Why AI is a boon for local media, not a threat

Paul Gewuerz runs LocalPod and hosts Small Press, Big Ideas, so he talks with legacy local publishers every week. He came back on the show to catch up, and the conversation kept returning to one question: should a local newsroom be afraid of AI? His answer is no. Then he shared his screen and showed how he turned our own articles into a podcast episode.

Why Paul says AI is Return of the Jedi

Paul says the legacy publishers he interviews are nervous for a reason. Over the last 20 years, Google and Meta wrecked their business, and Craigslist took classified revenue early on. In many towns the Facebook message board is now the town square. So when AI shows up, many of them assume it will finish the job.

Paul disagrees. He tells local publishers that this isn't The Empire Strikes Back. It's Return of the Jedi, and the AI tools are lightsabers. He points to the wave of tech layoffs, including 8,000 at Meta that he mentioned, as proof that you can do more with less. For a paper that is barely scraping by, he says, that is how you get your moat back.

The moat is local information

Paul lives in a town of about 20,000, and he says a newspaper there is the only reason people learned about a corrupt city commissioner who was voted out. No AI company is going to find that story. His point is that the reporting and relationships a local publisher has built are worth a lot, and the AI companies need that kind of material.

He mentioned Paul Myers of Mineral King Publishing in California, who runs a startup called Lantrn.ai. Paul's summary of the idea is that if a publisher can put up a hard gate and keep that content from AI companies, it holds something very valuable.

TJ added the cost side. Local papers shrank mostly by cutting people, which meant less got done. Now the same work can still get done without as many hands. If half of a reporter's week goes to things that don't move the needle, freeing that time means better reporting without much extra cost. AI isn't writing the stories. It handles the annoying parts, like bookkeeping or digging through city council meeting databases.

Journalists, ego and the business side

Paul interviews a lot of career journalists. He describes them as mission-driven and not very business-minded, which he says is great because they expose things that need exposing. But he sees some ego in how they respond to AI. His view is that you can still do high-quality investigative work and use these tools.

He gave one example on the other end. A guy in Palm Springs, California runs three publications that he started during the pandemic, and AI writes the content with one quick look-over. Paul thinks he's making multiple six figures. Paul added that out of about 130 episodes, that is the only person he has talked to who sends things out without checking them.

TJ made a related point about the money side. Journalists who assumed good writing would carry them forever now have to learn business, or partner with someone who knows it.

Community Impact and the $300 subscriber

To show how strong a local brand can be, Paul told the Community Impact story. John Garrett and his wife quit their jobs and put $40,000 on credit cards to start it. Paul says it now reaches about 2.5 million households through roughly 40 publications in Texas. His favorite detail: after a fender bender, a stranger spotted Garrett's Community Impact polo, said he read it every day, and the argument calmed down.

TJ followed with a number for people who run Meta ads to newsletters. Some legacy brands charge $10 a month, and he says their lifetime value can be around $300. Fulfilling one more subscriber costs them almost nothing, so paying $200 to get one still leaves $100 in profit.

Why local podcasts pay

When Paul talks to older publishers, he says, they ask whether anyone really listens to podcasts. He tells them to count the earbuds at the gym or the grocery store. His case is that audio is close to distraction-free, because you're not competing with a screen when someone is doing dishes or taking a walk. That is why, he says, podcast ads command higher CPMs than most media.

Paul puts the industry standard at 30 to 35. For local ads, like a steakhouse a few miles from the listener, he has seen publishers get 50 to 75. He has been bringing costs down for a couple of years with AI narration. TJ's point was that in most towns nobody could justify recording a local show by hand, so the choice is this or nothing, and listeners will take it.

What LocalPod Studio does, live

Paul built Studio with Claude Code and some software knowledge from the last 20 years. Then he showed it on screen. He copied a few paragraphs from one of our articles, pasted it in, and chose an AI voice. His favorite is Brian, and he says more voices are coming. He added show notes, or you can upload a PDF.

The first audio take sounded off, so he regenerated it. TJ's verdict: you can tell it's AI by the cadence, but it sounds real enough that he'd still listen to find out what was going on in his community.

Then Paul set up an ad. He made a campaign for a made-up steakhouse, set it as a pre-roll or mid-roll, typed the ad copy and generated it. TJ asked if the ads are the real gold, and Paul agreed. If you're good at selling ads, it's a profit center. Then he approved and published, and the episode goes live on Apple Podcasts, Spotify and the other apps.

Paul says the ElevenLabs-style pitch of "turn your stuff into a podcast" only gives you the audio, and you still have to upload it and set up the rest yourself. Studio does it all. You can also upload an MP3 of a recorded interview. He is working on a big red record button so reporters can record in the field and save it as a draft or publish it right away, which he expects in the next month or two.

Why a specialized tool beats a general AI

TJ compared it to newsletters. You can tell Claude to write you a newsletter, but it won't be publishable, and you'll be asking where the ads go. His view is that the whole economy will be people in every niche building specialized tools on top of the big models.

Paul said anyone could stitch this together themselves, but it took most of his time for a month and a half. TJ's version is that most business owners would rather pay Paul $99 a month. If you have more money than time, you pay. If you have more time than money, you can build it. Both paths exist now.

TJ also pointed out where this is heading. Instead of logging in, you tell Claude to pull your last four articles, run them through LocalPod in a chosen voice, and confirm when it's published. Paul said they aren't there yet, but that is where they're headed.

The Cleveland newsroom and a podcast as lead gen

TJ brought up a Cleveland newspaper whose reporters gather the facts in the field and then hand their notes to AI to write the story. Paul thinks it was the Cleveland Plain Dealer. TJ calls it a sandwich: a human does the first 10% and the last 10%. Paul says the reaction was milder than it would have been a year earlier.

On the lead gen side, Paul says a podcast gets you into rooms. If you want a real estate brokerage as a sponsor, interview their CEO or top agents, because everyone in that office will listen. He says he uses his own podcast this way all the time.