No Priors is hosted by Sarah Guo and Elad Gil, both venture investors, and the show is exactly what two thoughtful investors talking to founders sounds like. That framing is not a criticism, it is a description, and understanding it tells you whether the show is for you. The hosts are better prepared technically than most business podcasters. They can follow a technical answer, ask a specific follow up, and push on the part of a claim that matters.
Episodes with researchers rather than founders are often the strongest for this reason. There is a version of the AI business podcast where the host nods through a technical explanation they have not understood, and this is not that. The access is genuinely good. Founders of companies people want to hear from agree to come on, partly because of who the hosts are and partly because the venture ecosystem makes those introductions easy.
If you want to hear how the person running a company you use thinks about their roadmap, this is often where that conversation happens. The half hour to hour length is a deliberate discipline and I appreciate it. The AI podcast genre has drifted towards three hour episodes and No Priors gets in, covers ground, and gets out. If your listening time is limited that ratio matters.
What the show is genuinely good at is the commercial layer. Where does value accumulate in the stack, what does defensibility look like when models commoditise, how do you price a product whose marginal cost is inference, what does a services business look like when the model does the work. These are interesting questions, they are underdiscussed in technical venues, and the hosts think about them professionally. Anyone building an AI product will get something useful here.
Now the things to hold in mind. The investor lens is not neutral. Venture investors have a professional interest in the belief that this technology is transformative and that enormous value will be created quickly, and that belief shapes which questions get asked and which do not. Whether current capabilities are being oversold, whether a category is a feature rather than a company, whether the unit economics work at all, these come up and they come up less often and less pointedly than a disinterested host would raise them.
Conflicts of interest run through the guest list. Founders whose companies are in or near the hosts' portfolios appear regularly. Disclosure happens and it is not always prominent, and even with disclosure a conversation between an investor and a founder in their orbit has a different character from an interview. It will be more generous.
That is human and it is worth knowing. There is no educational content. You will not learn how anything works. Listeners looking to understand AI technically should go elsewhere, and this is a show about the industry, not the subject.
I want to be clear about that because the AI podcast category gets recommended as a bundle and these shows do very different jobs. The short format has a cost as well as a benefit. Interesting threads get opened and then abandoned because the clock is running. Where Dwarkesh Patel would spend forty minutes on one question until it is genuinely resolved, No Priors will spend six and move on.
Breadth over depth is a legitimate choice and it means fewer episodes leave you with something you did not already suspect. How I use it: when a company I am interested in has a founder on, or when the topic is market structure rather than technology. Not as a subscription I listen to completely. My three point four is for a well made, well hosted, disciplined show that is genuinely informative about the business of AI, marked down for a perspective that is structurally interested in the outcome and for a format too short to get all the way to the bottom of anything.