ň#
10,000 Fake Podcasts, $1 an Episode: One Company Is Flooding Your Feed With Robot Radio39% of new podcasts dropping right now are made by machines. Nobody’s telling you the feed is already half-fake.
One outfit called Inception Point AI is running 10,000+ shows. Dropped 800 of them in a single 48-hour stretch. Cost to make one episode: about a buck.
Look, this comes straight from Bloomberg (picked it up off a Slashdot thread, followed the trail). The word going around is “podslop” — cheap AI audio shoveled onto Spotify and Apple by the thousand. The Podcast Index says nearly 4 in 10 fresh shows over a recent stretch were probably robots. Nieman Lab covered it too if you want the media-nerd version.
🧩 Dumb Mode Dictionary (read this first, no shame)
| They say | What it actually means |
|---|---|
| Podslop | Cheap AI-made podcasts pumped out by the thousand, mostly junk |
| Inception Point AI | The company doing it at biggest scale — 10,000+ shows |
| Podcast Index | A free public list of basically every podcast that exists |
| Long-tail search | Super specific stuff people Google, like “back pain stretches for truckers” |
| AI disclosure | A little label saying “a robot helped make this” |
| Discovery | How you find a new show — search, charts, “you might like” |
| Voice cloning | Software copies your voice so you never have to record again |
📻 So what's actually going on here?
Real talk: making a podcast used to mean a mic, an hour of talking, and editing. Painful. Slow.
Now a company feeds a robot a topic — say “celebrity gossip” or “morning stretches” — and it writes the script, does the voice, and posts the episode. All of it. For about a dollar.
- One company (Inception Point AI) is running over 10,000 shows at once.
- They dropped 800+ new shows in 48 hours. Humans can’t type that fast, let alone talk.
- They aim at what people search for — health, wellness, celeb bios — so their junk shows up first.
They don’t need any single show to be good. They win on volume. Thousands of tiny shows each earning pennies = real money.
📊 The receipts (the numbers that matter)
| Thing | Number |
|---|---|
| New podcasts that are likely AI (recent stretch) | ~39% |
| Shows one company runs | 10,000+ |
| Shows dropped in 48 hours | 800+ |
| Cost to make one episode | ~$1 |
| Episodes a network can pump weekly | thousands |
Do the math. A dollar an episode, thousands of episodes, tiny ad payouts stacking up across 10,000 shows. That’s the whole play. Apple Podcasts now asks creators to disclose AI use — but the robots post faster than any human can check.
🗣️ What the timeline's saying
- Real podcasters: furious. Their honest shows are getting buried under a landfill of robot audio.
- Listeners: half don’t even notice. A calm AI voice reading Wikipedia sounds “fine” at 2x speed in the gym.
- Platforms: sweating. Spotify and Apple built their charts assuming humans make shows. That assumption just broke.
- The hustlers: already printing. Same thing that happened to blogs (2010) and YouTube (2015) is happening to audio right now — and the door’s still open.
⚠️ Why you should care even if you never make a podcast
Here’s the thing — this isn’t just a podcast problem. It’s a preview.
Every place that pays out based on “content exists = money” is about to get flooded by robots working while you sleep. Audio’s just first because it’s the cheapest to fake (no faces, no video).
The people who made money on this stuff early weren’t the smartest. They were the earliest. The window between “wait, this works?” and “everyone’s doing it” is where the whole bag is. That window is open right now for audio.
Cool. Robots Are Reading Wikipedia Into a Mic For Money… Now What the Hell Do We Do? (⊙_⊙)
🎣 The Long-Tail Lurker
The big robot farms all pile onto the same fat keywords — “true crime,” “celebrity news.” Brutal fights, no room.
So don’t fight there. Go where they’re NOT looking: hyper-specific stuff in languages and towns they ignore. “Fishing spots in Kerala.” “Small-business taxes in Nigeria.” Boring to them = wide open for you. Make a tight cluster of 20-30 laser-niche shows, let ad money trickle in from a corner nobody’s flooding yet.
Example: A 24-year-old media student in Manila spins up 25 ultra-local Tagalog “neighborhood news roundup” shows using ElevenLabs for voice and free Podcast Index data to spot gaps. Each show barely moves — but 25 of them stacked pulls a few hundred bucks a month in ad splits nobody else even saw.
Timeline: First cents in ~3 weeks. Good run for 4-6 months until the farms notice the niche and pile in. Then move to the next gap.
🕳️ The 'Made By a Human' Stamp
Look, when the feed fills with fakes, the rarest thing becomes proof you’re real. People will pay for that badge.
Build a simple directory that verifies and lists genuine human-made podcasts — a trust stamp. You become the first name people search when they want the real thing. First mover owns the search word forever (ask anyone who grabbed a category name early).
Example: A 27-year-old in Lisbon launches a free “Certified Human Podcast” directory on a $0 Carrd page, verifies shows over a quick video call, charges $12 for the badge + listing. 400 anxious indie podcasters sign up in two months = ~$4,800, plus the site becomes the SEO anchor for “real podcast.”
📡 The Backlog Flip
Grey-hat angle: there are mountains of free, legal, public-domain text — old books, expired-copyright essays, government reports. Boring on paper. But turn it into a clean audio series and you’ve got a niche show for $0 in content cost.
The robot farms chase trends. You chase the free vault nobody’s mined. Old classic literature, translated + narrated for a language that has zero audiobook options? That’s a real audience with no competition.
Example: A 22-year-old in Jakarta grabs public-domain classics from Project Gutenberg, runs them through Google NotebookLM + a cloned voice, and drops Indonesian-narrated versions no one else offers. Builds a loyal niche, monetizes with mid-roll ads = ~$600/month by month four.
🪟 The Compliance Sprint (patch-window play)
Apple and others just started demanding AI-disclosure labels. Thousands of these robot networks now technically need to tag their stuff — and most have no clue how, or they’re too lazy.
There’s a 2-4 week gap where being the person who cleans up their metadata and slaps on proper AI labels is a paid service nobody’s offering yet. Boring? Yes. That’s exactly why it pays.
Example: A 25-year-old freelancer in Cairo builds a dead-simple checklist + template for AI-disclosure tagging, DMs small AI-podcast networks on LinkedIn, charges $150 a batch to make them compliant. Ten clients in three weeks = $1,500, near-zero cost.
🎰 The Voice Ghostwriter
Blackhat tone, fully legal play: local businesses want a podcast for clout but will NEVER sit down and record one. So don’t make them.
Offer “your own branded weekly show, you never touch a mic.” You clone a friendly voice (with permission), script it around their business, publish it. They get a shiny “our company podcast” to brag about. You pocket a monthly fee to press one button.
Example: A 26-year-old in São Paulo pitches local dentists and real-estate agents a “done-for-you branded podcast” using ElevenLabs voices + Spotify for Podcasters hosting. Charges R$400/month per client. Eight clients = ~$650/month for maybe two hours of work a week.
🛠️ Follow-Up Actions
| Want to… | Do this |
|---|---|
| See who’s making what | Browse the free Podcast Index |
| Clone a voice cheap | Try ElevenLabs free tier |
| Turn text into audio talk | Test NotebookLM |
| Get free legal content | Raid Project Gutenberg |
| Host + publish free | Set up Spotify for Podcasters |
Quick Hits
| If you want… | Then… |
|---|---|
 Easy money corner |
Hit the niches the farms ignore — local + non-English |
 A long-term moat |
Be the “verified human” stamp before someone else is |
 Free content forever |
Public domain is a goldmine nobody’s mining |
 Fast cash now |
Sell AI-label cleanup while the rule’s brand new |
 Recurring income |
Run “done-for-you” branded shows for local shops |
The robots are already reading Wikipedia into a mic for a dollar. You can either scroll past the slop — or set up a booth right next to it and sell shovels.
!