Michael Hawley of MIT's Media Lab has died
nytimes.com
nytimes.com
"I really think the act of making and doing things is the salvation of our crumbling society. It has to be as someone commented over lunch -- creation is the opposite of destruction."
RIP
https://twitter.com/mjhawley/status/1173273979894796297?s=19
I don't think it's too unusual or bad for someone to have a brush with cancer and start getting philosophical about it in this way.
People either ignored it, or folks like Steve Job decided to rationalize it.
It doesn't mean it's a good thing.
Most of the transhumanists I've met don't bother justifying this belief beyond the selfish "I don't want to die." Personally I think it would lead to the ossification of culture and an astronomically enormous class divide (particularly between those who can afford immortality and those who cannot.) I'm not eager to live in a society where the ultra-billionaires like Steve Jobs can buy their way out of death, which thus-far, has been the great equalizer that even the most wealthy cannot bribe forever.
It isn't a selfish desire. Do you want your SO to die? Your parents? Think about anyone you ever care about.
Personally I think it would lead to the ossification of culture and an astronomically enormous class divide (particularly between those who can afford immortality and those who cannot.)
Based on what? This society isn't the only way to organize society or the only way we ever live, assuming it is the kind of society where we get deep and rapid ossification.
Sure, immortality have its own problems, but can you say that the new problems are so undesirable that we should not even attempt to solve it?
I'm not eager to live in a society where the ultra-billionaires like Steve Jobs can buy their way out of death, which thus-far, has been the great equalizer that even the most wealthy cannot bribe forever.
If that is the case, then it is a colossal failure for such a technology, since it only alleviated human suffering for a very handful of people.
Currently, our healthcare system is suffering from a massive demand on its capabilities due incredibly ill people being needed to be cared for.
Biomedical technology here, should be the great equalizer. Instead of being bankrupted by ultimately ineffective medical care, you now have more wealth to spend on things that's important to you, like living with your loved ones.
Colon cancer is most often avoidable. All one needs to do is go for a colonoscopy by age 45. Heck, why not just start when you are 25? (I had a great friend die at age 32 from colon cancer).
So you get a colonoscopy every 5 years.
Every year you get the colon cancer marker poop test. It looks for bits of RNA I believe. Or maybe it's PCR. Maybe those are the same. I dunno.
Colon cancer typically grows slowly and the precursors are easily seen in the colonoscopy & the colon cancer poop test does a good job too.
Also this: https://www.medicalnewstoday.com/articles/blood-test-can-det...
The answer to that is almost certainly that balancing the pros and cons of preventive screenings is a surprisingly complicated science. Screenings can cause harms by causing unnecessary treatment of patients with false positive diagnoses. If you increase screening of patient groups with a very low incidence of something then you overwhelmingly will get more false positives.
It requires careful evaluation of the scientific evidence to design screening programs that do what they should do: Improve patient outcomes and not cause more harm than good.
It has to be an individual decision. Give me the false positive rate and the harm of an unnecessary treatment and I'll weigh that against the risk of my premature death.
> "More harm than good" cannot be assessed objectively. How could you possibly weigh "one extra death" against "ten unnecessary treatments"?
Well, a very simple way would be to see how many deaths are caused by those ten unnecessary treatments. If those ten unnecessary treatments cause two deaths, I think the outcome is fairly clear.
More specifically, you wouldn't assess this in terms of "deaths" because 100% of people die. Instead, you'd study this in terms of units of time alive, with a probability distribution and a quality factor applied. If a treatment reduces pain and but causes blindness, pain and blindness both correspond to quality factors.
The quality factors, right now are generally treated as subjective and left to the patient as you suggest, but this is largely because we haven't gathered enough evidence for a general theory of quality of life based on health factors. Your claim that this "cannot be assessed objectively" is very much not in evidence--a great many scientists are assessing it objectively, looking at happiness, life satisfaction, satisfaction with care, societal effects, etc.
There is an entire, mature field of study built around objectively assessing the thing you are claiming cannot be objectively assessed. Hundreds of billions of dollars are invested in the idea that this can be objectively assessed. You might have some criticism of medical research, but throwing out a claim like "this cannot be assessed objectively" with no evidence when millions of people and billions of dollars are invested in the opposite, sounds an awful lot like you aren't aware that strong counterarguments to your point exist which need to be addressed if you make such a claim. This is further supported by you talking about "deaths" when in fact that's not really the metric people use for assessing interventions.
Source: worked a at a company that did medical study data warehousing.
In a more general sense, I would caution you that just because you can't assess something objectively doesn't mean it cannot be assessed objectively.
No, it isn't. Deaths are not fungible. My life is not someone else's life. If I get a diagnosis, there is always a chance that it is wrong, that the treatment will kill or injure me, and so on. That doesn't mean I shouldn't get a diagnosis. I'm the person to make the decision here, I can always choose not to get a diagnosis, or to ignore it.
> More specifically, you wouldn't assess this in terms of "deaths" because 100% of people die. Instead, you'd study this in terms of units of time alive, with a probability distribution and a quality factor applied. If a treatment reduces pain and but causes blindness, pain and blindness both correspond to quality factors.
Again, this can not be objectively assessed. You can not put a number on how bad (risk of) blindness versus pain is for an individual.
> There is an entire, mature field of study built around objectively assessing the thing you are claiming cannot be objectively assessed. Hundreds of billions of dollars are invested in the idea that this can be objectively assessed.
Just because something can, by simple reasoning, not be done objectively, doesn't mean it can not be done at all. Of course you can "objectively" add up a bunch of numbers and say that this represents the best course of action by some arbitrary metric. That doesn't mean it is the best course of action for the individual. See:
https://en.wikipedia.org/wiki/Mere_addition_paradox
https://en.wikipedia.org/wiki/Utility_monster
> In a more general sense, I would caution you that just because you can't assess something objectively doesn't mean it cannot be assessed objectively.
It should stand to reason that no one can assess something objectively that is inherently subjective, like my own value perception of my life and (risk) of death.
It sounds like you're imagining that looking at your individual case will necessarily break you out of the broader statistical pattern because you have special information, and you should act on that without regard for the broader pattern.
This is an error because it's not the case that your individual info. will lead you to be able to escape the false positive rates.
TL;DR: If you buck the rule and get screened earlier than recommended, you may end up diagnosing something earlier, but you also may end up hitting the false positive treatment which has its own consequences. The general rates capture these individual possibilities.
That may or may not be informed by additional information, for instance a person close to me may have had an early death from colon cancer and therefore I value the relief of having done the screening higher than someone else.
What makes you think that personal preferences can't be studied? I'll tell you they can be studied because they are being studied.
There is no "recommendation system" that can predict the right course of action based on my personal preference. There isn't even a good system to record and measure personal preference, because expression of perception is by itself highly subjective.
A doctor might ask you "how bad is the pain on a scale of one to ten?" or something of those sorts. That is informed by your entire history of pain experience. You're likely to rate the worst pain you ever felt as 10 - every time.
Then, pain does not equal suffering. A child might start crying and get upset about a finger prick. A trained martial artist might be able to emotionally ignore the sensation of most pain stimuli, even though it's felt just as strongly.
Could a system measuring and accounting for all this exist in principle? Perhaps. It doesn't exist though, so your point is moot.
Okay, referencing studies showing that people don't accurately report pain in pain assessments is the second time you've obliquely referenced well-known groups of studies (the first time being a roundabout reference to How Doctors Die[1]).
Think about how bad this logic is: you're referencing papers which gather objective evidence on personal perceptions around treatment, and using those papers as evidence that it's impossible to gather objective evidence on personal perceptions. Do you not see how the very evidence you're referencing refutes your point?
One fix being studied for pain assessments is to additionally add a question about what the worst pain the patient has ever suffered previously is, and compare that to their rating of the current pain.
So you have a statistic that says doctors choose to die differently than the patients they are treating. That is interesting, but merely descriptive, not prescriptive. It does not tell you which treatments works best for someone who happens to be a doctor.
If you are talking about fixes being studied for problems borne out of subjectivity, you admit that objectivity is at odds with the individual.
I really don't even know what we're talking about any more. What are "problems borne out of subjectivity"?
There is no pain meter that gives an objective reading of the patients experience and outlook and then tells you if some medication is statistically justified, relative to the harm it causes.
Yes, pain and suffering are subjective experiences, but they are also objective phenomena. We know, to some extent, how pain works, and can see it happening on brain scans.
I think the core of our disagreement here is that you seem to think that experience is somehow supernatural and therefore can't be studied, but it isn't: experience happens in the physical world, via physical and chemical processes, and can absolutely be studied via normal scientific means.
A lot of the things that indeed are possible (like brain imaging for pain response) are only feasible in a lab setting, they're not in widespread clinical use. Maybe they should be, maybe doctors shouldn't be handing out fentanyl like candy and instead look at what's going on in their patient's brain first - but then how do you prove that what's showing on the screen reflects the experience of the patient?
That's not to say that such studies are entirely useless. They're one more data point, but ultimately, the decision should be up to the individual. This is also closer to my experience of how patients are actually treated. There is no oracle system that can tell you what the right way to treat an individual is.
> I think the core of our disagreement here is that you seem to think that experience is somehow supernatural
Not at all. I don't think we disagree on what's possible "in theory". I'm not a dualist.
> No, it isn't. Deaths are not fungible. My life is not someone else's life. If I get a diagnosis, there is always a chance that it is wrong, that the treatment will kill or injure me, and so on.
Yes, and we can objectively measure what those chances are. Also don't forget, the process of diagnosis often also has consequences. Some percentage of people die on the table during surgical biopsies, for example.
> That doesn't mean I shouldn't get a diagnosis. I'm the person to make the decision here, I can always choose not to get a diagnosis, or to ignore it.
Sure, if you wish to ignore all objective evidence when making life-and-death decisions for yourself as a way of showing off just how gosh-darn free you are, that's your prerogative.
> It should stand to reason that no one can assess something objectively that is inherently subjective, like my own value perception of my life and (risk) of death.
I'm not sure how you've concluded that your perceptions cannot be objectively measured.
That's not being questioned. You're missing the point.
Let's say you have cancer: You statistically have a 10% chance of surviving if you do treatment, but it will make your remaining time worse, no matter what. What do you do? Delegate your decision to some abstract scoring system? Of course not. It's a personal decision, influenced by all the variables in your life.
As matter of fact, our system is set up to endorse more treatments in order to limit liabilities. How are you going to factor that one in? It has nothing to do with you and everything to do with the people treating you.
> Sure, if you wish to ignore all objective evidence when making life-and-death decisions for yourself as a way of showing off just how gosh-darn free you are, that's your prerogative.
Choosing to get a screening early is probably about as risky as riding a bicycle to work. What if the average person really likes riding a bike so much that it's worth the risk? Shouldn't you also absolutely ride a bike? Of course not. You should ride the bike if you like to.
> I'm not sure how you've concluded that your perceptions cannot be objectively measured.
Can my perception, which by definition is subjective, be measured objectively, by some technology, at least in theory? Perhaps.
Practically speaking, is it going to be measured in this context? Am I going to get a prediction on how I personally will feel when choosing one thing over another? No.
If by "some technology" you mean "your doctor asking you what your perception is", sure.
Yes, patients lie, which is why there's a whole field of study around when and why patients lie, and detecting that.
> Practically speaking, is it going to be measured in this context? Am I going to get a prediction on how I personally will feel when choosing one thing over another? No.
Yes. Do your really think you're the first person in hundreds of years of medicine to think, "Hey, maybe we should personalize treatment based on what patients want."?
This is a science that is still far from mature, and we're a long way from being able to make accurate predictions in every situation, but at this point we can definitely predict patient satisfaction with a high degree of certainty for common cases, based on what the patient says they value.