Eric Jang  🇺🇸 🇹🇼

1/n Adding some detail as to why this is would be a good benchmark. The set of ingredients & materials is small (eggs, bowl, plate, frying pan, stove, rubber spatula, seasoning)

2/n Despite limited set of raw materials, the cooking techniques are highly varied and still exercise many kinds of dextrous manipulation techniques

3/ Cooking eggs is very timing-sensitive - forces robots to be quick. Slow robots (academic manipulation research does not optimize for this) would fail instantly.

4/ current computer vision algorithms have a very pose-centric, object way of understanding scenes. Having to manipulate an amorphous scramble in a pan or watch a poached egg disappear under a swirling mass of egg white would likely break those abstractions.

5/ it's still hard to imagine general purpose robots with our current software and hardware, i.e. defining useful tasks for robots without solving AGI. I think an "eggs any style" robot would be a concretely useful, well-defined problem scope with real utility to real people.