> it's more of a "ughhhh my stomach is stealing all my energy, better sit/lay down and let it do its thing" feeling. Is that what you mean by "fatigue?"
I think so, yes, you've just taken in a massive load of carbs, your body is trying to switch over to burning them (the Randle Cycle), but it's not working properly, so the weird thing where you've just taken on board a lot of easily accessible high-energy fuel but you have less energy as a result happens.
If you feed dogs shortbread they go mental with excess energy (I have one sample here, the law may not be universal)
People always used to complain that children went hyper after eating sweets.
A long time ago I remember cycling half-way to Oxford and back, fast, and on the way back I just completely ran out of energy. Not out of breath or anything, just my legs wouldn't move very fast and I was crawling along. And we found a garage that sold little glucose-enriched chocolate bars. And every time I ate one, I would suddenly speed up for about fifteen minutes and then slow down again. It was pretty comical! Like they were power pills.
We moderns seem to get tired after taking in carbs. I don't think that's normal, or at least I don't think it's a sign of a system working properly.
You keep saying "one of the things I believe must be fiction." I don't really get it. I think it's fair to see that "we don't completely understand what's going on" and there's enough uncertainty in there that nothing you strongly believe is necessarily false?
On the recent fatigue thing: when I swamp hard on refeeds, I do get sort of a feeling that I suppose could be called "tired" or "fatigue." I don't usually get "fatigued" so I don't know what exactly you're referring to. That swampy feeling to me is less "tired" in the yawning & sleeping sense like lack of sleep, it's more of a "ughhhh my stomach is stealing all my energy, better sit/lay down and let it do its thing" feeling. Is that what you mean by "fatigue?"
> On the recent fatigue thing: when I swamp hard on refeeds, I do get sort of a feeling that I suppose could be called "tired" or "fatigue." I don't usually get "fatigued" so I don't know what exactly you're referring to. That swampy feeling to me is less "tired" in the yawning & sleeping sense like lack of sleep, it's more of a "ughhhh my stomach is stealing all my energy, better sit/lay down and let it do its thing" feeling. Is that what you mean by "fatigue?"
It is hard, I think, for a person in good health to even notice the difference between fatigue and sleepiness. I remember years ago having my doctor explain this difference to me, and it only really made sense because I used to be an endurance athlete, and was familiar with the idea that you can be utterly out of energy after hours of hard work, even though if you hadn't done a bucket of exercise you wouldn't be even thinking about lying down.
Sleepiness is the thing where your brain realizes it has been awake for too long, and there is less light around, and it needs to switch into its overnight state and do whatever mysterious maintenance it does when we sleep.
Fatigue is when you run out of energy, and it usually happens only when you've used up most of your reserves and damaged yourself and your whole system including your brain just wants to do nothing until the situation improves.
One of the more interesting things about ketosis for me was that I seem to be able to get brain and body fatigue independently. I rarely see body fatigue in deep ketosis (I'm not an athlete any more, and rarely operate anywhere near the limit), but if I over-exert myself mentally I can get this thing where I've got loads of physical energy but my thought processes just aren't working properly. All my automatic skills are still there but I can't think.
I suspect I can run low on ketones quite easily, and my brain can't burn fat like the rest of me can.
I think this is all 'glycolysis is blocked/probably by PUFAs'.
I have noticed that I am confused by the fact that other people seem to have what I think are symptoms of PUFA poisoning, but don't complain about fatigue much.
If you can get fatigued while swamping, but not in deep ketosis, then I am less confused than I was!
> You keep saying "one of the things I believe must be fiction." I don't really get it. I think it's fair to see that "we don't completely understand what's going on" and there's enough uncertainty in there that nothing you strongly believe is necessarily false?
It's a habit I'm trying to develop: If your predictions fail or the world surprises you in some way, you should remind yourself that reality is lawful, and that confusion exists only in the mind, and must be explained by some imperfection in your model of the world.
That's how mathematicians think, and physicists, and I think it would probably be a helpful attitude for other people trying to figure things out.
Whenever I hear medical people say "not completely understood", I hear "we have no fucking clue, but we are going to carry on believing what we already believe even though we keep being surprised by things".
Better to say: "I *notice* that I am confused. At least one of the things I believe must be fiction". And then you should review all the assumptions that went into the prediction you have just seen go wrong, and work out which ones are dodgy.
It's an impossible standard and I'm crap at it because I am some hopeless human thing with a jerry-rigged mind that is fundamentally incapable of reason. But I can at least remind myself what my ideal actually is, and try to notice when I need to think.
I think cosplaying as mathematicians and physicians is how modern "science" got into this mess. It seems to not work well outside of those domains.
I don't make predictions much for my experiments, because it's just not that stage of sciencing. I suppose I could write "I predict to lose 10lbs" before every experiment, and most would fail. But what would that get me?
Largely we're still at the "throwing poop at the wall" stage of science, and clearly we don't understand enough to make acute, specific predictions like "lose 10lbs." We can't control our variables enough, or maybe we don't know which the real causal variables are, because even when I attempt to do the exact same thing, the results are often quite varied. Trends/patterns are just about the best we have, so far.
I agree that admitting confusion is good. But "at least one of the things I believe must be fiction?" I think I've repeatedly commented how I disagree that your results mean you believe a fictitious thing; they merely mean you don't understand the system perfectly yet.
We believe that water flows downhill, and that you can make hills and dams and trenches. If we couldn't perfectly predict the flow of any given river when it floods, that doesn't mean we believed a thing that was fiction.
> I don't make predictions much for my experiments, because it's just not that stage of sciencing. I suppose I could write "I predict to lose 10lbs" before every experiment, and most would fail. But what would that get me?
You have models, and you make predictions. All the time! And you act on your predictions! It is *always* that stage of sciencing.
For instance "PUFAs are probably bad, if I eat them bad things will happen", or "on ex150 I tend to end up weighing about 210lbs", or "if I weigh more than 210lbs doing ex150 for a month will bring my weight back towards 210lbs".
If you did ex150 for a month and ended up 50lbs heavier you would be very surprised!
Surprise is the feeling where you have made predictions and they have failed.
But I think a lot of your predictions and models are subconscious, and result in feelings about what will happen. That's fine, that's how we work. My subconscious is much more powerful than my explicit reasoning and I tend to trust it more and rely on it entirely where I am not very curious or very motivated to get better at things.
But by trying to bring those subconscious models into the light and think about them, we can sometimes spot where there are incoherencies and errors, and by setting up a dialogue between conscious and subconscious reasoning and allowing them to talk out their disagreements we can sometimes persuade both processes to be more accurate.
It is not even much of an analogy, but I try to think about the difference between writing a program to tell me what will happen, explicitly made out of formal rules and calculations, and training a neural net.
You should be doing both! Neural nets are great, and often do better. But formal rules are also great, for different reasons. And often the best approach is to do both, and use both programs to improve each other.
But I mostly don't think "something I believe must be wrong" only "I must be missing some details." Because it's a different stage of sciencing. I didn't abandon PUFA or low protein the first time it didn't work as the only required & sufficient factor.
I think you're missing my point. I think you can do science very wrong, and I think the way you phrase it is basically wrong for the level of confidence you should have given your information and where we are "as a field."
It's just not that phase, mainstream science is way too rigorous (in a bad way) and inflexible and has actually long abandoned the scientific method, see also Feyerabend.
The reason we make so much more progress is that we do NOT follow this orthodoxy and dogma, but perform various different methods when appropriate.
> I think you're missing my point. I think you can do science very wrong, and I think the way you phrase it is basically wrong for the level of confidence you should have given your information and where we are "as a field."
I don't have very much confidence at all. But often I think that a thing is more likely than another thing. When the second thing happens I try to *notice*. That's all.
> But I mostly don't think "something I believe must be wrong" only "I must be missing some details."
These things are perhaps not too different?
'Heavenly objects move in circles with epicycles' is quite hard to tell apart from 'Heavenly objects move in ellipses' if you're only looking at the planets.
I think almost everyone would have agreed that Ptolomy was only missing some details. In fact I think his theory was more accurate.
> The reason we make so much more progress is that we do NOT follow this orthodoxy and dogma
Well it is my orthodoxy and my dogma, and I follow it. It is not a widespread orthodoxy or dogma. I don't know many fellow believers! Although I have hopes to spread it.
> If we couldn't perfectly predict the flow of any given river when it floods, that doesn't mean we believed a thing that was fiction.
In this rather more complicated case it still works! If I know what the height of all the land is, and how the water drains away, and what the recent rainfall has been and where it fell, and how much water capacity each channel has, and how various crops and forests and flood defenses impede the water as it flows, then I should be able to predict very accurately where the water will end up.
And if I know what sort of weather East Anglia typically has, and which bits of land are designated flood plain and will be fine if they are occasionally underwater, and how strong my weirs and locks and pumps are, I can design the whole system so that Cambridge and Ely never flood.
The Environment Agency does this all the time. And every time Cambridge and Ely get expensively flooded, they notice that they are confused, and work out how to make their beliefs more accurate. And make their models better. And then they work out how to strengthen their systems so it's less likely going forward.
And when I was a student, I got shown pictures from the 1960s of my very expensive and very historically significant college in various states of glub.
As an adult, I noticed that the houses near the river were regularly inundated, but it rarely did any damage that couldn't be fixed (insurance companies have models too!)
And as an old man, the worst that ever happens is that the river goes up a couple of feet, and the poor old Cambridge Evening News puts a photographer on Victoria Bridge to take a photo of my boat sitting in a slightly bigger river than usual and put it on the front page with the headline "Boaters in Trouble as River Floods".
But the trouble is mostly that I am floating a bit higher than usual and have to wear wellies to get on and off, and even I am not terribly bothered by this, especially since it only happens once a year or so and is quite fun.
But even then, I think there is some bit of the Environment Agency noticing that things did not go quite according to plan, and adjusting some false belief.
And these days I think the Environment Agency is not actually very confused at all about this sort of thing. Mostly they are doing cost-benefit analyses and working out how much they will have to spend to make it so I don't need to own wellies and wondering whether it's worth it.
> I think cosplaying as mathematicians and physicians is how modern "science" got into this mess. It seems to not work well outside of those domains.
I think cosplaying the authority of maths and physics without actually doing the work is a lot of the problem.
Building models that make predictions and then noticing when they don't quite fit is the work. If the useless bastards were cosplaying that I'd have some respect for them.
But it's a broader problem than that in the sense that it's mostly publicly funded, and nerds and politicians do not like each other much or understand each other much. And designing systems where intelligent actors are acting in their own interests but nevertheless do even roughly what you'd want is very hard.
I suppose it could all be downhill from "the greater good" becoming "my personal welfare check" and this corrupting everything, including all basic science.. the only scientists that, by necessity, stayed honest, were the ones that had to show results at the end of every week.
Pretty much everything is like that. It's not just science or just government funded stuff.
But 'skin in the game' does tend to help, and 'public funding' does seem to make it worse.
The problem is that a man who can afford to live free can indulge his curiosity for its own sake. Whereas a man who needs someone else's support has other things to worry about.
But without public funding, we don't have any scientists except the idle rich and people with other jobs doing it as a hobby in their spare time.
One question is: 'how do we design a system that reliably turns public money into truth'. And that question is very very hard.
And another one is 'given that we think such-and-such a system might work better, how do we get it funded', which is even harder.
That is a totally excellent and coherent thing to believe, and every time we see water flowing downhill we should be slightly more smug about our correct belief, and if we *ever* see water flowing uphill then we should say "I notice that I am confused. One of the things I believe is fiction!". And then work out what the false assumption is.
I have seen a river change direction while I was watching it. I was very surprised indeed!
Most people say "Oh, an anomaly! But we are still basically right it's all just a bit unpredictable", and then never think about it again.
I think this is epistemically unvirtuous, and I am trying to teach myself not to do it.
This is why I (try to) like surprises and (try to be) grateful when I am wrong! It is always an opportunity to get stronger.
> every time we see water flowing downhill we should be slightly more smug about our correct belief
I don't even agree with that part. I think Bayes was wrong.
And I think we're like that guy watching the river change direction. This doesn't invalidate one of the things he believes ("water flows downhill"), it just means "it's a bit more complicated than JUST that and we don't quite understand all the complications and how they interact."
Do you actually think that if you've seen something happen thousands of times, you shouldn't be more confident that it will happen again than if you've never seen it happen?
Did you not get more confident that ex150 worked (on you) after you'd seen it working? And not more confident than that after you'd seen it working twice? And so on...
No, but that's not what you said. I think there should be an asymptotic approach to certainty. If you hit the button and you get shocked, you learn a lot the first time. Quite a bit the next 5 times. But the 50,000th time you learn almost nothing.
It's not a linear increase in knowledge. And technically all of this implies "all variables held equal" which is the tricky part that Bayesians like to overlook.
> And I think we're like that guy watching the river change direction. This doesn't invalidate one of the things he believes ("water flows downhill"), it just means "it's a bit more complicated than JUST that and we don't quite understand all the complications and how they interact."
Which is to say, he is surprised, he should examine the implicit assumptions that have lead him to be surprised.
If 'water flows downhill' then it can hardly start flowing uphill, can it?
The fact that he has already seen a very great deal of evidence that 'water flows downhill' is why he needs to be quite careful about what he's going to replace that with. Because the new thing has to explain all the old observations too.
> it's more of a "ughhhh my stomach is stealing all my energy, better sit/lay down and let it do its thing" feeling. Is that what you mean by "fatigue?"
I think so, yes, you've just taken in a massive load of carbs, your body is trying to switch over to burning them (the Randle Cycle), but it's not working properly, so the weird thing where you've just taken on board a lot of easily accessible high-energy fuel but you have less energy as a result happens.
If you feed dogs shortbread they go mental with excess energy (I have one sample here, the law may not be universal)
People always used to complain that children went hyper after eating sweets.
A long time ago I remember cycling half-way to Oxford and back, fast, and on the way back I just completely ran out of energy. Not out of breath or anything, just my legs wouldn't move very fast and I was crawling along. And we found a garage that sold little glucose-enriched chocolate bars. And every time I ate one, I would suddenly speed up for about fifteen minutes and then slow down again. It was pretty comical! Like they were power pills.
We moderns seem to get tired after taking in carbs. I don't think that's normal, or at least I don't think it's a sign of a system working properly.
You keep saying "one of the things I believe must be fiction." I don't really get it. I think it's fair to see that "we don't completely understand what's going on" and there's enough uncertainty in there that nothing you strongly believe is necessarily false?
On the recent fatigue thing: when I swamp hard on refeeds, I do get sort of a feeling that I suppose could be called "tired" or "fatigue." I don't usually get "fatigued" so I don't know what exactly you're referring to. That swampy feeling to me is less "tired" in the yawning & sleeping sense like lack of sleep, it's more of a "ughhhh my stomach is stealing all my energy, better sit/lay down and let it do its thing" feeling. Is that what you mean by "fatigue?"
> On the recent fatigue thing: when I swamp hard on refeeds, I do get sort of a feeling that I suppose could be called "tired" or "fatigue." I don't usually get "fatigued" so I don't know what exactly you're referring to. That swampy feeling to me is less "tired" in the yawning & sleeping sense like lack of sleep, it's more of a "ughhhh my stomach is stealing all my energy, better sit/lay down and let it do its thing" feeling. Is that what you mean by "fatigue?"
It is hard, I think, for a person in good health to even notice the difference between fatigue and sleepiness. I remember years ago having my doctor explain this difference to me, and it only really made sense because I used to be an endurance athlete, and was familiar with the idea that you can be utterly out of energy after hours of hard work, even though if you hadn't done a bucket of exercise you wouldn't be even thinking about lying down.
Sleepiness is the thing where your brain realizes it has been awake for too long, and there is less light around, and it needs to switch into its overnight state and do whatever mysterious maintenance it does when we sleep.
Fatigue is when you run out of energy, and it usually happens only when you've used up most of your reserves and damaged yourself and your whole system including your brain just wants to do nothing until the situation improves.
One of the more interesting things about ketosis for me was that I seem to be able to get brain and body fatigue independently. I rarely see body fatigue in deep ketosis (I'm not an athlete any more, and rarely operate anywhere near the limit), but if I over-exert myself mentally I can get this thing where I've got loads of physical energy but my thought processes just aren't working properly. All my automatic skills are still there but I can't think.
I suspect I can run low on ketones quite easily, and my brain can't burn fat like the rest of me can.
I think this is all 'glycolysis is blocked/probably by PUFAs'.
I have noticed that I am confused by the fact that other people seem to have what I think are symptoms of PUFA poisoning, but don't complain about fatigue much.
If you can get fatigued while swamping, but not in deep ketosis, then I am less confused than I was!
> You keep saying "one of the things I believe must be fiction." I don't really get it. I think it's fair to see that "we don't completely understand what's going on" and there's enough uncertainty in there that nothing you strongly believe is necessarily false?
It's a habit I'm trying to develop: If your predictions fail or the world surprises you in some way, you should remind yourself that reality is lawful, and that confusion exists only in the mind, and must be explained by some imperfection in your model of the world.
That's how mathematicians think, and physicists, and I think it would probably be a helpful attitude for other people trying to figure things out.
Whenever I hear medical people say "not completely understood", I hear "we have no fucking clue, but we are going to carry on believing what we already believe even though we keep being surprised by things".
Better to say: "I *notice* that I am confused. At least one of the things I believe must be fiction". And then you should review all the assumptions that went into the prediction you have just seen go wrong, and work out which ones are dodgy.
It's an impossible standard and I'm crap at it because I am some hopeless human thing with a jerry-rigged mind that is fundamentally incapable of reason. But I can at least remind myself what my ideal actually is, and try to notice when I need to think.
I think cosplaying as mathematicians and physicians is how modern "science" got into this mess. It seems to not work well outside of those domains.
I don't make predictions much for my experiments, because it's just not that stage of sciencing. I suppose I could write "I predict to lose 10lbs" before every experiment, and most would fail. But what would that get me?
Largely we're still at the "throwing poop at the wall" stage of science, and clearly we don't understand enough to make acute, specific predictions like "lose 10lbs." We can't control our variables enough, or maybe we don't know which the real causal variables are, because even when I attempt to do the exact same thing, the results are often quite varied. Trends/patterns are just about the best we have, so far.
I agree that admitting confusion is good. But "at least one of the things I believe must be fiction?" I think I've repeatedly commented how I disagree that your results mean you believe a fictitious thing; they merely mean you don't understand the system perfectly yet.
We believe that water flows downhill, and that you can make hills and dams and trenches. If we couldn't perfectly predict the flow of any given river when it floods, that doesn't mean we believed a thing that was fiction.
> I don't make predictions much for my experiments, because it's just not that stage of sciencing. I suppose I could write "I predict to lose 10lbs" before every experiment, and most would fail. But what would that get me?
You have models, and you make predictions. All the time! And you act on your predictions! It is *always* that stage of sciencing.
For instance "PUFAs are probably bad, if I eat them bad things will happen", or "on ex150 I tend to end up weighing about 210lbs", or "if I weigh more than 210lbs doing ex150 for a month will bring my weight back towards 210lbs".
If you did ex150 for a month and ended up 50lbs heavier you would be very surprised!
Surprise is the feeling where you have made predictions and they have failed.
But I think a lot of your predictions and models are subconscious, and result in feelings about what will happen. That's fine, that's how we work. My subconscious is much more powerful than my explicit reasoning and I tend to trust it more and rely on it entirely where I am not very curious or very motivated to get better at things.
But by trying to bring those subconscious models into the light and think about them, we can sometimes spot where there are incoherencies and errors, and by setting up a dialogue between conscious and subconscious reasoning and allowing them to talk out their disagreements we can sometimes persuade both processes to be more accurate.
It is not even much of an analogy, but I try to think about the difference between writing a program to tell me what will happen, explicitly made out of formal rules and calculations, and training a neural net.
You should be doing both! Neural nets are great, and often do better. But formal rules are also great, for different reasons. And often the best approach is to do both, and use both programs to improve each other.
But I mostly don't think "something I believe must be wrong" only "I must be missing some details." Because it's a different stage of sciencing. I didn't abandon PUFA or low protein the first time it didn't work as the only required & sufficient factor.
I think you're missing my point. I think you can do science very wrong, and I think the way you phrase it is basically wrong for the level of confidence you should have given your information and where we are "as a field."
It's just not that phase, mainstream science is way too rigorous (in a bad way) and inflexible and has actually long abandoned the scientific method, see also Feyerabend.
The reason we make so much more progress is that we do NOT follow this orthodoxy and dogma, but perform various different methods when appropriate.
> I think you're missing my point. I think you can do science very wrong, and I think the way you phrase it is basically wrong for the level of confidence you should have given your information and where we are "as a field."
I don't have very much confidence at all. But often I think that a thing is more likely than another thing. When the second thing happens I try to *notice*. That's all.
> But I mostly don't think "something I believe must be wrong" only "I must be missing some details."
These things are perhaps not too different?
'Heavenly objects move in circles with epicycles' is quite hard to tell apart from 'Heavenly objects move in ellipses' if you're only looking at the planets.
I think almost everyone would have agreed that Ptolomy was only missing some details. In fact I think his theory was more accurate.
I think these 2 things are quite different.
> The reason we make so much more progress is that we do NOT follow this orthodoxy and dogma
Well it is my orthodoxy and my dogma, and I follow it. It is not a widespread orthodoxy or dogma. I don't know many fellow believers! Although I have hopes to spread it.
> If we couldn't perfectly predict the flow of any given river when it floods, that doesn't mean we believed a thing that was fiction.
In this rather more complicated case it still works! If I know what the height of all the land is, and how the water drains away, and what the recent rainfall has been and where it fell, and how much water capacity each channel has, and how various crops and forests and flood defenses impede the water as it flows, then I should be able to predict very accurately where the water will end up.
And if I know what sort of weather East Anglia typically has, and which bits of land are designated flood plain and will be fine if they are occasionally underwater, and how strong my weirs and locks and pumps are, I can design the whole system so that Cambridge and Ely never flood.
The Environment Agency does this all the time. And every time Cambridge and Ely get expensively flooded, they notice that they are confused, and work out how to make their beliefs more accurate. And make their models better. And then they work out how to strengthen their systems so it's less likely going forward.
And when I was a student, I got shown pictures from the 1960s of my very expensive and very historically significant college in various states of glub.
As an adult, I noticed that the houses near the river were regularly inundated, but it rarely did any damage that couldn't be fixed (insurance companies have models too!)
And as an old man, the worst that ever happens is that the river goes up a couple of feet, and the poor old Cambridge Evening News puts a photographer on Victoria Bridge to take a photo of my boat sitting in a slightly bigger river than usual and put it on the front page with the headline "Boaters in Trouble as River Floods".
But the trouble is mostly that I am floating a bit higher than usual and have to wear wellies to get on and off, and even I am not terribly bothered by this, especially since it only happens once a year or so and is quite fun.
But even then, I think there is some bit of the Environment Agency noticing that things did not go quite according to plan, and adjusting some false belief.
And these days I think the Environment Agency is not actually very confused at all about this sort of thing. Mostly they are doing cost-benefit analyses and working out how much they will have to spend to make it so I don't need to own wellies and wondering whether it's worth it.
But it would be a mistake to say "One of the things I believe is wrong, water does NOT flow downhill!" and start performing rain dances.
Actually in the case I to which refer, that I personally witnessed and had to figure out on my own, that was actually the false assumption!
I didn't do any rain dances. Just because you're wrong about something doesn't mean you should jump to literally any random action!
> I think cosplaying as mathematicians and physicians is how modern "science" got into this mess. It seems to not work well outside of those domains.
I think cosplaying the authority of maths and physics without actually doing the work is a lot of the problem.
Building models that make predictions and then noticing when they don't quite fit is the work. If the useless bastards were cosplaying that I'd have some respect for them.
But it's a broader problem than that in the sense that it's mostly publicly funded, and nerds and politicians do not like each other much or understand each other much. And designing systems where intelligent actors are acting in their own interests but nevertheless do even roughly what you'd want is very hard.
I suppose it could all be downhill from "the greater good" becoming "my personal welfare check" and this corrupting everything, including all basic science.. the only scientists that, by necessity, stayed honest, were the ones that had to show results at the end of every week.
Pretty much everything is like that. It's not just science or just government funded stuff.
But 'skin in the game' does tend to help, and 'public funding' does seem to make it worse.
The problem is that a man who can afford to live free can indulge his curiosity for its own sake. Whereas a man who needs someone else's support has other things to worry about.
But without public funding, we don't have any scientists except the idle rich and people with other jobs doing it as a hobby in their spare time.
One question is: 'how do we design a system that reliably turns public money into truth'. And that question is very very hard.
And another one is 'given that we think such-and-such a system might work better, how do we get it funded', which is even harder.
I think the answer includes "weak men -> bad times"
> We believe that water flows downhill.
That is a totally excellent and coherent thing to believe, and every time we see water flowing downhill we should be slightly more smug about our correct belief, and if we *ever* see water flowing uphill then we should say "I notice that I am confused. One of the things I believe is fiction!". And then work out what the false assumption is.
I have seen a river change direction while I was watching it. I was very surprised indeed!
Most people say "Oh, an anomaly! But we are still basically right it's all just a bit unpredictable", and then never think about it again.
I think this is epistemically unvirtuous, and I am trying to teach myself not to do it.
This is why I (try to) like surprises and (try to be) grateful when I am wrong! It is always an opportunity to get stronger.
> every time we see water flowing downhill we should be slightly more smug about our correct belief
I don't even agree with that part. I think Bayes was wrong.
And I think we're like that guy watching the river change direction. This doesn't invalidate one of the things he believes ("water flows downhill"), it just means "it's a bit more complicated than JUST that and we don't quite understand all the complications and how they interact."
> I think Bayes was wrong.
Do you actually think that if you've seen something happen thousands of times, you shouldn't be more confident that it will happen again than if you've never seen it happen?
Did you not get more confident that ex150 worked (on you) after you'd seen it working? And not more confident than that after you'd seen it working twice? And so on...
No, but that's not what you said. I think there should be an asymptotic approach to certainty. If you hit the button and you get shocked, you learn a lot the first time. Quite a bit the next 5 times. But the 50,000th time you learn almost nothing.
It's not a linear increase in knowledge. And technically all of this implies "all variables held equal" which is the tricky part that Bayesians like to overlook.
> And I think we're like that guy watching the river change direction. This doesn't invalidate one of the things he believes ("water flows downhill"), it just means "it's a bit more complicated than JUST that and we don't quite understand all the complications and how they interact."
Which is to say, he is surprised, he should examine the implicit assumptions that have lead him to be surprised.
If 'water flows downhill' then it can hardly start flowing uphill, can it?
The fact that he has already seen a very great deal of evidence that 'water flows downhill' is why he needs to be quite careful about what he's going to replace that with. Because the new thing has to explain all the old observations too.