Comment by ErrantX | Hacker News Reader

Comment by ErrantX | Hacker News Reader

Parent Full thread

I understand where your ideas come from, I wish they were true. But sadly they are misinformed on many levels, and the root of the problem is you are playing a statistical game where reality will win.

- There are not enough Billionaires to meet your needs, statistically you'll need quite a few to find any decent cures. Especially with 30 teams in play on each case; figuring out "what worked" is a bad numbers game in that case.

- Science shows there is unlikely to be one "magic bullet" for cancer, which makes it hard to cure individuals and then take that science into a wider populace; "something here worked, but does it work generally?"

- Cancer treatments are harsh and invasive - by design! 30-40 competing drugs is unsustainable, you would be, to put it mildly, royally fucked. So death could come simply by exhausting the body with treatment, as easily as anything else.

- You invoked Godwin's law, which is never a good sign.

I agree with your general premise; which is that it takes too long to iterate ideas into actual production. But the reason that scenario exists is not just "quack medicine" and other hopeful (but useless) remedies, it is for the case of "oh shit, I just killed a million people because we didn't investigate long term effects".

And that is where your theory really fails; because you are trading an unknown collateral now, for an unknown collateral in the future. I'm not arguing against either as a good or bad thing; but it does highlight the fatal flaw in the argument "you are killing X people right now by not releasing/iterating fast enough"

I have some questions about your points... How many people do you think one would need to find any decent cures/treatments? How many patients are current researchers making use of to accomplish the feat? If moldbug had said something along the lines of "you can increase the numbers by having billionaires fund people other than themselves, such as their parents or other people they care about, or just as donations, in ways that don't bankrupt them but still provide lots of experimental data", do you think the increased number could still not possibly be enough, would you still object?

If moldbug had started with 5 concurrent treatments instead of 30, would you object to that? (Though considering death is coming I personally would rather die of treatment exhaustion from X treatments than the actual cancer, knowing I tried many potential solutions. (This conditional on a cancer that actually threatens my life.))

Lastly, are the "unknown collaterals" equivalent unknowns? Shouldn't we try to estimate them and decide whether ramping up the release/iterate process is worth it? That people are dying because the research system sucks doesn't seem that controversial to me or very unknown in truth-value. How likely do you think it is that someone kills 1 million people with some remedy, and how likely do you think it is that the same 1 million people would have lived much longer without such a remedy being attempted on them?

I am going to echo Jach comments a bit. Millions of people are dying from cancer, and will die at the pace of current research. The argument is that if you turn up the pace, you will kill more people but in the longer term you may save a lot more. Instead of decades from idea to human trial you push into a few years.

Millions of people are dying from cancer, and will die at the pace of current research. The argument is that if you turn up the pace, you will kill more people but in the longer term you may save a lot more.

You can't predict these numbers though; which is my point. Neither is a very useful argument.

There are a lot of improvements that could take place; the moving of more drug research from the US to other countries, for example (that's probably a controversial opinion in itself).

Statistically you'll need quite a few

This is exactly the problem: statistics. In a problem where every case is unique, any use of statistics is an abuse of statistics.

The roots of Western medicine are arguably pre-scientific, and certainly pre-statistical. We knew and could do a lot of things before we knew what a p-value was.

If what "advancing knowledge" means to you is "accumulating p-values so I can publish papers," this kind of experiment can produce no information whatsoever.

If what "advancing knowledge" means to you is "we think we know what's going on and we think we can do it better next time," the answer is very different. Of course this is what most people, indeed most industries, mean by "learning." We don't have to know absolutely that we know - we don't have that luxury. Instead we need to get things done.

Modern scientific medicine is supposed - no, required (with the increasingly narrow exception of surgery) - to learn only in the first way. Guess what? It doesn't seem to be learning much. Or at least, much of any use. Maybe medicine isn't as much like physics as we'd thought.

it is it is for the case of "oh shit, I just killed a million people because we didn't investigate long term effects".

And you're lecturing me about Godwin's law?

For what it's worth, long-term effects of OTC non-cancer drugs aren't investigated for shit. I'm permanently achlorhydric from permanently solving my GERD with daily omeprazole. As recommended by actual doctors. What the hell is the effect of taking omeprazole for a decade? Who the hell knows?

And that is where your theory really fails; because you are trading an unknown collateral now, for an unknown collateral in the future.

Did you ever read Asimov's Foundation? Remember the bits set on Trantor? Your whole mindset is straight out of Trantor, I fear - and it's certainly not exclusive to you.

For what it's worth, long-term effects of OTC non-cancer drugs aren't investigated for shit. I'm permanently achlorhydric from permanently solving my GERD with daily omeprazole. As recommended by actual doctors. What the hell is the effect of taking omeprazole for a decade? Who the hell knows?

Omeprazole has been about for nearly quarter of a century; and is fairly well understood - hell I can walk down to my local drug store and pick up a load from the shelves.

What you are describing, IMO, is either one of two things:

- a systemic failure in healthcare provision (not drug development)

- or a valid trade off between cure and side-effect

If what "advancing knowledge" means to you is "we think we know what's going on and we think we can do it better next time,"

The problem here is that you're going to kill a larger number of people in-trial due to mistakes or missteps. We know this because that is what early medicine did. Whether this is more than those who would die without aggressive research is something of a moot argument. Arguing the numbers is largely pointless because it's impossible to predict what will happen (that, after all, is the point).

Now; I see your argument that this is a valid tradeoff. Indeed, I think we should take more risks than we do. But not quite to the extent you are arguing.

Omeprazole has been about for nearly quarter of a century; and is fairly well understood - hell I can walk down to my local drug store and pick up a load from the shelves.

You say this like it follows.

I can too. I can read the label as well, which says to take it for no more than 28 days "without the advice of a doctor." Presumably this is the period for which it's been studied.

The idea of doing a double-blind study of the various systemic effects of a decade of induced achlorhydria is... absurd. How can you say it's understood? It's not understood, at all, either by my standards or yours.

So, without perfect information - I decide. I don't like the fact that all my ancestors for 100 million years had HCl in their guts, and I don't. I do like the fact that I'm not in the process of getting Barrett's esophagus and my abdomen doesn't hurt all the time.

Weighing risks and deciding without perfect information: fact of life. Trying to outlaw this process: like legislating that pi equals 3. Classic failure mode of the utopian 20th century.

Presumably this is the period for which it's been studied.

So, without perfect information - I decide.

The thing about the modern world is that you can make informed choices - information is only a Google away.

For example; my specific knowledge of that particular drug is not high. But this morning I found out:

- The drug is well researched as being tolerated in the short to medium term

- There is concern that long term use may increase the risk of intestinal or stomach cancers (based on mouse studies)

- So far the longest term test has been 6 years (somewhat longer than your 28 days theory).

In fact, 28 days is a standard recommended limit for drugs available over the counter. It is fairly hard to do damage to yourself, with those sorts of drugs, in that time - but after that you really need the advice of a doctor who understands all of the risks involved. Again, this detail is available online with a little searching.

There is a long list of other side effects, risks and theories that are studied - mostly based on 5 year studies from what I can tell - and the current view is that those risks are pretty low.

What you've proven, here, is that you've not made an informed choice, because you appear unaware of the actual long term concerns. You've not weighed risks; you've weighed one obvious risk that you understand (and I presume the doctor told you about). That is only the very tip of the iceberg.

- There is concern that long term use may increase the risk of intestinal or stomach cancers (based on mouse studies) - So far the longest term test has been 6 years (somewhat longer than your 28 days theory).

In fact, 28 days is a standard recommended limit for drugs available over the counter. It is fairly hard to do damage to yourself, with those sorts of drugs, in that time - but after that you really need the advice of a doctor who understands all of the risks involved.

I can't believe you just said "understands all the risks involved." On the basis of mouse studies and some completely uncontrolled, non-blind 6-year longitudinal study. Or meta-analysis. Or whatever.

Is this the best information we have or can get? Sure. Are the words understands all appropriate? No, they are not.

Aiming for this kind of certainty, and constantly, unconsciously boasting of it, is how postmodern medicine works itself into a box where it's completely unable to think. The information available in this case is shite. But the decision still needs to be made.

>Modern scientific medicine is supposed - no, required (with the increasingly narrow exception of surgery) - to learn only in the first way. Guess what? It doesn't seem to be learning much. Or at least, much of any use. Maybe medicine isn't as much like physics as we'd thought.

Hmmm: http://www.ined.fr/fichier/t_paragraphe/27734/paragraphe_img...

It seems we've doubled our life expectancy in the last 50 years. Medicine isn't working you say?

Look, I'm bitter as much as anyone else about the state of academics. I'm ex-academic-biology and have made no bones about it, here in this thread or elsewhere. But you are seriously characterizing the entire field.

Do you know why people are dieing from cancer? Because we've cured all the other diseases that killed them off at age 40.

Think about that for a minute.

Cancer is simply your body falling apart, going off the rails and destroying itself. It's doing that because, frankly, we were never intended to live 100 years. Thanks to modern medicine, we live well past the age that normal diseases killed our ancestors.

Our reward: battling cancer, instead of tuberculosis, pnemonia, cholera, gangrene or simply fever.

This point is so insightful and important, and I'm pointing it out in the hopes that more people will see it.

It seems we've doubled our life expectancy in the last 50 years. Medicine isn't working you say?

Mostly gains in infant mortality, plus antibiotics. We're coasting on the mid-20C golden age of medicine. Modern medicine is great - postmodern medicine sucks.

Replace "50" with 20 or even 30 and ask the same question.

Reply on news.ycombinator.com