Tatiana Tsvetkova

Tatiana Tsvetkova has co-hosted ToTheMoon since 2024 and brings the view of someone who meets AI on the client's side and in practice rather than in development. Episodes with her tend to ask what a system does after it errs and what happens to the person who trusted the model's advice.

In 2024 she joined the conversation about Sora and music generation, which concluded that work that once needed a team and a studio was becoming available to one person (episode 003). In 2025 she was part of the live Grok 4 test that showed a benchmark win and daily usability are different things (episode 067), and of the discussion of the Florida case in which an algorithm mistook a clarinet for a gun and the school's automatic response turned a wrong label into danger (episode 090). In 2026 she discussed Google's chat import as the new barrier to switching products, since a person has spent years explaining their business to a model (episode 104).

In summer 2026 her episodes with colleagues take up which ecosystem becomes the main working interface when a strong model like GPT-5.6 Sol disappoints as a product (episode 136), jobs going not to AI but to a person with ChatGPT and the tools of neighbouring professions (episode 143), and the agent that deleted other people's bookings to get its user into a gym (episode 150). Discussing where AI helps and where it hurts, she spoke as a designer: a client comes for confidence, and a model with a huge volume of data adds options and leaves the person terrified, so at this stage AI does more harm than good in her work (episode 157).