David Mears

Team lead @ EAGxCambridge
1104Joined Aug 2021


By the way, the finding of an IQ gap isn’t (or shouldn’t be?) what is under contention/offensive, because that’s a real finding. It’s the idea that it has a significant genetic component.

I think both Bostrom and I claim that he does not believe that idea, but I’ll entertain your hypothetical below.

I think that, in the world where racial IQ gaps are known not to have a significant genetic component, believing so anyway as a layperson makes one very probably a racist (glossed as a person whose thinking is biased by motivated reasoning on the basis of race); and in the world where racial IQ gaps are known to have a significant genetic component, believing so is not strong evidence of being a racist (with the same gloss). There are also worlds in between.

In any of these worlds, and the world where we live, responsible non-experts should defer to the scientific consensus (as Bostrom seems to in 2023), and when they irresponsibly promote beliefs that are extremely harmful and false, through recklessness, they should apologise for that.

I don’t think anyone should apologise for the very act of believing something one still believes, because an apology is by nature a disagreement with one’s past self. But Bostrom in 2023 does not seem to believe any more, if he ever did, that the racial IQ gap is genetically caused, which frees him up to apologise for his 1996 promotion of the belief.

As a reminder, the original description I took issue with was:

Someone who is clearly not racist accidentally said something that sounds pretty racist, decades ago, and then apologized profusely

It ‘sounds pretty racist’ to say “blacks have lower IQ than mankind in general” because that phrasing usually implies it’s congenital. In other words, in 1996, Bostrom (whose status as a racist is ambiguous to me, and I will continue to judge his character based on his actions in the coming weeks and months) said something that communicates a racist belief, and I want to give him the benefit of the doubt that it was an accident — a reckless one, but an accident. However, apart from apologising for the n-word slur, I haven’t seen much that can be interpreted as an apology for the harm caused by this accident.

Now, if Bostrom, as a non-expert, in fact is secretly confident that IQ and race correlate because of genetics, I think that his thinking is probably biased in a racist way (that is to say, he is a racist) and he should be suspicious of his own motives in holding that belief. If he then finds his view was mistaken, he may meaningfully apologise for any racist bias that influenced his thinking. Otherwise, an apology would not make any sense as he would not think he’d done anything wrong.

The lack of apology for promulgating accidentally (or deliberately) the racist view is wrong if Bostrom does not hold the view (/any more). He is mistaken when in 2023 he skates over acknowledging the main harm he contributed to, by focusing mostly on his mention of the n-word (a lesser harm, partly due to the use-mention distinction).

One of the main complaints people (including me) have about Bostrom's old_email.pdf is that he focuses on the use of a slur as the thing he is regretful for, and is operating under a very narrow definition of racism where a racist is someone who dislikes people of other races. But the main fault with the 1996 email, for which Bostrom should apologise, the most important harm and the main reason it is racist, was that it propagated the belief that blacks are inherently stupider than whites (it did not comment on the causation, but used language that is conventionally understood to refer to congenital traits, 'blacks have lower IQ than mankind in general'). Under this view, old_email.pdf omits to apologise for the main thing people are upset about in the 1996 email, namely, the racist belief, and the lack of empathy for those reading it; and it clarifies further that, in Bostrom's view, the lower IQ of blacks may in fact be in no small part genetically determined, and moreover, as David Thorstad writes, "Bostrom shows no desire to educate himself on the racist and discredited science driving his original beliefs or on the full extent of the harms done by these beliefs. He does not promise to read any books, have hard conversations, or even to behave better in the future. If Bostrom is not planning to change, then why are we to believe that his behavior will be any better than it was in the 1990s?"

So in my view: in total, in 1996 Nick endorses racist views, and in 2023 he clarifies beyond doubt that the IQ gap between blacks and whites may be genetically determined (and says sorry for using a bad word).

A more detailed viewpoint close to my own from David Thorstad: https://ineffectivealtruismblog.com/2023/01/12/off-series-that-bostrom-email/

Please explicitly make the argument, rather than linking a definition. I'm inclined to take this as a tribalistic "that's communism!" remark, which is, indeed, unhelpful.

I guess that the motivation for OP was that people were referring to Bostrom's apology as evidence that he sincerely repented, and deserves to be welcomed back into the fold already; whereas in fact the apology provides scant evidence of sincere introspection and remorse, and so we should not treat him as redeemed, yet. OP describes the way the apology fails to provide this evidence, without which there's no cause for redeeming him yet.

Perhaps unlike OP, I don't want Bostrom to write a false apology by following those rules. Nor do I want a lazy or perfunctory apology to be accepted by the community. We should welcome Bostrom back into the fold on certain conditions, namely, that he is sincerely remorseful; and writing a dysfunctional apology doesn't get him closer to meeting that condition.

If we successfully coordinated to withhold our acceptance until he makes serious amends, we may in fact succeed in causing him to introspect and change more than he otherwise would. Or he might just lie about his remorse. But accepting a bad (or non-) apology throws away the possibility of Nick introspecting.

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.

Someone who is clearly not racist accidentally said something that sounds pretty racist, decades ago, and then apologized profusely.

Registering strong disagreement with this characterisation. Nick has done vanishingly little to apologise, both now and in 1997. In the original emails and the latest apology, he has done less to distance himself from racism than to endorse it.

While I think it can make sense to model whole organisations as having traits like 'truth-seeking' or 'having integrity' or 'transparent', particularly when they are small and homogenous, it's always worth remembering that organisations are made up of people, and those people can vary a lot along all those traits. For example, CEA's character could change rapidly after hiring a lot, or if they lose one exceptionally conscientious person, etc.

Not totally sure, but IIRC characters like 'a' or 'z' are about 8 bits each, depending how the text is encoded. So 48 bits would give you 6 characters.

It’s embarrassing for Bostrom to claim this as an apology.

Load More