Hacker News Reader

Comment by maratc | Hacker News Reader

This kind of science leaves me with lots and lots of questions.

E.g.: As our understanding of DNA progresses, and we link more and more of it to different traits, will we eventually get into a situation where a swab on the water glass tells potential employer much more than the interview?

Or: imagine that in addition to the discovery "viral DNA in the human genome linked to major psychiatric disorders" there's a discovery that "this viral DNA is much more prevalent in [a certain population]". What will be the consequences of this?

View on HN

> E.g.: As our understanding of DNA progresses, and we link more and more of it to different traits, will we eventually get into a situation where a swab on the water glass tells potential employer much more than the interview?

For almost anything that an interview is designed to discover: no. DNA manipulates probabilities of certain things in one direction or another, but then that probability shift is amplified and attenuated through layer after layer of immensely chaotic biological, then psychological, then social processes.

The connection between genetic code and "things we care about" is extremely finicky, with the exceptions being certain types of diseases which emerge much closer to the genetic code than any sort of social behaviors do.

Once we get past the diseases research, what's stopping us from going further? Could we discover e.g. a DNA sequence in human genome "linked" to higher intelligence, or better leadership qualities? And if yes, what can guarantee that instead of the Nazis' "master race" based on pseudo-scientific scull measurements we won't discover "master genes" based on actual DNA science?

What frightens me is not what we can discover, it's what we as human society will do with that information once we discover it.

I think we're past that. Setting aside politically hot topics like IQ, Height is genetically determined and makes a huge difference in life outcomes. Now it's all about what we do with that information.

You can tell a person's height just by looking at them — unlike other examples, you don't need any advancement in DNA science for that.

That makes height an easy topic to talk about, but it's still challenging to think through the implications. If you can increase your child's height by 20% and thereby increase their likely income by 25% - is it morally right not to? Does it matter whether others can afford that intervention? Is it morally right not to fund public access to such interventions? Does the answer change if relative height is all that matters?

On height, that is already a reality.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5956020/

https://www.statnews.com/2019/11/18/a-new-treatment-promises...

I mean, read up on the history of eugenics. Mainstream establishment science has come up close to doing exactly this sort of thing in the past (including sterilizing people who were considered "unfit").

It seems almost certain that we will continue to find more DNA sequences that are correlated with increased intelligence.

Because DNA won't tell you whether the genetically-genius person smashed their head on a rock and is now totally unfit for your high-paying engineering job.

Nor will it tell you whether the genetically-average intelligence person spent 14 hours per day for the last 8 years in the code mines building awesome stuff.

I.e. it won't tell you the types of things that interviews are designed to suss out.

Any test mixes signal with noise (Maybe that superstar candidate is just awful at interviews? Maybe the super lazy coder puts their laziness aside only for the interview day?)

Combining many noisy factors into one, with the correct weighting, will always, on average, give better results than a single test.

That means that as soon as DNA tests give even a weak prediction of 'iq' or 'leadership abilities', companies will integrate them into hiring processes if legal and cheap.

If there were single genes that produced high intelligence scientists probably would have found one by now. Those sorts of traits are almost certainly highly polygenic. Any single gene is likely to have both small and unpredictable contributions to the traits in question. Even totting up all your "good" and "bad" genes to try to come up with a genetic score is fraught due to complicated interactions; perhaps two "good" genes interact badly and make a bad result, or similar.

> where a swab on the water glass tells potential employer much more than the interview?

More likely RNA, because DNA doesn't mean traits are actually active. Also, only if you believe nature>>nurture and that experience doesn't matter. The level of understanding of your issues and how to counteract them makes a difference too.

In practice you probably want someone with issues they know how to cope with and lots of enthusiasm for $job rather than a rando with great generic potential and hating every minute at work.

> only if you believe nature>>nurture

Imagine a situation where there are two presidential candidates, who are both somewhat bad, and then it becomes a matter of public knowledge that one of them has "Viral DNA in [his] genome linked to major psychiatric disorders". I don't think what I do or don't believe would matter in the larger scheme of things.

The assumption is that they're comparable with nothing that makes one stand out ("both somewhat bad"). In that case, sure, it could mean something. But you rarely see that situation, because it doesn't make sense to run against someone not very different from you.

I guess on the DNA level there is not much difference between any two candidates for any $job ... yet. The questions I'm asking are not about the current situation, they are about potential (not yet existing) situations we can find ourselves in with the advancement of DNA science.

Well, in the US at least, if the presidential candidate is well supported by their establishment party, it seems like there would be little-to-no difference. (much like my voting activity in first-past-the-post systems)

> will we eventually get into a situation where a swab on the water glass tells potential employer much more than the interview?

So GATTACA (1997), a recommended movie treating the genetic discrimination, genoism, as main plot.

I don't think we will live to see it technologically speaking, but eventually our descendants may live through such hell of statistical speculation.

Thanks for the recommendation, I really enjoyed it!

I think we already know that, it's just there's usually a tradeoff for these things. Subgroups with higher average IQs also have higher rates of psychiatric disorders etc... If society's healthy there'll be lots of opportunities for subgroups to niche down. If it's not, everyone will be chasing the same rabbit and the inherent unfairness is going to compound that problem.

Please don’t give anybody ideas, hiring practices in tech are already so shitty

What makes you think this hasnt already been done?

Reply on news.ycombinator.com