Potato diet: A post mortem and an answer to SMTM's article

post by Épiphanie Gédéon (joy_void_joy) · 2022-07-14T23:18:26.691Z · LW · GW · 34 comments

Contents

  Personal experience
    Diet overview
    Reasons for the diet
    Before the diet
    Start of the diet
    Switching midway
    After the diet
    Making sense of all this
  SMTM's methodology
    Preregistration and conflation of the experiments
    Clustering
    Safety
    Tunnel-vision
    Foregoing the conclusion
    Testimonials
    Congratulations
  Lessons learned
    Data tracking
    Stable baseline
    Safeguards
    Low risk assessment
    Mood management
    Generalizing
  Conclusion
None
34 comments

Slime Mold Time Mold (SMTM) has just published the results of their studies on the potato diet. The idea is to eat nothing but potatoes for four weeks, and report how it went.

You can read SMTM's reasoning here, the basic idea is that they've observed some annectdotal data point of some people losing substantive weight while eating nothing but potatoes and want to reproduce the result.

Natália Coelho Mendonça has written in a very detailed and well-written article [LW · GW] why it was probably not lithium, SMTM's leading hypothesis of why obesity is so prevalent.

As SMTM's potato diet is not directly linked with lithium, some readers might think that even if SMTM was wrong on that one, the potato diet could be interesting to try.

Having done the diet, I thought I could give a subjective data point into all this. I wanted to be pseudo-anonymous, but this post has been nagging at me ever since finishing the diet.

This post will not be as sourced as Natália's article, I do not intend to explain the mechanism or absence thereof of the potato diet. I am not knowledgeable in nutrition or biology to really have any insights into the mechanisms of this.

This article will serve both as a report of how it went down for me, a postmortem of mistakes I made, and a general impression of the SMTM methodology that results from this.

While I am glad I tried the potato diet, and it gave me great insights into what can affect my energy and mood levels, I feel that my following of it was dangerous in a way SMTM did not account for. I am also finding their methodology has been very poor, and their recent article is loose in the way it draws its conclusions.

There are some things I either completely blundered and that were my fault, or that I calculated the risks of and decided they were worth the gamble. For instance, I asked SMTM if my bipolar anorexia would muddle the data, to which they said that they advise against bipolars taking the diet. Seeing as I am quite atypical in my bipolarity (my manias are never a problem for me, I actually look for ways to trigger them), I decided it was worth a shot.

So I am not trying to hold SMTM responsible for what happened or what might have happened. I am trying to

• warn other users here who might be tempted to follow the diet of precautions they might want to take into considerations
• narrate my experience as a whole in case there is something interesting for other people.

A quick TLDR of the (long, sorry) following post:

• Personal experience:
  • First started the diet eating self-made, oilless fries and chips/crisps, as well as potato soup. Then, as there were no effects, switched to whole boiled potatoes on their recommendations
  • Got some bruises/maybe oedema and general dizziness after stopping the diet, which I think is mild refeeding syndrome
  • Lost 8 pounds in 4 weeks, regained them completely 2 weeks after the diet
  • I was in a quite low energy state for months, I am now having bipolar swings and low sleep for two weeks (as I knew about being bipolar, those were risks I knew of and somewhat expected)
  • I still do not understand why eating the potatoes whole and boiled had an effect when nothing else worked. Especially seeing that I seem to be the only one. Solanine poisoning? Placebo effect? Weird timing?
• SMTM's methodology:
  • The design of the study was very weird, sometimes it would be about trying to reproduce the potato diet, and other time to search for a diet that would work.
  • The features of the change they were looking for were themselves not fixed in time
  • I do not think SMTM took enough precautions in safety communication
  • SMTM seemed focused in confirming their hypothesis rather than challenging it
  • Weird felt sense that might just come from my shift in perspective as a whole
• Lessons learned:
  • Start taking daily pictures of myself as a way to log changes. Diversify what kind of data to track.
  • If you want to track the effect of an intervention, ensures your life will be stable enough during to avoid confounders.
  • If something seems potentially dangerous, open a manifold market asking "Will I in retrospect think X was hazardous for me to try?"
  • Do not go into something with such a high investment as one month before you either have a very thorough understanding of it, or that you have already experimented with smaller variants of the experiment
  • Alimentation seems to have high leverage over my mood and experimenting more with this could be a solution to trigger my hypomania more reliably, and just to manage my mood in general.
  • This might also reflect more of this type of experiment within the rationalist community as a whole, and I still have to ponder more about this. - Conclusion:
  • If you want to try the potato diet, please don't be reckless. Low risks high rewards only apply if the risks are low, and I don't think there's been enough evidence of this for doing the diet a full month.
  • Please don't take up on SMTM for doing a study on the Onion Diet unless you can have way more responsibility in your approach.

Personal experience

This section is given for context, feel free to skip it.

For full transparency, I feel that it is important to give my datasheet and emails I've sent them. I think one of the answer they've sent back as a mail is crucial in how I felt and that I should include it as well. As it is anonymous, I believe this is fine, and I believe the inclusion of my data should not be problematic now that the study is over.

If you feel a link or text I give is not ethical/correct for me to put here, please do signal it and I will remove it posthaste

I am number 02142044, here is my data sheet . I also reported some of my experience on a manifold market

Diet overview

The gist of my diet:

• Did not use oil at all during the whole month.
• Used potatoes from the mall
• Started with oil-less self-made chips (crisp), moved on to whole boiled potatoes at their suggestion
• Some occasional breaks during the weekend.

Reasons for the diet

My main motivations for taking the diet were threefold:

• I want to lose weight for aesthetic and/or social reasons. Because I have had induced anorexia two times before, during which I lost 10 kg in a month, I was looking for a way that would not trigger it and that would be safe for me. It seemed to me that this diet could fit the bill.
• As mentioned before, my manias have never been problematic for me. I am looking for a way to be more frequently in a hypomanic state, as it is very pleasant and productive.
• Mostly, I am very much into applied transhumanism and looking for new ways to do science and data-collection, so this idea seemed very appealing to me. I estimated the risks were not high based on the accounts they had and other information I gleaned.

So I was excited to explore this diet.

Before the diet

Before starting the diet, I was in a quite low point in mood-space for months (might have been depression). I was a bit apprehensive of what dangers the diet could have and was talking frequently to my brother about this.

Start of the diet

As I started my diet, I was mostly eating the potatoes in chips/crisps or soup, and was seeing no effects at all. It put me in an uncomfortable state of constant hunger, where I would always want to eat. I tried several variations to remediate that: eating them cold, adding salt, stopping seasonings...

I was starting to get a bit desperate from this, and I was wondering if I should just cut the diet altogether as I saw no changes at all. One of the main reasons I was unsure about quitting is that I thought they might not include my data if I did not continue until the end, and I was relatively certain, seeing how much my experience differed from the others, that it would not see any improvements at all.

12 days in, I thought I might quit and sent them this email:

Hi!

I am on the 11th day of the potato study (my spreadsheet: https://docs.google.com/spreadsheets/d/1x_s--xhkr5OZJU0_jcX5XRxdm_m99mVC-gECX6MEaEE/ ), and am not feeling much different, either in terms of energy, mood or weight. Would it give you more interesting data if I continued the study?

I do not really mind, the biggest problem for me is that I do not know what you would recommend, like if I should only eat them whole and not sliced&broiled, or unseasoned. If I had to eat them only boiled and nothing else without any seasoning I might start to find it untenable, though I can try.

Otherwise, I might stop the diet, as it leaves me quite hungry and does not seem to have any benefits.

Thanks a lot!

To which they answered

Thanks for checking in. If you're feeling poorly you should end the diet and finish the study, we don't want anyone to take risks with their health and comfort. But if you're feeling ok here are a few ideas of things you could try that might make a difference:

• Start peeling the potatoes, eating them without most or all of the peel. Some people seem to find this makes a difference.
• Try cutting out all oil from the diet. The SMTM author who tried the diet didn't start losing weight until trying the diet without oil.
• Eat more whole potatoes. Someone mentioned that cooking potatoes in smaller pieces might cause some of the potassium to leave the potato and end up in the cooking water or whatever, so if potassium is part of why the diet works, that could be a problem. You could also do sliced potatoes in a soup and just drink the water they were cooked in.
• Try avoiding tomatoes. We see on your sheet that you sometimes have had tomato sauce. Other participants have mentioned that they seem to stop losing weight when they use ketchup, and given that we've recently found sources suggesting that tomatoes may be lithium concentrators, they might be worth avoiding.
• We recommend that you keep eating your potatoes seasoned, that seems to work for almost everyone, and it's more enjoyable. A lot of people swear by Frank's sauce. When the SMTM author who tried this was on the strictest form of the diet, they were just having potatoes, Frank's sauce, salt, and black coffee, and that worked surprisingly well for being so restrictive. Salt might even be important if the mechanism is something like a potassium/sodium balance thing.

Let us know if you have any questions!

(I should note that on day 7, I started working on my (non-physical) projects a lot more, but I did not notice it until the diet was almost done. Before the diet fluctuation in my concentration and procrastination was very high, so it did not seem unusual, and it did not fit any cases of the data sheet - I did not feel any more energetic. This productivity is still ongoing a month later, and it is very rare for me to maintain such a high streak. However, I am skeptical this is related to the potato diet, as I was also ongoing a lot of relationship/life changes.)

Switching midway

Following their recommendation, I stopped broiling the potatoes and only started eating them boiled whole from day 12 forward. Weirdly enough, this seems to be the moment a change started to kick in, I was getting less and less tired of the diet, and also eating way less. The hunger sensation started to dissipate, and this was a pivotal moment in the diet for me.

On day 17, hypomania kicked in, and lasted for about 4 days, my mood kept swinging after that.

I do not feel comfortable describing this fully, however after switching to boiled potatoes, there were signs of body dysfunctions. Because of complicated reasons, this did not alert me while it should have. This was my fault and caused by a faulty assumption of mine.

I do not yet know how to make sense of this, except to have better tracking in general. The signs resorbed swiftly after stopping the diet.

After the diet

You can read the full report I sent SMTM here. This was written just after I was done with the diet.

Two days after stopping the diet, I noticed a bruise and maybe oedema on my forearm, about 10 cm wide. I did not sadly take a picture of it. I realize now that this was likely mild refeeding syndrome which is what happens when the body has a sudden influx of calories. I also became significantly tired for a few days.

To the best of my knowledge, SMTM does not say anything about how to ease back out of your diet, just "you can eat whatever you like". I believe, now that I have read more about it, that this is a mistake. If by any mean you are doing the potato diet, please do consider ways to ease back out of it.

Something very interesting that was happening to me that might be related is that I noticed my hunger was offset. In the sense that I would start eating, and my body would just feel full, but the sensation of wanting to eat was still there. I thought it was a sign that the diet worked and that my "wanting to eat" sensation would adjust, but now that it didn't, and my hunger has rekindled even greater, I am wondering if it was a way to regulate against the refeeding mentioned above.

2 weeks later, I have now fully regained the weight I lost and feel quite hungry all the time. I am noticeably more fluctuating in my mood, with 2-3 mood swings per day and sleeping ~6 hours a day. I am now still working a lot more, with noticeably less procrastination/watching youtube/etc.

Overall the outcome is positive for me, I highly prefer frequent mood shifts to constant low. I cannot however say that

• the outcome was positive in expectation
• that it was related fully or partly to the potato diet (my prior is that it is a loose link).

Making sense of all this

I still do not understand why eating the potatoes whole and boiled had an effect when nothing else worked. Especially seeing that I seem to be the only one. Solanine poisoning? Placebo effect? Weird timing? The heat wave?
The fact I fully regained my weight while feeling insatiable makes me lean toward "it was just malnutrition", with solanine poisoning lingering in my mind.

The bipolarity symptoms are still present. This could be a very weak update for lithium, though I do not think they are discriminant enough to not be explained by "just malnutrition+a lot of life changes".

SMTM's methodology

Preregistration and conflation of the experiments

One of my biggest gripe with this experiment is that it was very unclear what they were going for. A lot of the general vibe, which I can appreciate, is that they were trying to explore the space of diet: "We see that some persons are very successful in a diet made only of potatoes, is there something to it, or is it just anecdotal myths? Let's explore around eating only potatoes."
This could explain why they told me to change the way I was doing the diet, while in other cases it could have be seen as data manipulation.

But those features clash with the design of the study: It only makes sense to go for a month if you have a well-prepared plan in mind. The affirmed purpose of the study is also to reproduce a result, and their article presents it as such.

I think the general intent of "let's explore around this" is very interesting, but I regret there has not been any kind of preregistrations and the guidelines were so loose. Another point which perplexed me is that they gave feedback on how other participants were doing, but not opening the data to everyone. It gives an air of studying it together, while all the insights are centralized by SMTM.

I believe the best way to have gone about this would have been to recognize the features of the anecdotal successes around the potato diet and to search for variations of the potato diet that would give those features.

In particular, that it seems to occur very rapidly (less than 5 days), and to open a Scott Alexander style of study: Preregister what you are going to try and for how long (probably just a week), and report on your own trial. Then only try to reproduce when there is already a well-defined plan you want to ask your participants to stick to.

Clustering

This is related to my latter point. In the answer SMTM sent me, they say

• Try avoiding tomatoes. We see on your sheet that you sometimes have had tomato sauce. Other participants have mentioned that they seem to stop losing weight when they use ketchup, and given that we've recently found sources suggesting that tomatoes may be lithium concentrators, they might be worth avoiding.

This was I believe the first moment I started being uneasy about this. When asking for advice, it had been 4-5 days since I had tomatoes, and was still not seeing any changes. I was also not seeing any changes for the beginning of the diet. This, I believe, was atypical of the diet's effect they were looking for, which usually kicks in less than 5 days and is fairly robust to eating something else.

I am not saying they were necessarily wrong in this. It might be that the "kicks in less than 5 days" data points were actually outliers. I am saying that the effect they were trying to find seems to itself have changed over time.

Safety

While SMTM does say to consult a doctor, I believe there are several points they have not been cautious enough in the impact of:

• A month is a lot. Why start with a month? Is there any basis for this? This is especially weird when their data sheet goes up to 9 weeks (!!) and they say "We think you probably should stop here (or at least email us to let us know how it's going). But if you really want to keep going, just keep adding more columns!" which seems a mild way to put it at best.
• Not warning about refeeding syndrome and ways to ease back from the diet as mentioned earlier.
• In general, very little data about what kind of impact such a drastic diet could have, what to expect, and what to look for in terms of problems.

EDIT: I'm not saying they should have made sure it was 100% completely safe. I am saying that the same way they instructed to supplement in vit A and B12, they should have quickly checked for ways to secure this experiment, and communicated on this degree of safety and certainty in it.

For instance, saying clearly (and in neon signs) "We did not do any research into what kind of impact this can have and there may be very non-obvious negative impacts you will not notice and be aware of"

Tunnel-vision

I think a general issue is that their experiments seems to try to confirm their leading hypotheses rather than refute it.

As an example, here are two things that puzzled me:

• SMTM seems to take prevalence of hypomania and agitation in participants as a strong update toward lithium clearing. I believed that too, as discussed in the next section. However, this now seems wrong to me and I think they could have discounted it a lot more at that point:

  • Hypomania and agitation is a common answer, especially during a physically stressing time. For instance, sleep deprivation is quite known to cause hypomania or full-on mania.
  • For another anecdotal example that is the closest I could find to the potato diet, Mr Beast also experienced hypomania when fasting for two weeks six days after start

• SMTM does not seem to really consider alternative diets. They do mention mono-diets, and other type of diets, but since they were working from anecdotal data point in the beginning, it seems weird to me they did not contrast it with others (I mean wikipedia gives George Sitwell who ate only chicken and Howard Hughes who ate only canned soup as examples. I do not understand why they did not contrast the anecdotal successes of the potato diet with them and try to see if there was something very specific about potatoes from the start)

Foregoing the conclusion

Related to the previous point, a lot of the effects Slime Mold Time Mold discusses in their more recent article seem weird to me, in that I am not sure most of them distinguishes with a potato diet as opposed to simple malnutrition. Weight loss, hypomania, cholesterol, libido, ...

One way this could resolve is if it actually was safe to follow through. Then the effects themselves would not matter as much: Reproducing malnutrition's benefits without its drawbacks would be awesome enough.

However, I don't think that SMTM has done enough to show that it was safe. There is little about potential solanine poisoning and ways to avoid it (especially since they said to eat the peels at the beginning), about what to track before stopping, etc

Testimonials

This is going more in the "diffuse feeling" part, i.e. issues that may not be issues and depends on your prior in how you interpret them.

SMTM quotes me as saying:

You probably already know this, but I find it credible a potential reason as to why the diet works, if it does, is that it is helping clear lithium, which would also help explain the mild hypomanias people experience. https://jasn.asnjournals.org/content/10/3/666 seems to indicate that potassium and sodium can help with clearing lithium. That is also why I started salting more.

(I no longer support this for reasons mentioned above)

The thing is they already had this as a potential hypothesis, quoting from their mail

Yes, this is our suspicion as well! This might also explain why some people get feelings of dread/despair instead of hypomania.

I am confused as to why they quote me. It could be a simple explanation like they want to give credit, or that they want to emulate discussions and problem-solving. It could be that since the it's probably not lithium [LW · GW] article released, I no longer support this hypothesis, and so seeing it here is strange for me.

One thing however that I am wary it could be read as - whether or not that is their reasons for doing so - is an update toward lithium, i.e. "we are not the only one to think this!"
So just in case, and because I am still feeling uneasy about all this, please disregard my comment as an update, the fact a random person who undertook the study think/thought that lithium contamination is the source of obesity should not be an update in itself, only the link I gave.

Congratulations

This last diffuse feeling point is not about SMTM itself, but about comments I've read generally. I still add it as something to be wary about if you are doing the potato diet.

Something that rubbed me the wrong way is the fact that some people would congratulate me for "sticking through it", and others would more generally say "congratulation to everyone who made it through".

It might be that one problem I see is that this directly contradicts the safety precautions one should take about quitting if feeling sick or unwell, and added irony in that the potato diet was supposed to be easy in the first place.

Lessons learned

Data tracking

In general, I am not tracking data regularly, and I should fix this.

But on the study itself, one of my shortcoming in all this is that I relied on SMTM's template while I should have made my own rows for tracking. I should have asked myself general categories to put in, especially anything I was worried about (like agitation or physical weakness)
Also, I should have diversified what I tracked, track variables I do not know the impact of yet, or that are holistic in nature. Taking photos of myself regularly is a good way to notice any progressive change for instance, and to have a global view of what is happening.

Stable baseline

Relatedly, I notice a lot of things are changing in my life at the same time. I should wait for all of it to calm down before experimenting more if I want to have data, as an unstable baseline makes the data confused.

Safeguards

Another failure of mine is that, while I did see the diet was potentially dangerous, I did not set enough fallbacks for me to track how I was doing. In particular, I knew that change could be progressive and that the jump from "this feels weird" to "I should quit" is not automatic.

One way I see how I could have solved this is to open a manifold market asking "Will I in retrospect think that my doing X was hazardous?" and update it frequently. Then, if the probability gets too close to YES, stop immediately and resolve it later.

I did open a market, but it was asking the wrong question with the wrong incentives, as it was about how long I was going for.

This is another thing I will have to solve more generally and do not have yet a very clear answer to.

(A friend of mine remarked while proofreading this post that I should also have taken into considerations my friends' and relative's reaction to my announcing that I was starting the potato diet. He himself had a severe disgust reaction when I mentioned it to him.

I am still unsure how I should have factored this into my decision)

Low risk assessment

I was already doing self-experiments a lot, as it fits my general vision for applied transhumanism. For instance, I take nootropics like ALCAR or MSM, my reasoning being that it seems generally safe and that I want to explore my own configuration space.

I still stand by this. The mistake I made for the SMTM potato diet is not that I should not have tried something like the potato diet. It is that I should not have committed to a full month of this. I think what I should have done is to try a week of it with a very specific approach to it, then note, and take at least a month of break.

In general, I should fragment my life interventions more, trying different costless variants, then easing more into the ones that seem successful.

Mood management

I learned that some things, such as where I live, what I eat, who I talk to, etc... can have drastic impact on my mood and my bipolar symptoms. I will focus more into way I might impact my mood with alimentation, for instance maybe fasting a day could help improve my mood a lot when needed.

Generalizing

This was already something that was on my mind for a while. Scott writing against pascalian medicine for instance has been a big update I am still processing. It seems to me that creating a narrative about some isolated points (re: lithium and hypomania) is very easy to do, and it is easy to miss the big picture.

It might be that I should be way more cautious about weird hype memes in the lesswrong community in general, and that I am not grounded enough in reality. I will be more cautious in joining such experiments or trying 'weird hacks' from the lesswrong sphere for the time being until I have a more definite answer.

Conclusion

Please DO NOT run your own Onion Diet. At least not the way SMTM has done their potato diet. Be responsible about how the participants are going to follow your study, and how hygienic you are with the resulting data.

If you want to try the potato diet, please please please:

• Find evidences that it's safe-ish. While I am all in favor of risk computation and trying things that have little risks high rewards, I think the risk of one month of eating nothing but potatoes has been very understated.
• Have some safeguards. Talking regularly to my brother about the diet was not enough. Have some clear signs you look out for that you are at your limit and should stop now.

I do thank SMTM for having run this study, and I am glad I started it, though I should have stopped it sooner.

Thanks to arthurrainbow [LW · GW], my Big Brother (sic) and others for proofreading this post

34 comments

Comments sorted by top scores.