I’d updated a bit towards the move fast and break things / super high ambition / chaotic philosophy of org management due to FTX (vs. prioritise carefully / be careful about overly rapid growth / follow best practice etc.), but I undo that update. (Making this update on one data point was probably also a mistake in the first place.)
...But undoing the update based on one data point is not a mistake?
I’m not being pedantic. I don’t think anyone’s had the bravery yet to publicly point out that the SBF story—both pre-scandal and post-scandal—is just one data point.
A massive ******* data point, to be sure, so let’s think very seriously about the lessons to be learnt—I don’t mean to trivialise the gravity of the harm that’s occurred.
I’m saying that I think the updates we’re all making right now should be much more similar in magnitude to finding out that SBF had almost succeeded but for some highly contingent obstacle in committing an $8bn fraud, than to finding out that 8 self-proclaimed EAs had each committed a $1bn fraud.
I feel like the typical community member is over-updating on the FTX scandal. I don’t really get that sense with you and I love that you’ve included an “Updates I’m basically not making” section.
(Noting that I think there’s a good chance my reasoning here is confused given that I can’t recall this point having been made anywhere yet. Putting it out there just in case.)
I don’t think the update is just one data point. Some of the issues with FTX—over reliance on alignment over domain expertise, group think, poor governance, overemphasising speed over robust execution at times straying into recklessness also exist within part of EA. FTX brings that into very sharp focus but is not the only example.
Yes I think if you make update A due to a single data point, then you realise you shouldn’t have updated on a single data point, you should undo update A. Like your original reasoning was wrong.
That aside, in the general case I think it can sometimes be justified a lot to update on a single datapoint. E.g. if you think an event was very unlikely, and then that event happens, your new probability estimate for the event will normally go up a lot.
In other cases, if you already have lots of relevant points, then adding a single extra one won’t have much impact.
One extra point is that I think people have focused too much on SBF. The other founders also said they supported EA. So if we’re just counting up people, it’s more than one.
...But undoing the update based on one data point is not a mistake?
I’m not being pedantic. I don’t think anyone’s had the bravery yet to publicly point out that the SBF story—both pre-scandal and post-scandal—is just one data point.
A massive ******* data point, to be sure, so let’s think very seriously about the lessons to be learnt—I don’t mean to trivialise the gravity of the harm that’s occurred.
I’m saying that I think the updates we’re all making right now should be much more similar in magnitude to finding out that SBF had almost succeeded but for some highly contingent obstacle in committing an $8bn fraud, than to finding out that 8 self-proclaimed EAs had each committed a $1bn fraud.
I feel like the typical community member is over-updating on the FTX scandal. I don’t really get that sense with you and I love that you’ve included an “Updates I’m basically not making” section.
(Noting that I think there’s a good chance my reasoning here is confused given that I can’t recall this point having been made anywhere yet. Putting it out there just in case.)
I don’t think the update is just one data point. Some of the issues with FTX—over reliance on alignment over domain expertise, group think, poor governance, overemphasising speed over robust execution at times straying into recklessness also exist within part of EA. FTX brings that into very sharp focus but is not the only example.
Thank you!
Yes I think if you make update A due to a single data point, then you realise you shouldn’t have updated on a single data point, you should undo update A. Like your original reasoning was wrong.
That aside, in the general case I think it can sometimes be justified a lot to update on a single datapoint. E.g. if you think an event was very unlikely, and then that event happens, your new probability estimate for the event will normally go up a lot.
In other cases, if you already have lots of relevant points, then adding a single extra one won’t have much impact.
One extra point is that I think people have focused too much on SBF. The other founders also said they supported EA. So if we’re just counting up people, it’s more than one.