Use AI to mass produce podcasts, reduce cost per episode to $1, turn successful podcasts into influencers...profit.

The company is able to produce each episode for $1 or less, depending on length and complexity, and attach programmatic advertising to it. This generally means that if about 20 people listen to that episode, the company made a profit on that episode, without factoring in overhead.

Inception Point AI already has more than 5,000 shows across its Quiet Please Podcast Network and produces more than 3,000 episodes a week. Collectively, the network has seen 10 million downloads since September 2023. It takes about an hour to create an episode, from coming up with the idea to getting it out in the world.

The company has been recently experimenting with short-form videos and creating social media profiles for the AI personalities, in the hopes of eventually turning some into influencers. Wright hopes to create thousands more personalities in the near future to see what personalities stick.

I like to think that such podcasts won't ever be terribly successful, that they will quickly feel dull and not have the staying power of human podcasts...

But who am I kidding? A lot of podcasts are not very interesting or just rehash what everyone else is already saying. If one were to use a company like this too a-b test the shit out of topics, I have no doubt they could narrow in on the perfect mix of opinions, looks, and attitude that would be wildly popular.

This company doesn't make podcasts, it makes influencers.

Final note:

The shows are also spot-checked periodically.

They don't even listen to the shows they create...

view all related items

239 sats \ 3 replies \ @freetx 13 Sep

I like to think that such podcasts won't ever be terribly successful, that they will quickly feel dull and not have the staying power of human podcasts...

I think there is a niche market where this can succeed. "how to" podcast.

This has been going on for years and years on youtube. Look up any type of technical instructional video and you will find dozens of how-to vids that are entirely AI generated.

The reason why this works is people are: (a) interested in a narrow subject and that video/podcast hits exactly on those terms, and (b) is not interested in compelling discussion...its literally just "tell me how to do this thing".

121 sats \ 2 replies \ @Scoresby OP 13 Sep

Later in the article it says one of the people behind this got his start by doing a podcast that just read CDC reports during the pandemic, then expanded to reading weather reports.

You make a great point. Name your niche skill -- here is a weekly podcast that explains how to do it and delivers the latest on updates to the skill. This is something even I might listen to.

111 sats \ 1 reply \ @freetx 13 Sep

This is something even I might listen to.

Agree. Imagine a customized newscast on the daily/weekly updates for your specific hobby or interest. Listen while driving to work or whatever.....

28 sats \ 0 replies \ @optimism 13 Sep

Discussions work well too!

https://m.stacker.news/107979

100 sats \ 3 replies \ @kepford 23h

I played with Notebook LM's podcasting feature a while back. Tricked some friends into thinking it was a real podcast talking about their business.

Personally there's no way I could listen to it but several people I played it for thought it was great. But, I don't think they actually listen to any podcasts.

I have a feeling that the knowledge that it's AI will limit the success regardless of how real it sounds. That is not to say it can't have success.

186 sats \ 2 replies \ @optimism 22h

The main problem is depth, but as @freetx says above, there's a niche area of no-nonsense podcasts and explainer videos that are very doable without needing specific insights.

I'm not sure about the podcast format though - why not just do narrated videos? I made the script for the little joke "podcast" above within a minute and the voice rendering took 280 seconds on my 5 year old MacBook M1 with the small 1.5B vibevoice model (once I figured that it works better to batch it by voice, not try to render it all in 1 shot). Alibaba has released a novel pictures + script to video model that would enable anyone with a 24GB GeForce card to just roll out explainer vids.

202 sats \ 1 reply \ @kepford 18h

Gonna be interesting to see how this tech is used.

102 sats \ 0 replies \ @optimism 9h

I expect floods of scams. I wish it wasn't so but I have little hope. 15 years of Bitcoin's code being used for scamming / self-enrichment is probably a good measuring stick?

102 sats \ 0 replies \ @kepford 23h

They don't even listen to the shows they create...

LOL

0 sats \ 0 replies \ @cryotosensei 10h

This might repulse you, but I actually attended a workshop on how to use NotebookLM helmed by a secondary school educator. The premise was that since we upload our teaching materials and other curated resources as Databases, the podcast generated by NotebookLM will engage and intrigue our students. Cue Joy of learning, my nation’s current emphasis

0 sats \ 0 replies \ @OT 10h

I might listen to one that compresses the Bitcoin news of the week into about an hour. I currently have more than 30 hours of podcasts in my queue and will probably end up clearing it. Too much time, too much info. As long as it's not sloppy AI BS I think it could be a thing for me.