From a practitioners perspective, I view this type of behavior as tuning the model more towards the exploitation side of the exploration/exploitation trade-off. I think a lot of recommendation engines do this (looking at you YouTube) because it’s more profitable.
Which at its core is probably an alignment problem in the way the models are evaluated: they are measured on their short-term effects, and there exploitation rules. But if you look at the long-term effect of recommendations you really need a healthy dose of exploration to keep your users around.
My wife and I visited Paris last summer and biked everywhere. It felt very safe, not as safe as Copenhagen but still safe. I hope that with more time, as more drivers get used to bikes on the road it will get safer and more people will bike, creating a positive feedback loop.
I'm a strong believer in road culture change happening. Biking in DC changed enormously when our bike share program showed up in 2008. Seeing casual chill people on slow bikes, often a little lost or uneasy, forces a bit of a reset from the classic fast cars vs spandex wheel-stander bikes dynamic.
It makes sense with some extra punctuation: Great. Socialize healthcare. As it, is blah blah.
If you want even cheaper car insurance, you can even go beyond socialized healthcare. You can have socialized car insurance! Or in some cases a hybrid system where you insure your vehicle but not all trauma and damage to the other party. But I don't think Americans would be a-okay with seeing a government employed doctor when they make their injury claims. Even if outcomes are better than a purely private system.
If a large portion of the cost of auto insurance is to pay for potential injuries to someone that you hit, then we can conclude 1) single-payer healthcare would significantly lower the cost of auto insurance, and conversely, 2) the lack of single-payer healthcare is a significant contributor to the current state of auto insurance markets/pricing.
I completely agree with your assessment and logic, but the parent posted suggested that socialization of healthcare (e.g. ObamaCare) is contributing to auto rates. That I don't understand.
Looking again, I think their comment can be parsed in multiple ways.
> Socialize healthcare as it is forcing all sorts of market distortions in unrelated markets
I interpreted that as a call to socialize healthcare, not a description of market effects that "socialized healthcare" has to the extent that exists in the US.
My suspicion has been that SNP array data is not that useful for drug discovery. They’re targeting the most common/variable SNPs, which I suspect don’t have a large health effect (except for maybe late in life, otherwise how would they get passed down). I would suspect the more valuable targets would be rarer, or arise de novo (as is common in cancer, eg driver mutations).
The effect size of common SNPs is not informative about the effect of drugging their related genes. For example, the common variants near HMG-CoA reductase have very small but significant (confidently nonzero) effects. Yet drugging HMG-CoA reductase can reduce LDL cholesterol by ~40-50% (statins).
deCODE Genetics, whose history is very interesting and worth reading [1], was bought by Amgen based on this premise.
Note, however, that SNPs like the one you pointed out are relatively infrequent. Amgen was expecting a two digit % improvement in their pharma pipeline by using GWAS insights.
> Note, however, that SNPs like the one you pointed out are relatively infrequent.
If you mean that SNPs with small effect sizes don't always point to useful drug targets with big druggable effects, that is possible, but this remains an open question and is the subject of intensive research right now.
Yes, I agree. The trick is probably to find cell-specific SNPs located in regulatory regions so that there are no off-target effects. Massive screens using single-cell perturbations will help to gain some insights.
They could help in stratifying patients and strategically targeted therapy. They can't be easily targeted (drugged) drirectly but they could lead to therapeutics if I understand correctly.
You’re right that rarer SNPs have big effect sizes. However the problem is that they’re rare so you don’t make much money treating ppl with rare diseases.
> supports that the CTO was a better judge of the candidate.
No, it doesn't, because the hiring manager's choice never got a chance to show what they could do. Besides, the fact that the nepotistic hire worked out could have been just dumb luck. After all, hiring is a crapshoot, especially hiring interns.
Regardless, this was clearly nepotism, and the question isn't whether the CTO could judge the candidate, the question is whether favoritism was shown toward a family friend, which is indisputably the case.
That is a factor in why nepotism is so endemic - people are much better judges of the character of people in their family. This is a clear-cut case of nepotism, although personally I don't see a problem here. Nepotism isn't a bad thing in small doses. This instance is a good example of why not.
