no further, and she wrote the system off, although it has changed since then.

In the argument about the Higgsfield film made by AI, Tatyana suggested treating manipulation as a given: the model already knows what she will like. But she will buy a two-thousand-dollar pendant only if there is a story behind the object — a thing made by a machine has no soul. In the same episode she sided with classical education: five years at university build connections in your head, and afterwards you take in even a course on launching an Instagram account in a structured way.

Where to start listening to Tatyana Tsvetkova

To hear Tatyana at full voice, start with the episode on the benefits and harms of AI in different professions: it lays out her position as a designer in the most detail, along with the argument that a model raises the lower bar but not the upper one. Then come the Sunday episode on the US and China, with the conversation about the Higgsfield film, the soul of things and education, and the episode on the pause in Astra training, with the story of how she got to know Codex. Voice assistants and interpreting on a trip are in the episode about agents that shop and haggle instead of a person.

Questions

Tatyana Tsvetkova: in brief

Who is Tatyana Tsvetkova?

Tatyana Tsvetkova is the CEO of PT Interiors and an interior designer whose work has been published in AD, Elle Decor and Vogue. She is one of the hosts of ToTheMoon, a Russian-language podcast about artificial intelligence from Silicon Valley, and brings to the conversations the view of a professional who cares about taste, the client's trust and the story behind an object.

Which ToTheMoon episodes feature Tatyana Tsvetkova?

She has more than a hundred episodes to her name. Recent ones include conversations about agents that haggle instead of a person, the release of Fable 5.1 and Claude's open memory, the benefits and harms of AI in different professions, the pause in Astra training at OpenAI and the rivalry between the US and China.

What does Tatyana Tsvetkova think about AI in interior design?

In her words, at this stage AI does her work more harm than good: a client expects confidence, and a model only adds more options. Presentations from Claude Design struck her as monotonous, and explaining the task to the model, as she puts it, takes longer than doing it herself in a familiar tool.

Does Tatyana Tsvetkova use AI herself?

Yes, and she is candid about it. In China she used ChatGPT as a near-simultaneous interpreter, and it helped a lot. She tested Claude Design on two urgent client presentations and found the result monotonous, and she never did start using Codex: its first suggestions matched what she had already worked out herself.

Recent episodes

Latest episodes with this host

Open the page →
Other hosts

More hosts

All hosts →