[paulryanrogers]

Survivorship bias could be at play. I'm a fan of 20% time. Though I think we'd need more data than some anecdotes to draw broad conclusions.

[kergonath]

It can be good for your engineers’ sanity even if it does not end up a unicorn, though.