how to come up with a business idea: 6 tactics, from your card statement to the next exponential

Lex asks the obvious question: how do you find a good idea? Levels, on the transcript: "I think you need to be able to spot problems." start with a normal morning: "What is stuff that I'm really annoyed with that's like in my daily life that doesn't function well?" then the method: "make a list of all these problems you have, and an idea to solve it, and see which one is viable." he says travel helps because "travel is uncomfortable" and you see apps abroad that "You can go back to your country" and build. "There's a lot of problems you start seeing, just getting out of your house."

the tactic: collect annoyances before you collect ideas. the problem list comes first. the idea is just the second column.

try this week: keep a note open for five days. every time something in your day does not work, write it down with one line on how you would fix it. on day five, circle the one you could ship a first version of yourself.

Hormozi remembers a college class where almost every business plan was an invention. his point, on the captions: "the vast majority of businesses are not inventions." think lawn care, laundry, car sales, insurance. so where do you look? "look at your credit card statement. Look at everything you spend money on. You are the customer." for each line he asks: "are any of these things are things that I'm interested in or things that I could do?" and he kills the pressure to pick perfectly: "your first business is not your last business. It is your first stepping stone."

the tactic: start from what you already buy. you know the customer, the price and the bad parts, because you are the customer.

try this week: export the last three months of your card statement. highlight every service you pay a person or a small business for. next to each, write one thing you would do better. keep the three you could start with little money.

Heller starts with the old YC line, make something people want, and says it just got easier. on the captions: "what do people want? For example, things they're paying for right now." his examples are customer support, insurance adjusters, paralegals, personal trainers and executive assistants. "we already know what people want because they're paying people to do it." so the search is simple: "now you just have to look what are people paying other people to do." then he sorts each job into a category, which is the free tactic below.

the tactic: start from paid work, not from a gadget. a job someone already pays for is demand you do not have to prove.

try this week: write 20 jobs people pay others to do, at work and at home. for each, mark assist, replace or unthinkable (the free tactic below). pick one and ask two people who pay for it what they hate about it.

Pincus calls this the play he would encourage anyone to consider. on the captions: "Find a mature market that's over, that's done, that's dead, that's been played out." his examples are online dating, eBay-style listings and unglamorous analog businesses. "They're almost not investable. Like VCs won't like them." the filter: "Find a market like that, but has a lot of money in it and has a proven behavior in it." for him that was video gaming in 2007, which he says was barely growing and not fundable, and he started Zynga there.

the tactic: go where the money is old and the attention is gone. a boring market already proved people pay. you only need a new angle on it.

try this week: list five markets people call dead or boring, where you know people still spend. for each, write the habit that never went away. pick one and write one new angle in a sentence.

Tim reads Ryan's own description of the exercise, on the captions: "list 20 ideas on the board of trends that I believe are going to exist for 10 years and then what are the implications that come out of that." Tim admits he skips the second step: "what will exist that doesn't exist today." Ryan explains the horizon: "to build an important company, unfortunately, you can't do it in two or three years. It does take 10 years." and the warning: "if it's something that is hypy today and is going to be done in 2 years or started 5 years ago, you're too late." he is honest about the hit rate: "I've gotten it wrong many times, but sometimes you get it right."

the tactic: list long trends, then their consequences. the business sits in the second column, not the first.

try this week: take a whiteboard or a page. write 20 trends you would bet hold for 10 years. next to each, write one thing it makes possible that does not exist today. cross out anything that is hype this year.

Altman's advice to the room, on the captions: "figure out what the really big brand new thing is that is possible now that wasn't possible a few years ago." he borrows a Paul Graham line: "The world does not understand how to intuit exponentials." OpenAI's bet looked wrong to experts, and that kept rivals away. copying the hot startup has a cost: "you get a lot of hype and you can raise a lot of money," but he says those are less often the big outcomes. if your data keeps backing you, "be grateful that the world doesn't understand." and "you don't need a ton of people" to agree.

the tactic: hold a belief most experts call wrong, and keep testing it. if more data keeps backing you, the doubt is your head start.

try this week: write one thing that is possible now and was not three years ago in your field. list your evidence, and what would prove you wrong. look for the proof each week.