The bait, then the rug-pull.
Two operators sit down on a Sunday and draft picks from a list of thirty-plus AI startup opportunities. Greg Isenberg picks six. Jonathan Courtney picks six. By minute three they're showing you the actual Discord, the actual retreat invoice, the actual domain to go grab. The line that ends the episode — 'date the product, marry the niche' — is the only thing you need to remember.
Who's talking.
Where the time goes.
01 · Intro & the draft format
Greg frames the episode: 30+ viral startup opportunities, each host picks six off the list, pressure-test live.
02 · Idea 1: Live & unscripted creator shows
Jonathan's pick — Twitch model applied to business. TBPN sold for $100M+. Anti-AI authenticity is the bet as feeds get sanitized.
03 · Idea 2: Action apps (agent-first mobile)
Greg's favorite category. Apps that act on your behalf — inbox, calendar, expenses — instead of apps that wait for taps. Same setup as the mobile-first transition.
04 · Idea 3: Loneliness + third spaces + niche communities
Jonathan's Dads of Marathon Discord (13,900 members), his $90K painting retreat, two two two (222), Fabric. Membership models around hangout.
05 · Idea 4: Elder tech (65+)
70M boomers, underserved on hearing/mobility/memory/vision. Don't make the landing page look like a retirement home. Facebook ads still work — they're just hitting buyers in their 40-60s.
06 · Idea 5: Adult hobbies
Pottery, woodworking, paper-lamp workshops. People want screen-off joy. Tyler Lemko's Creative Club as template.
07 · Idea 6: AI employees / Juniors
Pick a vertical, pick a job title, list 50 jobs-to-be-done, ship at 1/10th salary. Brand as 'juniors' not 'senior replacement.' Greg gives away the domain junioremployees.ai live on the pod.
08 · Idea 7: Personalized nutrition by vertical
Jonathan has GERD, pays for blood work, gets nothing actionable, built his own Stomach Helper Claude project. 60M Americans with GERD. Zoe + AI + meal delivery, narrowed to one condition.
09 · Idea 8: Pet health + AI for animals
$140-150B pet industry, <2% smart monitoring. PetPace, Whistle. 'Look at who sponsors Huberman, apply to pets.'
10 · Idea 9: AI-native media (done right)
Rowan Chung built 400K followers in 18 months with AI avatar. Slop loses; top-1% AI-assisted media wins, especially with a product to sell.
11 · Stacking ideas: live show + retreats + entrepreneurs
Jonathan's real play — unscripted live show builds a creative-entrepreneur audience, monetized via high-ticket retreats. Same painting retreat = $5K to general public, $110K to entrepreneurs. Niche selection matters more than category.
12 · Final thoughts: date the product, marry the niche
Greg's portable rule. Product can change as you learn. Niche, you marry — pick something underserved, with disposable income, and a real pain point.
Lines you could clip.
"Date the product, marry the niche."
"Where are the grandpas of Marathon?"
"Fish where the fish are. Everyone is fishing... where there are no fish."
"Brand it as juniors, not senior replacement. Every company needs juniors."
"Someone go grab the domain. junioremployees.ai is available. Go and build this company."
"Today's apps assume a human taps and scrolls. The agent-first version just does the work and shows you the exceptions."
"Human beings are looking for the path of least resistance. We don't want to cook the steak. We want the steak."
"I cannot come to this event. I need to be on Claude."
"Almost a quarter of Americans have no close friends."
"Slop loses. The top 1% AI-native media wins — especially when paired with a product to sell."
"Same painting retreat is $5,000 to the general public. $110,000 to entrepreneurs. Pick your buyer."
Things they pointed at.
Word for word.
Steal the format.
Two operators draft picks from a public list, give away the domains live, and close on a portable rule the audience can repeat at a dinner party.
- Publish the master list FIRST (tweet, video, blog) so the episode IS the pressure-test, not the reveal. The viral post pre-sells the watch.
- Format = draft. Two people, alternating picks, six each. Built-in disagreement, no scripted segments needed.
- Give away one domain or one brand name live, by minute 40. The 'someone go grab junioremployees.ai' moment is the most-clipped 30 seconds.
- Land the episode on a six-word rule. 'Date the product, marry the niche' is the line. Make the title-card and thumbnail variant about THAT line, not the listicle count.
- Cut to screenshare only when the visual adds info (cards, charts, Discord screenshots). Otherwise stay two-cam — the friction is the entertainment.
- Make the CTA invisible. No newsletter pitch, no sponsor. Just hit subscribe buried in a self-deprecating 'that's why you listen' aside.
- Stack the offer: live unscripted show > narrow audience (entrepreneurs) > high-ticket retreats. Each lever multiplies the next.
What this could mean for you.
Most builders fail because they marry the product and date the niche. Flip it.
- Pick your buyer before you pick your product. Entrepreneurs vs. general public on the same retreat = $5K vs $110K. The price is the buyer, not the offer.
- If you're building anything with AI agents in it, brand them as juniors. Lower trust bar, more obvious value, no political baggage about replacing seniors.
- If you have a chronic health condition, build your own helper-project (blood work + symptoms + GPT) before you wait for a SaaS to solve it. Jonathan's Stomach Helper isn't a product yet — it's just a Claude project. That's enough.
- If you're over 40 and feel under-marketed-to, you're right — and you're also the most valuable customer most companies aren't going after.
- Loneliness has a Discord-shaped solution. Find the niche game / hobby / chronic-condition group you actually belong to, join the existing community, don't try to build a new one.
- Don't make AI content. Make AI-assisted content where the human still curates the top 1%. The slop is losing the floor; the top is opening up.



































































