A former energy trader is building an AI tool to help traders keep up with more market information than any person can read. Anekha Sokhal, CEO of Moshi, explained the idea in an interview on the Latent Space podcast, published October 22, 2026.
What Moshi does
Sokhal traded energy derivatives, including at JPMorgan. These are contracts whose value depends on energy markets. Moshi is the tool she says she wished she had. It is a recommendation engine that connects market information to what matters for a trader's book, meaning that trader's own positions.
She describes it as Bloomberg, Claude and Netflix combined: market data, an interface for working with it, and picks tailored to the user. Claude is an AI assistant. Her broader point is that useful AI has to fit the way people actually work.
She also spoke about engineers she met at the AI Engineer New York conference. Many feel overwhelmed, she said, because managers tell them, "you are the team." That means one person is asked to build a system that would typically take 10 people.
What is not known yet
The material is an episode summary and part of the transcript. It does not give pricing, customer names, availability or any measure of how well Moshi works. The episode also lists trusted sources, document extraction and whether to build or buy AI models, but the available text does not cover them. Sokhal says conference conversations filled her following week with meetings, including talks in Houston with people in energy.
The takeawayWhen picking AI for work, favor tools that fit how your team already works over tools that demand a new routine.