🗽 Live from NYSE

The conversation starts with the host and a guest on election night framing the midterms as a market event, with Texas data-center politics, prediction markets, and a few close Senate races carrying the most attention. The guest argues that prediction markets are worth taking seriously because they distill conventional wisdom into a usable probability, even if the outcome still looks unlikely. The pair also speculate about weird-chaos scenarios, like a narrow set of Democratic wins that would send social media into overdrive, and they treat weather, outages, and turnout as potential election variables rather than certainties.

From there, the discussion turns to AI, and the guest says the central question is not whether AI helps, but who actually captures the economic value. They argue that the AI trade still needs a major capabilities jump before most companies can justify large spending, and that today’s agents are still too constrained and too similar to one another. The host and guest also debate productivity, consumer strength in New York, and whether the economy is being carried by households, businesses, and government spending all at once. The guest is skeptical of easy claims that AI has already transformed productivity, and says the real breakthrough will come when software has true agency and budget authority.

Later segments broaden into media, newsletters, and real estate. Emily Sunberg describes Feed Me as a business that grew from an early Substack into events, podcasts, and a production deal, while noting that the broader Substack ecosystem is getting crowded. Ryan Sirant then makes a forceful case that real estate is being reshaped by social media, local studio teams, and AI-native marketing, while arguing that the market is tight, affordability is broken, and fake AI listing photos should be banned rather than normalized.