Much to my surprise, the SBF trial went ahead—and on time!
I’ve been mostly absent the last few months due to a personal/family situation, but one general reflection:
Much has been said about the importance of evaluating character in the wake of the FTX collapse. I’d also point out the importance of evaluating consistency of good judgment. From the outside perspective (but as a lawyer), SBF’s choices as a criminal defendant appear to reflect quite poor judgment.[1] I’d characterize it as well below average for college-educated defendants who are represented by counsel with enough resources to do a good job. I am less confident that this trend toward making a number of bad judgments could have been detected earlier, but I’d suggest that many people who make really bad decisions have a detectable history of making bad decisions in the past.
All that is to suggest that the community might want to screen for erratic judgment in addition to character and trustworthiness (C&T). Assessing C&T is hard; just ask any intelligence agency—which has polygraphs, FBI agents, intelligence information, and so forth that EA will never have. There are going to be cases in which the result is “mixed,” “middling,” or “not enough information to have decent error bars.” Assessing consistency of good judgment isn’t easy either, but at least we are trying to measure two things rather than just one.
I’d suggest that the combination of problematic character and a lack of consistent good judgment may be significantly more dangerous than either alone. I’m reminded of reports that 20% of business leaders may have traits associated with psychopathy. I’m not endorsing that statistic, and not equating untrustworthiness with psychopathy.
My thesis, rather, is that untrustworthy people who also display a lack of consistent good judgment may be particularly risky in synergistic ways. Thus, it may be prudent to treat a C&T rating of “mixed,” “middling,” or “not enough information” with a good-judgment rating of “poor” or even “mixed/middling” much the same as a C&T rating of “poor” as far as risk tolerance.
I can understand the decision not to plead guilty, especially since it sounds like it would have been a blind plea rather than a plea bargain. I would have made a blind plea in his shoes, but at least understand why he didn’t. And it’s hard to imagine a viable defense here that didn’t include taking the stand. But other stuff—the substack, personally contacting a witness, using a VPN, leaking derogatory information about a witness to the Times—is hard to justify as sound strategy, much less some galaxy-brained EV+ plan.
Do you have thoughts on how one might screen for “erratic judgment”? In particular: are there ways of screening for this which wouldn’t be well correlated with screens hiring managers already normally do?
I could imagine reference checks turning some of this up, for example, but reference checks are pretty standard thing that managers do anyway.
Much to my surprise, the SBF trial went ahead—and on time!
I’ve been mostly absent the last few months due to a personal/family situation, but one general reflection:
Much has been said about the importance of evaluating character in the wake of the FTX collapse. I’d also point out the importance of evaluating consistency of good judgment. From the outside perspective (but as a lawyer), SBF’s choices as a criminal defendant appear to reflect quite poor judgment.[1] I’d characterize it as well below average for college-educated defendants who are represented by counsel with enough resources to do a good job. I am less confident that this trend toward making a number of bad judgments could have been detected earlier, but I’d suggest that many people who make really bad decisions have a detectable history of making bad decisions in the past.
All that is to suggest that the community might want to screen for erratic judgment in addition to character and trustworthiness (C&T). Assessing C&T is hard; just ask any intelligence agency—which has polygraphs, FBI agents, intelligence information, and so forth that EA will never have. There are going to be cases in which the result is “mixed,” “middling,” or “not enough information to have decent error bars.” Assessing consistency of good judgment isn’t easy either, but at least we are trying to measure two things rather than just one.
I’d suggest that the combination of problematic character and a lack of consistent good judgment may be significantly more dangerous than either alone. I’m reminded of reports that 20% of business leaders may have traits associated with psychopathy. I’m not endorsing that statistic, and not equating untrustworthiness with psychopathy.
My thesis, rather, is that untrustworthy people who also display a lack of consistent good judgment may be particularly risky in synergistic ways. Thus, it may be prudent to treat a C&T rating of “mixed,” “middling,” or “not enough information” with a good-judgment rating of “poor” or even “mixed/middling” much the same as a C&T rating of “poor” as far as risk tolerance.
I can understand the decision not to plead guilty, especially since it sounds like it would have been a blind plea rather than a plea bargain. I would have made a blind plea in his shoes, but at least understand why he didn’t. And it’s hard to imagine a viable defense here that didn’t include taking the stand. But other stuff—the substack, personally contacting a witness, using a VPN, leaking derogatory information about a witness to the Times—is hard to justify as sound strategy, much less some galaxy-brained EV+ plan.
Do you have thoughts on how one might screen for “erratic judgment”? In particular: are there ways of screening for this which wouldn’t be well correlated with screens hiring managers already normally do?
I could imagine reference checks turning some of this up, for example, but reference checks are pretty standard thing that managers do anyway.