You can put this: > some text

at the beginning of a line, and it looks like

some text

Generally you can find info about editing text on LessWrong here: https://www.lesswrong.com/w/guide-to-the-lesswrong-editor

Anyway, I think it's going to be pretty hard to engage by text on this, because it seems like you're coming with background assumptions that are either misconceptions, or at least very different from where I'm coming from. So the right way to have this conversation would be synchronously, so we can clarify things quickly. Purely as an example, to illustrate that general fact, you say

if it is possible to install those values on a genetic level it would be extremely difficult to remove or contain them.

I don't know what you mean by "contain" them. And I don't see what you could mean by "difficult to remove"--if you can genomically vector a future child to go from a trait value of X1 to X2, it is (usually, roughly) equally easy to go from X2 to X1.

I used the word "personality" as a synonym for "values" or "preferences", so things like "liking math" count as personality while things like "bravery" don't.

Oh. (I'm not sure I totally believe you; e.g. you wrote "... create a generation of impressionable children ready for brainwashing, or unempathic psychopaths for military use", which sounds like it's about personality. But the lines are blurry.)

Then I think I agree with you more, though IDK if we fully agree. Like, at a simple level, yeah; I agree parents shouldn't be micromanaging their future child's preferences. I probably agree that we should not be doing research in order to enable that specifically.

Note that this is probably even harder in most cases than vectoring personality traits. Something is probably vectorable if and only if

1. It's fairly well-defined / a coherent thing (rather than a random grab bag of different things).
2. You can measure it pretty cheaply and sorta accurately, so you can collect data at scale.
3. You actually do collect the data (phenotype/genotype pairs).
4. It is significantly determined by many genetic variants.

A lot of preferences will fail some of these criteria. Some might meet the criteria. E.g.:

1. Some preferences will correlate at least a little with vectorable traits. E.g. probably smarter people are, at least on average / in aggregate, more likely to get really really into math. So parents could, if they wanted to, nudge preferences.
2. Some preferences are interesting enough to people that they get measured en masse. E.g. religious or political affiliation.

I think these will tend to be border cases, where the line between personality trait vs. narrow preference is unclear, and the effect on narrow preferences might be mostly mediated by effects on general personality traits. So, the remarks in my previous comment still apply.

Obvious enough but worth saying explicitly: It's not just impacts on epistemology, but also collective behavior. From https://www.lesswrong.com/posts/sTDfraZab47KiRMmT/views-on-when-agi-comes-and-on-strategy-to-reduce#My_views_on_strategy :

Suppose there is a consensus belief, and suppose that it's totally correct. If funders, and more generally anyone who can make stuff happen (e.g. builders and thinkers), use this totally correct consensus belief to make local decisions about where to allocate resources, and they don't check the global margin, then they will in aggregrate follow a portfolio of strategies that is incorrect. The make-stuff-happeners will each make happen the top few things on their list, and leave the rest undone. The top few things will be what the consensus says is most important——in our case, projects that help if AGI comes within 10 years. If a project helps in 30 years, but not 10 years, then it doesn't get any funding at all. This is not the right global portfolio; it oversaturates fast interventions and leaves slow interventions undone.

The idea of possible life paths, and genomic choices aimed at helping some vs. all life paths, seems important. It's been discussed some in the literature. See for example "The Illiberality of Liberal Eugenics", Dov Fox, 2009. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1072104

I'm going to be reading the first part of this essay in about an hour: https://www.youtube.com/watch?v=XllhegEy1K8 I estimate I'll be done very roughly around 12:50 Pacific Time, at which point I'll take questions or chat with anyone who wants. If you like, you could jump in and we could have a chat about your thoughts. I think your intuitions are only partially worked-out, or at least I only partially understand them, so I'd be interested in discussing (or debating), to help you and/or me understand your thoughts better.

A few things to note:

• What's the moral distinction between genetic vs. environment/parenting effects on children's personalities? Parents affect their children in all sorts of ways. Are you also saying parents should not, for example, try to nudge their children, through behavior, to be more kind, wise, brave, perseverant, etc.?

• Personality is less heritable and harder to modify than IQ and many disease traits. This means that for the time being, the profile of traits we can affect is more unambiguously good. Further, I argue that it's good to pursue reprogenetics with speed, partly to avoid an overhang of capabilities: https://www.lesswrong.com/posts/rdbqmyohYJwwxyeEt/genomic-emancipation#Appendix__on_safer_sooner_reprogenetics

• Consider bravery. To have bravery that is a bit above average, rather than exactly average--is this not also a good for most life plans? What about willpower?

• It is (almost certainly) true that if you can increase trait X, you can also, technically speaking, decrease it. But that doesn't mean this would be easily accessible. While I'm probably generally strongly in favor of few state regulations on genomic choices (https://berkeleygenomics.org/articles/The_principle_of_genomic_liberty.html), I am probably broadly in favor of clinicians / genetic counselors / self-regulatory organizations making more decisions about what sorts of genomic choices to make available to parents. I think clinics shouldn't offer "make your kid super obedient", and I think most of them wouldn't. Though, I don't know. (Reminder: "make your kid super obedient" is not something we know how to do, I'd guess not something we'll soon know how to do, and not even obviously doable any time in the foreseeable future.)

• You may be underestimating the controversialness of your judgements. Specifically, I've met plenty of people who are lukewarm, skeptical, or even anti-interested in increasing their own future child's intelligence. I've also met people who consider IQ / what is usually called "intelligence" to be mainly a particular personality trait, akin to openness / extroversion / bravery, rather than a general capability. Now, both sets of people are probably wrong, according to me. But part of what we're doing here is figuring out how to make a political coalition, and a political stance--that is, a group stance that society as a whole can take--that should be very desirable to most and acceptable to nearly all.

• As I argued, it's not clear to me how it's better to have my personality traits set by random genetic dice, or by weird alien optimization pressures coming from bioevolution and from Moloch-style human sociobiological pressures. There's no "fair" / "neutral" / "pure" background default. (I do think there's significant nuance here, but overall the conclusion stands. Happy to discuss.)

if parents get to chose their children's personalities this would make it easy for them to forcibly align such children with their own values

Again, you may have an intuitive picture that's somewhat wildly out of touch with what is practical. As I mentioned in the essay:

• Currently we can barely nudge any personality traits.
• Even if we could, we would not be able to greatly decrease the variance in a given future child's personality traits. Even if you make your future kid have an expected average "Obedience Quotient" of 130, you still have a few percent chance of having a kid with OQ = 100, and > .1% chance of OQ = 85.
• So Sally1 in a sense can speak for herself, even without Sally2. She must be informed of her parents's genomic choices for her, and then (with her randomly low OQ) can be outraged and say so. This does pose quite an interesting problem though--how to weigh different voices from children who had some type of genomic choice made for them.

.... Also are you calling Sneetches chickens??

Interesting. That is bizarre feedback to me, since I like images (at least I think I do?), so I wouldn't have imagined it being disruptive. Therefore it is very informative/helpful feedback, thank you.

(I might just keep using images because I like them. Note that the images are a bit less intrusive at https://berkeleygenomics.org/articles/Genomic_emancipation.html , though sounds like they'd still be intrusive to you. IDK how to use the LW markdown to resize images--I think it automatically full-widths them.)

I'll be reading the first part of this essay, excluding the appendices, on Sunday, 12:00 PM Pacific time (roughly 50 min), followed by discussion / Q&A: https://www.youtube.com/watch?v=XllhegEy1K8

E.g. "...and it seems like there could exist good paradigms for this area, and we probably want good paradigms for this area, and our current work in this area ought to be shaped by the fact that we're pre-paradigm, and...."

It's the same statement, plus an additional set of implications that come from reification. I want to say "mech interp (and AGI alignment and other things) is pre-good-relevant-paradigm". (Which people have been expressing as "pre-paradigm".) I want to say

• There's this category of areas called "pre-good-relevant-paradigm".
• Mech interp is one of those.
• This category has a bunch of features in common that make them a cluster / a Thing.
• This Thing has implications, e.g. about how to orient to research and fieldbuilding for that area.

This is much more easily done with a word for an adjective-like concept. It plants a flag and asserts its Thinghood. People can talk and coordinate about it. Words are good.

FWIW I also feel a bit bad about it being both commercial and also not literally a LW thing. (Both or neither seems less bad.) However, in this particular case, I don't actually feel that bad about it--because this is a site founded by Yudkowsky! So it kind of is a LW thing.