Taggar says that as he thinks about earlier examples of YC startups that tried a new product release or something else that had caused worry and anxiety before it was introduced, he cannot think of a single instance in which the outcome was “unfixably bad.” He reminds them that experiments offer an opportunity to learn from users. “You try out some

Graham has his own term but it is a clunker: a “quantum of utility,” which means, in his words, a product that would make the world “one incremental bit better.”

Paul Graham is fond of a Silicon Valley saying: “You make what you measure.”6 The act of measuring some aspect of performance leads to caring about that one thing and lavishing your attention on it, which, in turn, leads to improving it. He suggests that startups set a specific weekly target for growth, which can be measured in terms of revenue or

‘What do I wish someone would start a startup to do for me?’

“Here’s how to generate new ideas. Three things. One: founders are target users. Two: not many people could build it, but founders are among them. Three: few people realize it is a big deal.”

One meaning of the verb form “to hack” refers to finding a clever way to beat a system.

If you want to, you can spend all of your time refactoring your code or rewriting how you deal with database seeks or something—you could do that and that would be ‘work,’ in some sense of the word. But it wouldn’t be work in the sense of getting you new users or anything that was actually important.”

The session—or “batch,” which can refer to the session as well as to the group—culminates in Demo Day, at which the teams have the opportunity to make a pitch to several hundred investors.

“If we take 6 percent, we have to improve a startup’s outcome by 6.4 percent for them to end up net ahead,” he said. “That’s a ridiculously low bar. So the IQ test is whether they grasp that.”5 When Graham finishes his last