I went to the White House for a Cancer Moonshot meeting as part of VP Biden’s Cancer Initiative about 12 years ago. About 30 of us sat around a big table for a day and at the end of it, we agreed to have an intern try to improve the API for retrieving clinical trial descriptions. Not that that did much good because pharma companies gain competitive advantage from keeping the important details of their trials secret. That was our moonshot.
It's not all bad news though.
There is unlikely to be a single cure for cancer because cancer is lots of different diseases but targeted therapy and immunotherapy has been close to a miracle for some cancers. Kidney cancer, for example, has gone from practically no treatments to 50% five-year survival over the space of about ten years. Maybe not quite a moon landing yet but they are maybe orbiting the moon. The government rules do their best to thwart the people shooting for the moon with forms filled out in triplicate though.
I have terminal brain cancer and there have been no new treatments for gliomas for about twenty years. There’s a new targeted therapy — vorasedinib — that was so successful that they called an early halt to its phase 3 trials to get it out to real patients in a rush job. It was given Fast Track Designation in February 2023 and should receive actual approval later this year. Fast Track indeed.
There are people shooting for the moon but they are blocked by people with clipboards. It’s all very sad.
> “I gotta laugh at that because otherwise I’d cry.”
I would just like to mention Michael Levin's research which, if his moonshot research pans out, would in my opinion have a high chance of being a general method to cure cancer.
Great article. Another in the line of scientists with a great hunch who were howled down by the experts is Ignaz Semmelweis. When it was finally adopted, his radical notion dropped maternal mortality rates from 18% to less than 2% in Vienna General Hospital in 1847. The revolutionary idea which so upset the status quo? Clinicians should disinfect their hands before moving between patients.
And I think the reality was even wilder: clinicians should wash their hands after they’ve had them in dead bodies in the morgue before they work on the maternity ward. And the clinicians were like “you’re mental, Ignatz, get out of my way”
> Meta-analyses come out all the time that are like, “Unfortunately, all of the studies that try to answer this important question are so bad that we can’t learn anything from them.” Instead of running another 500 piddling, confounded little studies, someone with guts, taste, and money could run the actually good study.
The National Institute of Health and Care Excellence (NICE) is a UK body that determines clinical guidelines. When they publish their clinical guidelines, they also publish a list of research questions they would like to see prioritized. These questions are tightly focused on clinical relevance, they are well-defined and they are currently unanswered in the clinical literature.
Every few years NICE reviews the clinical literature, checks if their recommended studies have been run and, normally, move on disappointed.
It's not hard to design studies that answer NICE's research priorities. These studies wouldn't be unusually difficult to do. But somehow the studies which NICE claim would affect the clinical care they recommend UK hospitals provide do not get ran.
I relate to the boobytrap nature of academia. At one time I was young and full of ideas and experiments and thought academia would allow me to breakthrough, research and do some cool stuff. But it was as you describe here and I wanted no part of it.
If you read packing for mars by mary roach then you will find that much of the research for going to the moon was little things. How to get human poo to go out of a butt and into a bag without floating littlw bits everywhere. If humans will go crazy in a little tin can. Which screw is better. How to tile the reentry thing. Only some of the research in the moonshot is big, showy and obviousely important to lay people.
Cancer is a human cell gone rogue. Killing cancer cells without killing normal cells that the cancer cells is doing its damndest to imitate is tough. There are therapies that genetically engineer t cells to attack cancer cells. These therapies required a lot of background research that may not be obviously linked.
In reading this, I thought of the MacArthur Genius Grants and the Princeton Institute for Advanced Studies. Both seem to be in the ballpark, but I don't know if either produced anything meaningful.
I think both are right in their intentions, but both of them wait for people to become mainstream successful first. I'd love to see people do the same thing, but enabling success rather than rewarding it.
Waterpik versus flossing, please! I hate the feeling of food between my teeth but floss always felt to me like an assault on my gums. I switched to Waterpik maybe a decade ago and my gum health improved. Which is remarkable because late middle age isn't usually a time when gums tend to get better.
Also! When toothpaste was invented it was common knowledge that bacteria were bad - so pretty much all toothpaste includes harsh microbicidal detergents. Nowadays it's common knowledge that not all bacteria are bad, and yet we're still merrily nuking our oral microbiomes wholesale twice a day. My dentist once tried to convince me to scrub all the bacteria off my tongue. Is this really such a good idea? I'd sure like to have some detergent-free toothpaste data to look at. Or even just some detergent free toothpaste I could buy to try the experiment on myself at n=1 .
Great examples. Another related one (from 1 persons experience). Using an electric toothbrush. I switched many years ago and have not had a cavity since. (Side note, I actually use toothpastes without flouride and do not have issues - but do floss and my gums are the only thing the dentist comments about - maybe I need to switch to waterpik!!!)
You're reminding me that I switched to Waterpik around the same time I switched to electric toothbrush - so who knows which thing might have led to seemingly healthier gums.
My insurance on the fluoride front is to use a hydroxyapatite toothpaste in the morning and a fluoride toothpaste before bed. It sure would be nice to have actual data to go on, as opposed to the flimsy best-guess insurance betting we're forced to do here.
People inspired to reach for moon shots have to be prepared to be completely wrong and be ok with that. I’m trying for one and console my doubts with the thought that at best it will be a valuable advance, and at worst save someone else wasting time, energy, and precious funding going down the same rabbit hole. Win/win for humanity, not so much for the ego if it turns out to be a dud.
One comment: If you are looking for ideas that get repressed by scientists but sold individually as cures by people outside the usual circles - The whole naturopathy field fits this description. Walk through a Health supplements store... Go through the MLM marketing claims. Take a bunch of these and do some of the actual testing (like your sunscreen example).
I love this post and also it triggered a bunch of connections inside my head bones.
1. Crackpots by Acollierastro - there are LOADS of crackpots with ideas they really believe in but that are incoherent. Too many to review, let alone fund: https://youtu.be/11lPhMSulSU?si=O1T4C8SH-OMwZRTv
2. But maybe with a more distributed way of allocating funding, you could give 10,000 crackpots a little money and double down on the handful that get somewhere. Dave Snowden’s Cynefin Co is working on distributed structures for micro-funding and micro-lending using the panopticon effect to reduce fraud.
3. And Ken Stanley has a strong critique of the linear, objective-based approaches to research and innovation. In short: you should only do a moonshot when you have all the pieces basically figured out. Otherwise, research efforts should optimise for curiosity and novelty - the opposite of setting an instrumental goal. (Myth of the Objective)
I read this to think the ‘60s was the right moment for the literal moonshot. If they’d tried the same moonshot in the ‘50s or earlier, not only would it have failed, it would have also blocked some of the weird stepping stones that needed to be in place to make it possible to moonshot in the ‘60s.
Just like computers needed vacuum tubes to be invented, but you would never invent vacuum tubes if you were aiming at building a computer before they existed.
4. But surely some of the challenge is that humans tend to vote against tax money going to people who “don’t deserve it” - like the 999/1,000 crackpots, but also like the 1/1,000 who discovers a novel stepping stone that will be useful in 20 years but looks pointless today.
5. Even mainstream-ish people can be seen as crackpots. As I understand it, for example, David Deutsch has an explanation of quantum field theory that’s been seen as a crackpot outlier (I.e. there really is a multiverse) even though he used his explanation to make recognised contributions to quantum computing.
6. But of course some Nobel prize winning scientists go on to become wild conspiracy theorists. Our human ability with abductive reasoning enables both leaps of insight and leaps to conspiracy theories.
7. And yeah: string theory. Which seems to have lost ground at last, but did generate some cool maths along the way?
8. Which brings us full circle to crackpots, who have theories they *feel* are huge insights but which also tend to be unfalsifiable like conspiracy theories.
I have a PhD in molecular biology, but not a cancer person, but pro NIH bias
I think this blog post is a load of horse puckey.
if all you have is a hammer (writing blog posts that are critical of this or that) then everything looks like a nail (something bad that you can fulminate against)
The one very very very painful lesson we have learned, or should have learned in the last 50 years is that cancer is one effing fiendishly clever foe, and progress is slow, but
even a little bit of progress = people saying good morning to their loved ones, rather then their loved ones saying Kaddish
somewhere Erwin Chargaff remarks about DNA sequencing (in the Sanger dideoxy era) that it was bogus cause no one did wet chemistry to get an analytical measure that the pyrimidine / purine ratio of the sample was close to one !! <sarcasm/>
Your point regarding cancer is valid, but a more apt response to the article would be: "Cancer by its nature may not be amenable to a Moonshot initiative, because it's not just one thing." The article supports the value of incremental progress, so your denunciation is unjust.
But scientific honesty demands that if we call an initiative a "Moonshot," it is deserving of the name and not merely a reinforcement of the status quo with more research dollars.
Well, of course, in the moonshot realm Elon is building rockets that would never make it through a NASA design review to go to Mars. He is a huge risk taker, but has made some notable moonshot like advances. The key, in my mind, is no government review and approval of his ideas. Yes they funded some of his work, by buying specific services. Nothing cost plus. Government and academia are the most risk adverse entities on the planet.
This reminds me of the ‘Chaos in the Brickyard’ letter to Science by BK Forscher in 1963 - the money keeps going to making more bricks, but now we’re buried in them. How do we get back to actually building useful things?
Thanks for that link. In the age of machine learning (ML) and AI we have a problem which on the face of it is the opposite of the one Forscher describes: rather than too much raw data ("bricks"), we have not enough, especially in the biological sciences.
On a regular basis, we see delusional articles about how AI is going to cure cancer, diabetes, etc. Such articles neglect the fact that new data from well-designed experiments are essential for progress in these areas. And such data are expensive and time-consuming to acquire. No ML/AI model trained on pre-existing data can get around this fact.
ML models are data hungry. If such models are to lead to substantial medical advances, perhaps what we really need are clever scientists and engineers to come up with cheaper ways to generate the volumes of relevant data that such models require.
Meanwhile, I'm out here trying to convince cynics that they are naive, trying to dethrone Occam's Razor, reduce all heuristics and biases to a synthesis of 5 each (which are orthogonal to one another, and might spoil all the fun debate over whether something is a heuristic or a bias... [hint: if you can't tell the difference between a tool and a systematic error of that tool, then you probably have found neither]), and experimenting replacing "natural selection" as a "mechanism," with something that doesn't change the core theory, and doesn't get arbitrarily stuck personifying genes, reifying selection pressures, and gamifying complexity.
Sounds mad, I know. But don't worry. No one is mad enough to fund me, and/or I'm not mad enough to ask.
There is no room for serendipity in modern academia, and that is where most of the big breakthroughs came in the past. So the big question is- how do you go about creating any of the solutions you have proposed? Or is this a belling the cat kind of thought experiment?
I went to the White House for a Cancer Moonshot meeting as part of VP Biden’s Cancer Initiative about 12 years ago. About 30 of us sat around a big table for a day and at the end of it, we agreed to have an intern try to improve the API for retrieving clinical trial descriptions. Not that that did much good because pharma companies gain competitive advantage from keeping the important details of their trials secret. That was our moonshot.
It's not all bad news though.
There is unlikely to be a single cure for cancer because cancer is lots of different diseases but targeted therapy and immunotherapy has been close to a miracle for some cancers. Kidney cancer, for example, has gone from practically no treatments to 50% five-year survival over the space of about ten years. Maybe not quite a moon landing yet but they are maybe orbiting the moon. The government rules do their best to thwart the people shooting for the moon with forms filled out in triplicate though.
I have terminal brain cancer and there have been no new treatments for gliomas for about twenty years. There’s a new targeted therapy — vorasedinib — that was so successful that they called an early halt to its phase 3 trials to get it out to real patients in a rush job. It was given Fast Track Designation in February 2023 and should receive actual approval later this year. Fast Track indeed.
There are people shooting for the moon but they are blocked by people with clipboards. It’s all very sad.
> “I gotta laugh at that because otherwise I’d cry.”
I should have read your comment and liked it before making my own
I would just like to mention Michael Levin's research which, if his moonshot research pans out, would in my opinion have a high chance of being a general method to cure cancer.
Great article. Another in the line of scientists with a great hunch who were howled down by the experts is Ignaz Semmelweis. When it was finally adopted, his radical notion dropped maternal mortality rates from 18% to less than 2% in Vienna General Hospital in 1847. The revolutionary idea which so upset the status quo? Clinicians should disinfect their hands before moving between patients.
Poor Ignatz!
And I think the reality was even wilder: clinicians should wash their hands after they’ve had them in dead bodies in the morgue before they work on the maternity ward. And the clinicians were like “you’re mental, Ignatz, get out of my way”
> Meta-analyses come out all the time that are like, “Unfortunately, all of the studies that try to answer this important question are so bad that we can’t learn anything from them.” Instead of running another 500 piddling, confounded little studies, someone with guts, taste, and money could run the actually good study.
The National Institute of Health and Care Excellence (NICE) is a UK body that determines clinical guidelines. When they publish their clinical guidelines, they also publish a list of research questions they would like to see prioritized. These questions are tightly focused on clinical relevance, they are well-defined and they are currently unanswered in the clinical literature.
Every few years NICE reviews the clinical literature, checks if their recommended studies have been run and, normally, move on disappointed.
It's not hard to design studies that answer NICE's research priorities. These studies wouldn't be unusually difficult to do. But somehow the studies which NICE claim would affect the clinical care they recommend UK hospitals provide do not get ran.
Wow this is very interesting. If I was a billionaire I’d fund these studies.
(Speaking as an engineer and a father of a molecular biologist who went on to get a phd in Econ)
Cancer isn't something that is "cured." It isn't one thing. It isn't several things.
People talk about "curing cancer" to get money. And that works.
If you doubt it, ask yourself: would you rather be diagnosed with cancer today, or 30 years ago? Or even 5 years ago.
Progress is incremental and easy to ridicule. But the ridiculers aren't the one extending lifespans.
I relate to the boobytrap nature of academia. At one time I was young and full of ideas and experiments and thought academia would allow me to breakthrough, research and do some cool stuff. But it was as you describe here and I wanted no part of it.
If you read packing for mars by mary roach then you will find that much of the research for going to the moon was little things. How to get human poo to go out of a butt and into a bag without floating littlw bits everywhere. If humans will go crazy in a little tin can. Which screw is better. How to tile the reentry thing. Only some of the research in the moonshot is big, showy and obviousely important to lay people.
Cancer is a human cell gone rogue. Killing cancer cells without killing normal cells that the cancer cells is doing its damndest to imitate is tough. There are therapies that genetically engineer t cells to attack cancer cells. These therapies required a lot of background research that may not be obviously linked.
In reading this, I thought of the MacArthur Genius Grants and the Princeton Institute for Advanced Studies. Both seem to be in the ballpark, but I don't know if either produced anything meaningful.
I think both are right in their intentions, but both of them wait for people to become mainstream successful first. I'd love to see people do the same thing, but enabling success rather than rewarding it.
Waterpik versus flossing, please! I hate the feeling of food between my teeth but floss always felt to me like an assault on my gums. I switched to Waterpik maybe a decade ago and my gum health improved. Which is remarkable because late middle age isn't usually a time when gums tend to get better.
Also! When toothpaste was invented it was common knowledge that bacteria were bad - so pretty much all toothpaste includes harsh microbicidal detergents. Nowadays it's common knowledge that not all bacteria are bad, and yet we're still merrily nuking our oral microbiomes wholesale twice a day. My dentist once tried to convince me to scrub all the bacteria off my tongue. Is this really such a good idea? I'd sure like to have some detergent-free toothpaste data to look at. Or even just some detergent free toothpaste I could buy to try the experiment on myself at n=1 .
Great examples. Another related one (from 1 persons experience). Using an electric toothbrush. I switched many years ago and have not had a cavity since. (Side note, I actually use toothpastes without flouride and do not have issues - but do floss and my gums are the only thing the dentist comments about - maybe I need to switch to waterpik!!!)
You're reminding me that I switched to Waterpik around the same time I switched to electric toothbrush - so who knows which thing might have led to seemingly healthier gums.
My insurance on the fluoride front is to use a hydroxyapatite toothpaste in the morning and a fluoride toothpaste before bed. It sure would be nice to have actual data to go on, as opposed to the flimsy best-guess insurance betting we're forced to do here.
Consumer Reports for medicines! Now!
https://cbuck.substack.com/p/consumer-reports-for-medicines
People inspired to reach for moon shots have to be prepared to be completely wrong and be ok with that. I’m trying for one and console my doubts with the thought that at best it will be a valuable advance, and at worst save someone else wasting time, energy, and precious funding going down the same rabbit hole. Win/win for humanity, not so much for the ego if it turns out to be a dud.
Great Article!
One comment: If you are looking for ideas that get repressed by scientists but sold individually as cures by people outside the usual circles - The whole naturopathy field fits this description. Walk through a Health supplements store... Go through the MLM marketing claims. Take a bunch of these and do some of the actual testing (like your sunscreen example).
I love this post and also it triggered a bunch of connections inside my head bones.
1. Crackpots by Acollierastro - there are LOADS of crackpots with ideas they really believe in but that are incoherent. Too many to review, let alone fund: https://youtu.be/11lPhMSulSU?si=O1T4C8SH-OMwZRTv
2. But maybe with a more distributed way of allocating funding, you could give 10,000 crackpots a little money and double down on the handful that get somewhere. Dave Snowden’s Cynefin Co is working on distributed structures for micro-funding and micro-lending using the panopticon effect to reduce fraud.
3. And Ken Stanley has a strong critique of the linear, objective-based approaches to research and innovation. In short: you should only do a moonshot when you have all the pieces basically figured out. Otherwise, research efforts should optimise for curiosity and novelty - the opposite of setting an instrumental goal. (Myth of the Objective)
I read this to think the ‘60s was the right moment for the literal moonshot. If they’d tried the same moonshot in the ‘50s or earlier, not only would it have failed, it would have also blocked some of the weird stepping stones that needed to be in place to make it possible to moonshot in the ‘60s.
Just like computers needed vacuum tubes to be invented, but you would never invent vacuum tubes if you were aiming at building a computer before they existed.
4. But surely some of the challenge is that humans tend to vote against tax money going to people who “don’t deserve it” - like the 999/1,000 crackpots, but also like the 1/1,000 who discovers a novel stepping stone that will be useful in 20 years but looks pointless today.
5. Even mainstream-ish people can be seen as crackpots. As I understand it, for example, David Deutsch has an explanation of quantum field theory that’s been seen as a crackpot outlier (I.e. there really is a multiverse) even though he used his explanation to make recognised contributions to quantum computing.
6. But of course some Nobel prize winning scientists go on to become wild conspiracy theorists. Our human ability with abductive reasoning enables both leaps of insight and leaps to conspiracy theories.
7. And yeah: string theory. Which seems to have lost ground at last, but did generate some cool maths along the way?
8. Which brings us full circle to crackpots, who have theories they *feel* are huge insights but which also tend to be unfalsifiable like conspiracy theories.
Fun!
I have a PhD in molecular biology, but not a cancer person, but pro NIH bias
I think this blog post is a load of horse puckey.
if all you have is a hammer (writing blog posts that are critical of this or that) then everything looks like a nail (something bad that you can fulminate against)
The one very very very painful lesson we have learned, or should have learned in the last 50 years is that cancer is one effing fiendishly clever foe, and progress is slow, but
even a little bit of progress = people saying good morning to their loved ones, rather then their loved ones saying Kaddish
https://www.cancer.gov/research/key-initiatives/moonshot-cancer-initiative/progress
Sad but not surprising how far standards for PhD's in molecular biology have slipped these days.
yeah, right
somewhere Erwin Chargaff remarks about DNA sequencing (in the Sanger dideoxy era) that it was bogus cause no one did wet chemistry to get an analytical measure that the pyrimidine / purine ratio of the sample was close to one !! <sarcasm/>
C’mon, do better.
Woke boiler plate. Really?
Your point regarding cancer is valid, but a more apt response to the article would be: "Cancer by its nature may not be amenable to a Moonshot initiative, because it's not just one thing." The article supports the value of incremental progress, so your denunciation is unjust.
But scientific honesty demands that if we call an initiative a "Moonshot," it is deserving of the name and not merely a reinforcement of the status quo with more research dollars.
Well, of course, in the moonshot realm Elon is building rockets that would never make it through a NASA design review to go to Mars. He is a huge risk taker, but has made some notable moonshot like advances. The key, in my mind, is no government review and approval of his ideas. Yes they funded some of his work, by buying specific services. Nothing cost plus. Government and academia are the most risk adverse entities on the planet.
This reminds me of the ‘Chaos in the Brickyard’ letter to Science by BK Forscher in 1963 - the money keeps going to making more bricks, but now we’re buried in them. How do we get back to actually building useful things?
It looks paywalled (sigh), but it’s just the one page so it actually isn’t (for once): https://www.science.org/doi/10.1126/science.142.3590.339.a
And he thought it was bad 60 years ago….
Matteo Farinella made a beautiful comic adaptation of BK Forsher's letter:
https://massivesci.com/articles/chaos-in-the-brickyard-comic-matteo-farinella/
Thank you, I haven’t come across that before!
Thanks for that link. In the age of machine learning (ML) and AI we have a problem which on the face of it is the opposite of the one Forscher describes: rather than too much raw data ("bricks"), we have not enough, especially in the biological sciences.
On a regular basis, we see delusional articles about how AI is going to cure cancer, diabetes, etc. Such articles neglect the fact that new data from well-designed experiments are essential for progress in these areas. And such data are expensive and time-consuming to acquire. No ML/AI model trained on pre-existing data can get around this fact.
ML models are data hungry. If such models are to lead to substantial medical advances, perhaps what we really need are clever scientists and engineers to come up with cheaper ways to generate the volumes of relevant data that such models require.
Meanwhile, I'm out here trying to convince cynics that they are naive, trying to dethrone Occam's Razor, reduce all heuristics and biases to a synthesis of 5 each (which are orthogonal to one another, and might spoil all the fun debate over whether something is a heuristic or a bias... [hint: if you can't tell the difference between a tool and a systematic error of that tool, then you probably have found neither]), and experimenting replacing "natural selection" as a "mechanism," with something that doesn't change the core theory, and doesn't get arbitrarily stuck personifying genes, reifying selection pressures, and gamifying complexity.
Sounds mad, I know. But don't worry. No one is mad enough to fund me, and/or I'm not mad enough to ask.
Life in the hard mode lane 🛝🚆
There is no room for serendipity in modern academia, and that is where most of the big breakthroughs came in the past. So the big question is- how do you go about creating any of the solutions you have proposed? Or is this a belling the cat kind of thought experiment?