Communities, Buybacks and Gacha Mechanics: The Non-AI Business Ideas Shaan Puri and Sam Parr Pitch

Sam Parr opens by making a case that, even if you are not an AI person, there is still a lot of money to be made building businesses that will grow with AI without actually being AI. In his view, those businesses are communities, and he says he wants to show two examples that he believes have "crushed it."

He describes the two examples as businesses he finds AI-proof and fascinating, and the first of them is a peer group for auto dealers.

Parr introduces NADA, which he believes stands for the National Automotive Dealers Association, a large association for auto dealers. Its program, called NADA 20 Groups, is said to have started in the 1970s. Dealers who want to grow and improve apply to join, and Parr estimates they pay something like six or eight thousand dollars a year to get in.

Members are placed in groups of 24. The groups are non-competitive, meaning participants do not come from the same region, and they meet in the same city three times a year. Each member brings the P&L statement of their dealership and goes through it line by line with 20 others, who can praise or criticize the results. Parr notes that the association has since let in founders of HVAC companies and owners of marinas and boat repair shops.

Parr says that, based on the dues and pricing the group discloses, he would guess it earns around 13 to 15 million dollars a year in revenue. He also says the website looks horribly marketed, though the model itself rests on the simple premise of sharing P&Ls.

Shaan Puri suggests that e-commerce, which he had previously targeted through a community called Club LTV, is too competitive. He would pick a less crowded niche instead, such as dentist practices or lawyers, or any other boring industry where the average person earns about a million dollars a year in take-home income.

Puri says he would create one of these communities and believes it could be a great business that lasts for decades.

Puri asks about White Coat Investor, which he describes as a community of doctors, perhaps dentists too but mostly doctors, who talk about the finance side of their lives. According to the discussion, it is not about sharing research or clinical practice. Members instead raise questions such as whether to buy a rental property and what the cap rate is. The community uses media including Facebook groups and has been running for a long time, and it started as a blog.

The speakers say there are a few others in the doctor space, and Puri says he thinks there is a sizeable opportunity to build a 2.0 version of the same model, because media tools and the professional audience have changed since the blog era.

Puri also points to BNI, a network of local chapters that brings together people with different jobs, such as an accountant and an HVAC technician. The speakers note that BNI was recently acquired and that rumors put the price at over a billion dollars.

Parr explains that he has been learning about investing because, over the next handful of years, he expects the company to eventually buy other communities, since he believes they know how to operate them well. His stated goal is to build a multi-billion-dollar company by owning a variety of communities.

When the hosts joke about a possible slip, Parr clarifies that when he says hundreds of millions, he means profit, and billions refers to revenue. He says a business in this space counts as killing it when it crosses nine figures a year in free cash flow, and he believes several such businesses could be combined into something awesome.

The hosts note that building a business on Facebook groups is risky because the platform changes its rules often, which can cost reach or the ability to add new members. At the same time, plugging into such a large network brings benefits, so the trade-off goes both ways.

Parr says he likes the idea of aggregating super-enthusiasts, or whales, because they drive a shocking amount of the economy. In a mobile game such as Candy Crush, with millions of players, the speaker says 80 to 90 percent of revenue can come from a tiny sliver of players who spend thousands of dollars each. The same dynamic appears in Pokémon card collecting and in marketplaces.

On Twitch, the speaker recalls one of the platform's better features, the gifting sub, which lets a viewer give away 200 subscriptions from a channel's chat, roughly a thousand dollars at five dollars a month each. The discussion also cites a figure that Candy Crush has 82 million monthly players.

The speakers describe how mobile games can set their payout rate freely and trade off churn and growth against player frustration. Parr refers to a person who looks at the source code of apps, though Puri says he does not know exactly how that person gets the code and that the explanation may be partly inaccurate.

According to the account, a player who has paid before but has not paid in 30 days is given a run of success and then hits a wall designed to be frustrating. The wall is optimized to get the player to click the button to pay again.

The hosts say some of the smartest engineers and designers spend their careers figuring out how to get people to spend a little more. Puri calls this sad in a way but also fascinating from a business view. Parr adds that a community operator told him that whales fund almost everything, with the top 1 percent or even the top 0.001 percent paying to keep everything running.

Parr describes a trend of ripping packs, in which a buyer purchases a pack, such as a Pokémon card pack, and hopes to find a valuable card. He says these apps reduce friction by letting the user open a pack on the screen and, if nothing good appears, offer to buy it back. In his example, a 50-dollar pack might be bought back for 43 dollars, and the user can use the amount as store credit to buy another pack.

Parr says the app has done about a billion dollars in revenue in the last year, though he is not sure how that is calculated, since it is partly gross merchandise value. The buyback, he says, is what changed everything, because the user is still hunting for treasure and sees a 43-dollar return as cheap for another try.

For the company, Parr says, it books the full 50 dollars of revenue, gives back 43 dollars, and keeps the seven dollars. Because it never had to sell the card to the customer, it can put the card back into the deck and sell it again to someone else. He also says the founder was a quant at Citadel and Bridgewater, and that copycats such as Logan Paul's Rip It have followed.

Parr says Collector Crypt, which he believes may have been first, benefited from the same vending-machine buyback approach. He says it was doing something like 50 thousand dollars a month and is now doing about 25 million a month.

Parr says gacha is a Japanese term for mystery unboxing. A buyer purchases a mystery ball, opens it, gets something, and tries again. He calls it a toy form of gambling and says it became so popular in Japan that it was banned, or outlawed by law at this point.

American games began copying the mechanic. Parr points to Overwatch loot boxes, where a player might open a rare or common skin and be drawn to keep spending credits, and says FIFA and many other games followed. He adds that mobile games have refined the mechanic further.

Puri explains that an Overwatch skin is purely aesthetic, while a collectible card can be worth real money. He cites the top Pikachu card, which he says can sell for around 15 million dollars. By attaching the chance of making money to an addictive gacha behavior, collectibles tap a psychology that works across many forms.

The speakers say toys use the same approach, including LOL Surprise dolls, squishies, dumplings and mystery boxes. Parr says a friend in the toy business showed him data: when items sold on open shelves at Walmart or Target were placed in a mystery box, revenue was seven times higher. Parr notes that the data across five categories consistently favored the mystery box, even though consumers say they want choice.

The hosts return to an app they discussed three or four years earlier, called Private Photo Vault, which lets users hide photos. Parr describes it as basically storage: users move photos they want to keep private, often sensitive ones, off their camera roll into a hidden vault. The app is reported to have about a million reviews.

Parr says he met the founder at a Ham Pit event. The founder said the app had been discussed on the show, and shared numbers and headcount that Parr describes as a behemoth, though Parr says he cannot reveal the exact figures. The founder turned out to be a solo American rather than an Eastern European operation, as the hosts had expected.

The hosts joke about crypto enthusiasts who feel that AI has taken their thunder. Parr suggests that the true crypto believers are incredibly smart, high-conviction diehards who deserve to get rich, while others still act as if the old hype is ongoing and have not been told that the field has moved to AI.

Asked whether crypto meetups are empty now, Parr says he does not go to meetups and so cannot say for certain.

Parr says there is a lot of talk that AI hacking or quantum computing could break encryption protocols. He says Bitcoin started with about 100 risks and is now down to a couple, and that if Satoshi woke up and moved the coins, or if the actual encryption were hacked, it would cause extreme panic.

The conversation then turns to a reported hack of Binance for about 5 billion dollars. The hosts say Binance had a hack in 2018, and that after it the company created a fund, which Parr describes as a rainy-day fund for users, by taking 10 percent of trading fees. He says that fund reached about 2 billion dollars.

Asked who received the money, the hosts say no one obviously knows. Parr says the last thing he read claimed North Korea took it, and that the method reportedly involved a woman in a romantic relationship with an engineer who physically got his laptop. He says the hacker had enough in the hot wallet to transfer out a large amount, and that part of it was frozen by protocols.

Parr describes how the funds could be moved through services such as Tornado Cash, split across wallets, or placed in a DeFi protocol. He explains that DeFi works like a permissionless bank in which users can borrow up to about 50 percent of the collateral they deposit, and that if the collateral's price drops, the protocol sells it automatically. In the account he gives, the hacker deposited the stolen funds, borrowed stablecoins, and withdrew them, so that even if the collateral was frozen, the stablecoins were already out. The hosts treat these details as what they had read, not as established facts.