There have been calls in the community for more humility and less confidence in our conclusions, and one way to do that is to start valuing inside views less, and pay more attention to trends and reference classes with base rates, also known as the outside view.
An example in crypto
Nathan Young argues that the failure rate of crypto is 5% annually, and if this is right as a base rate, it would have been insane to be so over invested in cryptocurrency, because without unjustified assumptions of specialness, sooner or later EA would lose a whole lot of it's income, and that's what happened here.
Link to the content here:
What are the implications for EAs?
Focus more on base rates and trends, and less on your models.
Have a higher bar for suggesting that base rates won't apply to you or that a trend will end.
Do more outside view investigations and datasets so that you don't need to rely on your inside view here.