BLAINESWORLD · Blaine Greenfield

2.2.2026 - Mark Potts--CEO, LocAlmanac

February 2, 2026·44 min·3 clips

A senior software developer told Mark Potts his job is over because AI writes code better than he can.

1. Blaine's World Podcast host Blaine Greenfield interviews Mark Potts, retired journalist and CEO of Local Almanac, about AI, the collapse of print media, and the future of local information. 2. Potts spent decades at The Washington Post, Chicago Tribune, and other publications, and created one of the first digital newspaper prototypes for the Post in the 1990s; he is currently co-founder of Local Almanac, a six-month-old AI startup based in Lawrence, Kansas. 3. The episode's central question is whether AI is as transformational as advertised — and Potts' answer is that it is both overstated and underestimated at the same time. 4. Potts frames AI's current limitations using a direct analogy: 'Think of AI as an intern you have to manage' — capable of producing outputs but requiring constant supervision because it makes errors, including confidently misidentifying a musical performer before correcting itself when challenged. 5. He cites a recent conversation with a very senior software developer who described his job as 'over' because AI now writes code better than he can; the developer's new role is pure supervision of AI output. 6. Potts distinguishes the overhyped dimension — premature consumer launch, errors, unrealistic expectations — from the underhyped dimension: agent-based AI that can connect things within a computer and execute multi-step tasks like researching, booking, and paying for a trip has arrived in the last few weeks. 7. He compares the current moment to looking at the original iPhone before the App Store existed: a crude version of something that will be unrecognizable in 10 years. 8. Potts recounts being in the room when The Washington Post was offered 10% of AOL for $10 million and turned it down — an opportunity he calls a multibillion-dollar mistake. 9. He also describes a meeting where newspaper companies discussed acquiring the Excite search engine and declined on cost grounds; Excite sold for 10 times the price they were offered a year later. 10. Most strikingly, Potts details how a Dow Jones developer built a highly regarded search engine called RankedX at The Wall Street Journal, was not recognized for what he had created, left, and founded Baidu — now one of the largest companies in the world. 11. Local Almanac uses AI to create free community information dashboards that aggregate city documents, meeting minutes, permit records, and local media, starting with a launch on the Wichita, Kansas city website in the coming weeks. 12. For a single real estate development moving through a planning commission, Local Almanac can simultaneously surface the next meeting date, relevant zoning code, the developer's history, a contact for public comment, and a synthesis of prior coverage — information previously requiring hours of manual research across multiple sources. 13. One of Local Almanac's sources is Documenters, an organization that sends paid note-takers to local government meetings; the startup synthesizes those notes with other data streams. 14. Potts says Local Almanac is funded by a couple of small grants rather than venture capital, costs far less to build than traditional media companies because it has no reporters, and plans to make the core product free while monetizing premium versions sold to consultants and city governments. 15. He argues the startup is not replacing journalism but filling a void that already exists — most local newspapers no longer have the staff or budget to cover city meetings, and Local Almanac is surfacing information that exists in public records but has not been accessible to ordinary residents. 16. On Google's NotebookLLM tool, Potts describes uploading a three-page company plan and receiving a 20-minute AI-generated podcast of two computer-generated people discussing the company — produced in six minutes, complete with jokes, asides, and issues the founders had not explicitly discussed. 17. The episode has a conversational tone with the host drawing on personal experiences — canceling the Wall Street Journal, walking his dog while listening to the Mel Brooks audiobook — to ground the discussion in everyday AI use. 18. Potts shows comfortable fluency switching between consumer AI products (ChatGPT, Gemini, Claude, NotebookLLM) and structural industry analysis without treating them as the same thing. 19. People interested in the intersection of AI, local government information, and the decline of local journalism will find this episode directly relevant. 20. Listeners expecting technical depth on AI architecture, policy, or specific product comparisons will find the conversation anecdotal rather than analytical.

As heard by us

A plainspoken look at AI's useful side, its mistakes, and the places it deserves caution.

Mark Potts spends this conversation on AI's uncomfortable edge: a fake background that can be customized, tools that sound certain and still miss basic facts, and the easy habit of trusting them anyway.

Read the full review in PlayNext →

Why you'd press play

You want a plainspoken check-in on AI hype, mistakes, and the human cost.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from BLAINESWORLD