Are there any other key questions surrounding this debacle that I’m missing?
Ability for Public Figures to Repent? - Seems important over a span of time for a functioning society - Relevant that this happened once 25+ years ago and no pattern of behavior - Nobody seems to be questioning whether the initial email was wrong to send
Should Bostrom’s email/apology reflect negatively on EA? - Yes from a Bayesian mindset, but very slightly - Depends in part on whether one should disassociate Bostrom’s broader ideas on Longtermism / AI Alignment from his actions 26 years ago—seems reasonable
Does EA’s response reflect negatively on EA? - Depends on if you want EA to have a “strong view” towards these questions; if so, you should probably update negatively on EA regardless since there have been a variety of responses - If you prefer EA to have a diversity of viewpoint (seems reasonable) then maybe positive update on EA
What else should have been included in Bostrom’s apology? - More explicit language around group intelligence differences? (what if he doesn’t feel certain on that / doesn’t want to lean into an Info-Hazerdous-if-True topic?) - More empathy in tone? Seems like a solid criticism—what specifically ought to have been included? (And are we being welcoming to the neuro-divergent?)
Ought someone to convince themselves against “Info-Hazerdous if True” beliefs? - Epistemic integrity (seems unhealthy) - Relative benefits of 10-foot pole position (easier to not talk about it vs showing certainty?)
Ought someone to discuss “Info-Hazerdous if True” beliefs? - But doesn’t that give extremists free reign over ideas? - Seems like empirically it works alright to just not feed the fire? (do conspiracy theories/extremism disagree?)
Does this forecast future bad actions from Bostrom? - Likely not—would look to his pattern over last 25 years - Prediction markets seem to have faith in Bostrom
I don’t think people rejecting Bostrom’s apology are rejecting it on the basis that public figures can’t repent. We just don’t think it was a functional apology.
Yes I believe that would be covered in my 4th point—“what else should have been included in Bostrom’s apology”. I suppose it could be rephrased: “what else should have been included/excluded in Bostrom’s apology”, and one could argue he should exclude the latter paragraphs from his apology.
Are there any other key questions surrounding this debacle that I’m missing?
Ability for Public Figures to Repent?
- Seems important over a span of time for a functioning society
- Relevant that this happened once 25+ years ago and no pattern of behavior
- Nobody seems to be questioning whether the initial email was wrong to send
Should Bostrom’s email/apology reflect negatively on EA?
- Yes from a Bayesian mindset, but very slightly
- Depends in part on whether one should disassociate Bostrom’s broader ideas on Longtermism / AI Alignment from his actions 26 years ago—seems reasonable
Does EA’s response reflect negatively on EA?
- Depends on if you want EA to have a “strong view” towards these questions; if so, you should probably update negatively on EA regardless since there have been a variety of responses
- If you prefer EA to have a diversity of viewpoint (seems reasonable) then maybe positive update on EA
What else should have been included in Bostrom’s apology?
- More explicit language around group intelligence differences? (what if he doesn’t feel certain on that / doesn’t want to lean into an Info-Hazerdous-if-True topic?)
- More empathy in tone? Seems like a solid criticism—what specifically ought to have been included? (And are we being welcoming to the neuro-divergent?)
Ought someone to convince themselves against “Info-Hazerdous if True” beliefs?
- Epistemic integrity (seems unhealthy)
- Relative benefits of 10-foot pole position (easier to not talk about it vs showing certainty?)
Ought someone to discuss “Info-Hazerdous if True” beliefs?
- But doesn’t that give extremists free reign over ideas?
- Seems like empirically it works alright to just not feed the fire? (do conspiracy theories/extremism disagree?)
Does this forecast future bad actions from Bostrom?
- Likely not—would look to his pattern over last 25 years
- Prediction markets seem to have faith in Bostrom
I don’t think people rejecting Bostrom’s apology are rejecting it on the basis that public figures can’t repent. We just don’t think it was a functional apology.
Yes I believe that would be covered in my 4th point—“what else should have been included in Bostrom’s apology”. I suppose it could be rephrased: “what else should have been included/excluded in Bostrom’s apology”, and one could argue he should exclude the latter paragraphs from his apology.