Quibi
- We believe: people will watch premium 10-minute shows alone on their phones, between other things.
- To verify, we will: cut existing shows into 10-minute chunks and give a phone-only beta to 500 commuters, with no sharing or TV casting, same as at launch.
- And measure: how many come back in week four.
- We are right if: at least 40% are still watching in week four.
Cheapest test: the beta. COVID got blamed, but nobody ever proved the commute habit in the first place…
Humane AI Pin
- We believe: people will talk to a pin with no screen instead of reaching for their phone.
- To verify, we will: run a “Wizard of Oz” test where 20 people wear a clip-on mic for a week and a real person answers through an earbud.
- And measure: pin uses per day compared with phone pickups.
- We are right if: by day seven, the pin replaces at least 25% of phone pickups.
Cheapest test: a clip and a person. If people won’t use it even when a human gives perfect answers, better AI won’t save it.
Juicero
- We believe: you need the press to get good juice out of the packs.
- To verify, we will: give 10 people the juice packs and no machine.
- And measure: how much juice they squeeze by hand compared with the press.
- We are right if: hands get less than half of what the press gets.
Cheapest test: ten minutes and two hands. Bloomberg ran it for free in 2017, and squeezing by hand got almost as much juice.
