I paid them so that they can do useful things with my sample, like correlate it with research or do more research on their own. I then volunteered to share more of my data in their surveys, and collected more samples for them for further research. I think it's fantastic that they are sharing this with drug companies. In fact, my main question is: What took them so long?!
My DNA isn't some terrible secret I have to hide in the dark. I'm just one of like 8 billion variants of imperfectly copied ape. Maybe paying 23andme to do something useful with it will be my only lasting legacy in this world, lol.
It's not that I "didn't see it coming", it's that I hoped this would happen from the get-go.
Insurance companies will offer a "DNA discount" and ask for DNA directly, and simply charge more to anyone who refuses, and then grade the amount of "discount" based on the risk profile back from the DNA sample.
If you hold out and refuse, you'll simply get lumped in with the riskiest.
taken to the extreme, if complete data availability allowed to predict the future with enough precision, then everyone would pay an insurance premium corresponding to their insured negative events. In practice insurances wouldn't have any reason to exist. The foundational premise of the insurance concept is that risk is spread across every participant in an equal manner.
While this is an interesting idea, no doubt, epigenetics casts major doubt on whether knowing the DNA expression at a particular time is diagnostically useful.
Sure. There are some conditions that are purely genetic, but many genes can be switched on and off depending on the environment - or even recoded in the replication process. DNA isn’t as static as we once thought and knowing a person’s DNA is not quite as useful as we imagined.
This isn’t as big of a deal as we imagined. It’s going to take some incredible processing to uncover causal patterns, and a huge amount of experimentation to determine whether they are stable against epigenetic changes. In my opinion, this is quite exciting.
Why wouldn't insurance have a reason to exist? Even the healthiest people need medical insurance because anyone can experience accidental injury requiring hospitalization and/or on going medical care.
This is presuming perfect knowledge; it's a thought experiment. In this hypothetical world, your insurance company does know that in 13 months you're going to get in a car accident that flings a 1.27mm shard of glass into your eye, that it's going to take 27 minutes to complete the surgery that fixes it, that you'll have an infection after that and the first 2 antibiotics won't work but the third one will.
It's basically just pointing out that perfect knowledge makes risk 0, and that insurance doesn't make sense in a world with no risk. At that point the question is basically just whether you distribute medical costs evenly across everyone or let there be winners and losers on those costs.
Then only insurance for actual accidents would make sense, but even then insurance companies could start analyzing more meticulously the individual probabilities of having accidents: do you have an active lifestyle? you pay a premium; do you practice sports? you pay a premium; travel more than the average human? pay a premium; are you getting older? insurance rate grows exponentially; already had an accident? sorry, you no longer a desirable customer.
> The foundational premise of the insurance concept is that risk is spread across every participant in an equal manner.
Applying this broad principle to auto insurance, it isn't doing well. Auto ins is ~mandated. Participants are forced to pay for risk pools they aren't part of.
ex: Car repair costs are skyrocketing. Premiums are also skyrocketing, including for people who don't carry collision.
Sure. But insurance companies leverage that mandate; they make people who don't choose (or can't afford) collision policies subsidize people that do. With this, the bargain of shared risk is quietly killed and replaced with exploiting the poor (or frugal) to pay for the non-poor (or lavish).
People in the UK should take note of this thread: this is what the private health industry dearly, dearly wants for you too. Treasure what you have while you still have it.
UK has done the largest sequencing project on Earth, the UK BioBank, and shares the data with both academia and private companies to do drug discovery. NHS is also chronically underfunded.
But as no one (yet) needs private health insurance, and very few people have it, it is not (yet) a societal worry that the data will be used to "offer a choice" to people based on genetic misfortune (the choice is bankruptcy or death).
> NHS is also chronically underfunded.
That's very much by ideological design. A recent ex-health minister even co-wrote a book which contained this:
> Our ambition should be to break down the barriers between private & public provision, in effect denationalising the provision of health care in Britain.
Notably he became health minister after this, not before, so the appointment says a lot about what the Party has in mind.
Note: Compliance departments in insurance companies are basically intended to facilitate by rote presentations of trainings to check off that list item as a required business operation.
Actual teeth to bite back or remidiate non-compliant activity of a company that the executives have decided to file under the "fuck it, nobody's looking" risk heading are basically not there.
Point being, just because the law is on the books doesn't mean it isn't being elided at every opportunity. You have been warned.
I have not done extensive research, but I suspect this is as effective as banning the government from data collection, with the result that the government pays companies to do that and it's somehow legal that way.
Not OP, and I agree that research needs to get access to DNA for medicinal, personalization and drug-development research purposes. What I'm more worried about is this sort of data being used to train a neural network that maps genes to physical traits, which would be both a goldmine and Pandora's box.
If I could compare it to something recent, it's the whole Content -> LLM -> Content loop that was created and we're all upset over now to various degrees. Our DNA is the content and up until now has been an opaque black-box with only minor views into it for very specific genetic diseases. Once they open up the box that allows them to get to things like IQ, height, muscle-density, resistance to diseases, fertility, etc, then the human race is in for a wild down-hill ride.
I worked in insurance and they all seem intent on tailoring everyone's insurance policies based on any minutiae they can dig up on them. For example they found out for life insurance that people who live in greener areas live longer, so they were trying to obtain geographical data for "greenness".
I asked them what the end game is. What happens when everyone's insurance policy is exactly tailored to them? What is the point of insurance then? If they get it too right they put themselves out of business. They didn't have a good answer.
I expect that marketing, regulation, and accidents will keep insurance companies around indefinitely. The upside of near perfect information seems great, for the insurance companies. They can ensure that the premiums they charge are higher than the expected payout for 100% of their customers. That won't be so good for the folks who need insurance the most though.
>If they get it too right they put themselves out of business.
No, they don't. Getting it too right just means they are able to more effectively compete on pricing since they know the lowest price they can offer while still making profit.
The point you missed is that when insurance is hyper tailored to specific people, you are basically paying for your own healthcare and can cut out the middle man
Uh, no. Insurance is paying for unlikely events, as it was said above. Greener area or not, you can still get run over by a car, event for which you definitely didn't pay. Or you house struck by lightning, or any such insurance cases might arise. The green area only affects the probability of such events, thus the tailoring.
But my point is by hyper-optimising for individuals it will at some point cross a line where it's not really insurance as you understand it any more. The unlikely event is that the insurance company gets it wrong, for example your house floods in an area that was not predicted to flood for millennia. As the insurance company gets better, the chance of an unlikely event gets smaller. How small can that chance get before you decide to just take the risk yourself?
That is assuming all health related costs can be predicted 100% correctly. If there is a 1/3rd chance to incur a cost then insurance lets you get away with only paying 1/3rd of the cost.
In that hypothetical in the best case the insurance company will never lose money by charging someone less than what they will have paid out to them and they will be able to sell this service to other companies, in the worst case they become essentially a bank.
Insurance makes money by ensuring their income (premiums) is greater than their outgoing (payouts).
The value of insurance (and why we pay for their profits) is in spreading risk across populations. They already target particular demographics (when the law allows them to), but I'm talking about targeting individuals.
If the insurance company can predict exactly how much they will be paying out to you in the next 10 years, say, then they just become a glorified savings account with a cost attached (their profit). The value is then essentially telling you how much you need to save. Data companies like 23&Me could then offer this without bothering with the savings account part.
My feeling is people won't accept this and will want actual insurance policies that actually spread risk across populations because people don't want to be told "you need to save twice as much as your neighbour because of your cancer in 15 years time".
Unless they have a crystal ball and can tell you whether or not you will be involved in an accident leading to serious injury (and the nature of those injuries) then you still need health insurance.
I know a guy who was robbed and shot. Had to have surgery and spent a few weeks in the hospital.
I have a neighbor that was jogging, tripped over a bump in the sidewalk and broke his arm. Few weeks in a cast.
I know plenty of blue collar workers who had on the job injuries that required medical treatment.
Your DNA isn't going to help with that. And in America, a simple 4 hour visit to the ER can result in a multi-thousand dollar bill. And if you're unfortunate enough to need a surgery and a multi-week stay then we're talking 5 figures minimum.
As insurance risk estimation errors approach zero, the value of having insurance instead of just a medical savings account approaches zero. Rates for higher risk individuals will become prohibitively high and those people will have no choice but to be uninsured. Insurance companies would be essentially only covering black swan events, as individual premiums would match individual health costs with high fidelity.
As a result the incentive to get comprehensive health insurance rather than just disaster insurance would plummet, as an MSA would be more cost effective and you also get to take advantage of better than predicted health outcomes in the form of interest earned and at EOL a potential inheritance gift.
Sounds great, except you jhave to understand that this would mean that society would somehow need to deal with the burden of the least healthy, who would be very unlikely to purchase, or be able to afford, insurance under this scheme. It’s like trying to charge up front for building roads. You just end up with no roads. Socialism (or in the case of insurance,capital funded socialism) is extremely beneficial to society in certain specific verticals.
The Genetic Information Nondiscrimination Act (GINA) of 2008 already forbids them from doing anything.
>GINA prohibits health insurers from discrimination based on the genetic information of enrollees. Specifically, health insurers may not use genetic information to determine if someone is eligible for insurance or to make coverage, underwriting or premium-setting decisions.
"Efficiently priced insurance": do you mean so the insurance company can set your premiums based on a more-accurate estimate of your risks?
The whole point of insurance is (or was) pooled risk. The more the insurance company knows about my risk (even risks I don't know about), the less point there is in me buying insurance; I may as well just put my premiums into a private sinking fund.
So even if your insurance worked out cheaper you would now see it as not worth buying because it’s too accurately priced? Literally what are any of you talking about, this is crazy talk. You buy insurance to hedge against risk, that’s it, if the cost accurately represents the risk then that’s a good thing. If you have unlimited money then it’s always financially better to not buy insurance, that’s correct right now… People don’t have unlimited money to spend on massive, unexpected acute financial obligations, that’s why people buy insurance… ITT: people not even understanding why they’re spending money on insurance every month, this is mind blowing to me.
GSK doesn't have your well being as it's goal, it has it's own revenue stream as its goal. Without an ethical framework around consent for the use of this data the potential for abuse is substantial, upsides are optional. They need to make that money back somehow.
If my worthless genomic data contributes to a drug that one day helps even one other person, I'm fine with this.
Would I prefer all this data be free and open? Sure. Failing that, would I want to hold onto it? No. I'd still rather some mega evil pharma have it and be able to develop some overpriced medication that may one day be genericized. It's not doing me or anyone any good just sitting in my body.
Hell, I'll spit for anyone wanting to do research, for profit or not.
I think there are very few true "good guys" in the world. But if mega evil pharma corp has even a 1% chance of accidentally doing some tiny good as a byproduct of their evil... that's already more than would've happened if 23andme didn't share it with them. It's cool with me if they make some money in the process. My DNA isn't some creative work that I need royalties on. What nature giveth, I spiteth.
No kids and never will (vasectomy). Parents couldn't care less. They were intrigued by the ancestry stuff though. They're not long of this world anyway and probably won't live to see genetic medicines become an everyday reality :(
My partner's bigger family also contributed a lot of their DNA to this or similar services. I'll have to ask her what they think of sharing it with third (fourth?) parties. Be an interesting data point at least. I feel like the HN crowd holds a lot of uncommon opinions (not good or bad, just different from most people I know IRL).
Ok, that makes you a minority in the sense that I think that the potential negative fall out from your decision is limited enough that you were entirely free to make it. But the bulk of the people that sent their family data to 23andme isn't that well informed and likely has no idea about the potential for abuse of such data.
Maybe it is the 'dark side' in me or maybe it is simply seeing so many corporate abuses over a lifetime in IT that I can't see beyond the abuse to the possible good. The fact that it has Google backing is one major strike against it, as is the fact that they were going to commercialize the data itself right from the get-go whilst giving their paying customers something the amounted to expensive infotainment. It's clear that they wanted the data, not clear that there would be benefits to the participants, the data returned is just too coarse for that and akin to doing a lot of screening without symptoms, a practice that the medical profession has warned against repeatedly.
I don't disagree with what you're saying (that people don't always understand the implications of their casual dealings with big evil corps).
But can I offer another perspective, one of potential differences in values?
Privacy, for me, isn't a super high concern. Having some is nice, but it's never my top consideration. (I know this skews against the dominant HN sentiment, and Slashdot's too before it. That's okay.)
Science, however, IS huge for me -- even when the science is done by sub-optimal orgs like your Big Evil Corp du jour. When I learned that Google owned 23andme, my reaction wasn't one of horror.
Instead, it was "Oh cool! Maybe they'll actually have the resources to correlate this with all their other data on me." I then proceeded to sign up for Google Health, Google Fit, etc., and made sure to upload my health readings to their servers in the explicit hope that they would be able to merge it with 23andme and other data. My dream was that one day they would be able to identify genetic markers not just for health and lifestyle stuff but also behavioral data, like my YouTube preferences and Gmail conservational styles.
Gattaca to me was a utopia, not a dystopia.
I know that's probably not a common viewpoint. But I was really excited by the possibilities, and it doesn't bother me in the least that Google knows so much about me. I just wish they'd do something useful with all of it.
FWIW, I just asked my partner about the data sharing. She said she's fine with it and would opt-in if given the choice. Most of her family too (many were early adopters of genetic testing). One individual in particular skews conservative and is conspiracy-theory-prone and would probably not, though.
I wonder how much of this preference is political or personality or similar. Would be interesting to plot privacy needs vs the Big 5 personality test, maybe. But how would you get participants? Lol.
> It's clear that they wanted the data, not clear that there would be benefits to the participants, the data returned is just too coarse for that and akin to doing a lot of screening without symptoms, a practice that the medical profession has warned against repeatedly.
To this point, yes, you're probably right, but these things are rarely black and white. It's a false dichotomy. They can be harvesting all my data but STILL offer a useful service at a good value to me. And -- more importantly -- nobody else stepped up to offer the same thing with a better business model. Whether it's 23andme or YouTube, it's not necessarily the case that all their customers have their blindfolds on. Some of us are just OK with the tradeoffs and buy into it willingly. The medical profession isn't always able to meet consumer desires, to say the least. And even if 23andme didn't do it perfectly... at least they did it. If the medical profession or the government wants to offer something similar, I'd still be interested... they just haven't yet, AFAIK.
> My dream was that one day they would be able to identify genetic markers not just for health and lifestyle stuff but also behavioral data, like my YouTube preferences and Gmail conservational styles
You know that's unlikely to happen. What's more likely is you helping that in the future people similar to you will be paying more stuff because of their genes or being profiled against because of their genes. Yes, there are laws against just as they are against discrimination on the basis of protected classes but the discrimination is still happening. And with DNA info, you can do it better and will be harder to detect. At the end of the day, other things being equal, any of those companies using dna to assess candidates will be more accurate than those than don't and should effectively dominate the market in a red queen race. Surely, it won't happen now (dna sequencing is way too expensive now) but when it happens, it will be hard to stop.
I think this is inevitable too, but... maybe politically incorrectly... maybe we should acknowledge it and roll with it, and help people with worse genes however we can, but encourage genetic modification for future generations?
I don't think this idea of "gene blindness" can really work, long term. We're lying ourselves when we say all people are created equal. They very much are not. That's the whole point of genetics.
But we can still say "but they still deserve respect, compassion, and equal treatment." That applies whether theu are disadvantaged due to race (itself genetic, though more plainly obvious), hormonal (sex and some birth defects), upbringing, culture, whatever.
I'm totally for opening up the genome for cross-examination and all the hard social questions that will necessarily come wit it, rather than avoiding the topic and pretending like it's not a thing.
Social norms will just change over time. Better understanding and usage of our genes can actually advance the species (and hopefully societies with it).
Even I agree with this, despite not having kids. Their lives aren't mine to monetize or publicize. Let them have whatever semblance of a childhood they can have under the already too harsh spotlight of modern media, yeesh.
It was bad enough in the 2000s when my mom liked all my Facebook posts. Real travesty here.
> It's not doing me or anyone any good just sitting in my body.
That's funny dude!
Not laughing _at_ you, but you just nailed self-objectification and
reification better than a whole chapter of social scince jibberjabber
or pschology babble.
Makes sense! I'm also a "yeah they can poke and prod my body and harvest my organs, not like I have any use for them after I'm dead" kinda donor.
At least it's consistently materialist, eh? (as opposed to spiritual dualism, etc.)
This is probably not a controversial stance here on HN, but IMO humans are essentially wet, mostly hairless sacks of proteins and dreams. It's kinda cool that despite that, we are motile and can pass for sentient (some days of the week).
I'm not demanding anyone else give theirs. I'm just happy to give mine.
Different customers of the same company can respond differently to their actions. In this case my opinion is that this a net good. Others can and will disagree, of course!
I wish the data would be combined with things like Vanderbilt’s BioVU databanks, etc, for actual translational research and not for swift for profit research.
This is just another version of “socialized research privatized profit.”
> GSK doesn't have your well being as it's goal, it has it's own revenue stream as its goal. Without an ethical framework around consent for the use of this data the potential for abuse is substantial
I don't like that you seem to be trying to drive a wedge between revenue streams and ethical goods. Measuring how much money societies spend on things is the best measure we have of how much societies value those things.
Companies (and people, including employees and customers of companies) do all sorts of horrible, unethical things. But profit (surplus) is a good thing overall, as is symmetric information and competition.
>My DNA isn't some terrible secret I have to hide in the dark.
Considering we leave our DNA basically everywhere we spend more than a few moments, it seems like a rather impossible task to keep it secret anyway. I mean if we are fearing some dystopian world in which people are discriminated against based on their DNA profiles, why would that only apply to people who volunteered their DNA?
This is a matter of scale, though. There is a big difference between
- Someone wants to know if your DNA matches the killer's... so they get a hold of your DNA and test it
- The companies you interact with all have easy access to everyone's DNA and can make decisions based on things like who is the most likely to get addicted to gambling.
We've been fighting tooth and nail to prevent things like "people with hispanic surnames have a harder time getting a loan for a house". It's illegal to do so, and we still have problems with it. Now talking about adding in all kinds of "this person is more likely to have this condition" into those calculations. And they WILL be included if they're available; even if it's not obvious.
>- The companies you interact with all have easy access to everyone's DNA and can make decisions based on things like who is the most likely to get addicted to gambling.
Yes, that is exactly my point. In these hypothetical dystopias, this is going to apply to "everyone's DNA" and not just some subset of people. In order to fear volunteering your DNA, you need to fear a very specific level of dystopia in which this illegal DNA discriminating becomes common, but companies don't do anything illegal to acquire the DNA data.
It is like imagining that people could have escaped Nazi persecution if they just never admitted they were Jewish. Dystopias don't work like that. The evil people aren't going to give you a choice.
If we are going to live in a dystopia, it likely isn't going to be the Goldilocks dystopia in which just the right amount of evilness exists for this to be an issue.
Yeah, this. If an entity really wants to screen your DNA, there are easier routes than figuring out how to deanonymise you from a collection of thousands of profiles with basic demographic info they've just bought or paying enough to convince the entity with the very profitable lawful line in selling anonymised data to break the law for them.
No it isn't. I never said anything about making hiding DNA illegal, just that it is impractical. If you want to map it onto cryptography, I'm making the rubber-hose cryptanalysis argument[1][2]. People are worried about some evil entity going through proper channels for their evilness when there are more blunt and direct ways to get what they want.
> My DNA isn't some terrible secret I have to hide in the dark.
You DNA isn't as unique as you think it is, or it is, but specific parts aren't. The way law enforcement uses DNA tests only compares a small part of your DNA sequence. People have already been wrongly convicted based on DNA "evidence". If 23andMe is opening up access to their database of customers there's a very real chance for people to be misidentified and potentially implicated in crimes they didn't commit, again this has already happened using existing DNA databases.
So yes, you absolutely need to keep your DNA in the dark and only provide it when it's beneficial to you and when you can trust that it's kept safe or destroyed. Providing DNA to a company that might sell it in the future was always a stupid idea. 23andMe might be completely safe in what they are doing, but what stops them from selling your data to say Palantir in the future?
> I think protecting ones DNA is an uphill battle.
Sure, I'd agree with you on that, but that doesn't mean that you shouldn't try and just give up. Ideally it should matter and there is a lot of good and positive uses, but still little protection against misuse.
> there's a very real chance for people to be misidentified and potentially implicated in crimes they didn't commit, again this has already happened using existing DNA databases.
The solution there is more light, not more obscuration. You are right, the more samples you screen from the more false positives you will get. The solution is to corraborate the matches with other information, and being transparent about the levels of confidence.
Seems like growing pains for that methodology. More data can inform our confidence intervals.
Someday soon, it'll be as ubiquitous as fingerprints, facial recognition, iris scans, etc. It'll be up to our legislatures and law enforcement to keep pace (it does take decades, sadly).
I don't think pretending something isn't there has ever worked. Only when it's so open and transparent and undeniable does it force through change.
My DNA isn't some terrible secret I have to hide in the dark. I'm just one of like 8 billion variants of imperfectly copied ape. Maybe paying 23andme to do something useful with it will be my only lasting legacy in this world, lol.
It's not that I "didn't see it coming", it's that I hoped this would happen from the get-go.