Hacker News Reader

Comment by n2d4 | Hacker News Reader

Good article, but there's one thing that really bothers me:

> perhaps because of his serious and scientific background – he has a physics degree from Harvard and studied aerospace engineering at Stanford [...]

Which, for the thing he's doing now, is not a serious background! Knowledge isn't transferrable and there are plenty examples of people who transfer fields with unwarranted confidence and make totally bogus claims (see the Nobel disease [1]). I wish we would stop claiming the opposite, because especially in a field like nutritional sciences you can find some "evidence" for any theory (see eg. the Chemical Hunger drama [2]).

[1] https://en.wikipedia.org/wiki/Nobel_disease

[2] https://www.lesswrong.com/posts/NRrbJJWnaSorrqvtZ/on-not-get...

View on HN

It kind of is a serious background. The standards of intellectual and academic rigor in physics and aerospace engineering is far higher than nutrition science or whatever. What the domain specialists have is familiarity with the existing literature, but that is readily remedied as anyone with an interest and motivation can become familiar with literature like nutrition science in short order.

All things being equal, I would probably trust the judgement of someone that went deep in the hard sciences and engineering, and then educated themselves on the literature of some soft science specialty on the side, than the average person whose entire education was in the much less rigorous and more shallow soft sciences. I have seen far too much sloppy and motivated reasoning in the soft sciences to trust anything that comes from people that exclusively live in that world. I am sure there are proper scientists in those fields doing good work, but that isn’t the average.

I don’t know anything about this particular case, but it is much easier for someone with mastery of the hard sciences to become fluent in the soft sciences out of interest than vice versa. It may not be fair but I have seen so many real world examples of this, where hard science/engineering experts became recognized experts in softer science fields entirely out of interest, but never in the reverse, that it is difficult to ignore.

> The standards of intellectual and academic rigor in physics and aerospace engineering is far higher than nutrition science or whatever.

I'm not sure that's justified, at all. The domains are harder to experiment in - so the standards are more fitting for their domain. That doesn't make them better or worse, just different.

> All things being equal, I would probably trust the judgement of someone that went deep in the hard sciences and engineering, and then educated themselves on the literature of some soft science specialty on the side, than the average person whose entire education was in the much less rigorous and more shallow soft sciences.

You are, IMO, too dismissive in your language about entire professions. "Shallow, soft sciences".

I get where you're coming from - I'm more of a CS/Math/Physics person myself. But I think "our" side of hard sciences is far more dismissive than it should be - leading to implicitly trusting people who wave the right "hard science" credentials around instead of people working on the field for many years.

^ This. CS/Math/Physics people tend to have a massive blind spot in the awareness of just how much complexity exists in the biological sciences. Couple that with a background that makes these people have full confidence in their ability to understand complex systems, and you tend to have a larger issue than if the background and subject matter were reversed.

> I'm not sure that's justified, at all. The domains are harder to experiment in - so the standards are more fitting for their domain.

Not taking sides here, but I want to point out that there is a distinction between rigor in reasoning and diligence in testing (given feasibility constraints). The true answer need not be in the space of hypotheses that's feasible to test [1]. It's crucial that scientists appreciate that; sloppy reasoning tends to make strong conclusions from weak tests because there's a desire (bias) to emphatically drive outcomes even from weak conclusions.

[1] https://en.wikipedia.org/wiki/Streetlight_effect

So. I have a background in biophysics which is the part of biology where most physics folks go. It's highly quantitative, but still it struggles with the "assume a spherical cow" mental model of many physics and engineering folks. And the biology folks are stuck in a guild mentality (protecting their field by limiting the number of entrants) and very artisanal and resist increased quantitative methods.

Over time I've mentally recategorized entire parts of medical science into what I would say is non-science, or at least, a very different approach to science. In particular, nutrition, and many other areas, take extremely limited amounts of data and extrapolate it, making conclusions that far exceed what the data actually supports. And the folks publishing it got really good at convincing the larger press/medical world that they knew what they were talking about, which has led to all sorts of trends.

I would love to see more translation of quantitative physics into the medical sciences, but it's truly a challenging problem, especially because we (for good reasons) don't run the sorts of trials that would provide truly convincing evidence for causal mechanims related to disease, or whatever area of biology you care about.

It's like some kind of Gell-Mann effect corollary. Scientists and other educated experts quoted in a newspaper carry automatic authority unless you understand the target field well enough. But if you do, you realize they are no more trustworthy than a guy on the bus.

Titles don't help. An "environmental engineer" might not have sufficient knowledge about zoology or bridges to support whatever they are talking about, but a layperson might not realize that.

Coined as mann-gell if I'm remembering this video correctly.

https://m.youtube.com/watch?v=wBBnfu8N_J0

The Gell-Mann effect overrates having an academic degree and doesn't count the fact that people might become experts in something they haven't done formal studies for.

How can you tell whether I am as trustworthy as a guy on the bus on matters of software engineering? I am a high school dropout. Likewise, I'm no medical doctor, but knowledge is a spectrum, and I believe I know more than the average Joe on matters of nutrition because it is something that fascinates me and I've read and researched a lot about.

Given that I've read and heard about Taubes for 15 years, it is simplistic to say the least that he doesn't know anything because he's got a physics background.

tl;dr: the Gell-Mann effect only makes sense in a vacuum and if you really want to believe that having a degree automatically makes you an expert with worthwhile ideas.

The effect here is more about people claiming lateral authority, rather than the base claim of authority in the first point. But I take your point.

I’m also underqualified for my position and am wary of credentialism. But if I’m looking at a topic I don’t know well — e.g. seismic building codes- I have to start somewhere.

Software engineering is thankfully a field that is particularly accessible to people without formal training. Demand is high, functional skills can be proven on the spot, and employers are (usually) not in a position where their hiring decisions have potential legal consequences.

Bollocks, physics is an actual scientific field and aerospace engineering is actual engineering (compared to say "software engineering" where it's engineering in name only), similarly there's barely anything scientific in the "nutritional science".

If you have a real scientific background and real engineering background you are way more likely to do a proper job.

It frustrates me that people think nutritional "science" is a serious scientific field. And that the "nutritional scientist" is somehow better equipped - they have barely done any real science a single day in their life.

So what makes you an authority in defining what’s a scientific field? Care to share some actual data?

You don't need to be an authority to see how the majority of human nutrition science is barely a step above junk. Pick any journal in the field and go through the last few issues. You'll mostly find observational studies with subject reported data (known to be inaccurate). Very few long, term randomized controlled trials with meaningful endpoints.

It is worse than junk. most junk science is harmless, but bad diet advice can hurt your health.

If you can rigorously test your hypotheses in an experiment, it is a scientific field.

If you can't, or won't, it is not a scientific field.

There have been tons of nutritional experiments which are extremely rigorous.

Unfortunately, like most fields it’s the junk you read about because rigor is slow/difficult/expensive and therefore less common than BS. Just look at say fusion research and all it’s seriously flawed experiments.

And tons of nutritional experiments demonstrating the exact opposite result, that are also “rigorous”.

Sure, human biology is (hundreds of) orders of magnitude more complicated than say going to the moon.

But at this point in 2024 we “know” a lot more about physics than we do about nutrition.

And maybe that’s because physicists, when faced with contradictory data, keep digging until they figure out the basics once and for all.

That's just not true. We know a lot about the basics of nutrition. Actual nutrition scientists mostly agree on lots of different things.

The problem is that the public doesn't want to hear it, so the media keeps running articles that make it simple like "nobody knows anything". Which is just plain false.

Care to list some of the basic tenants of nutrition that are widely agreed upon?

About the only things I see as generally accepted are to eat a varied diet (whatever that means), keep total consumed caloric content near the amount expended, eat at least the RDA of micronutrients, and drink enough water to avoid overt dehydration.

And we're not even sure what the RDA should be, half the time.

At a practical level nutrition isn’t generally a problem, but that says more about the human body / food than the lack of scientific knowledge.

There is a long list of things that cannot be synthesized by people and must be consumed. (Elements like Potassium as well as specific nutrients like Vitamin C and Thiamine being well known.) There’s a longer list of things than can be synthesized but should largely be consumed ex:

Linoleic acid (LA), an omega-6 fatty acid, and α-linolenic acid (ALA), an omega-3 fatty acid, are considered essential fatty acids because they cannot be synthesized by humans.

The long-chain omega-3 fatty acids, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), can be synthesized from ALA, but due to low conversion efficiency, it is recommended to consume foods rich in EPA and DHA.

There’s also a lot of basic facts that get complicated when you dig into them. For example calories produced by burning foods isn’t 1:1 with the calories available to people. Cellulose for example burns but is effectively indigestible. One way to investigate this is to have some eat something and then burn their poop, followed by many studies that do so for various reasons.

There’s also a lot of nutrition science that straddles the line with medicine. Thus there exists many specialized diets for various conditions etc.

I'm not a nutrition scientist, so obviously not at all an expert that can probably chime in with some more findings.

But generally - yes, afaik there's mostly a consensus that total caloric intake is the driver of amount of adiposity, that maintaining a 10%-20% body fat percentage (for men) is optimal for long term health, that most diets are close from a health and fat-loss perspective - meaning, amount of adiposity is the main driver of long-term bad health, especially as compared to individual ingredients. There are some things that are known to be specifically problematic, e.g. trans-fats, at any amounts. And there are healthier and less healthy foods, some of which we know - but total caloric intake is the only driver of body composition, and a bigger driver of body health.

Most of this is basic - probably because, as I said, I'm not a nutrition scientist myself so only understand the basics myself.

That said, a large part of the population doesn't even understand these basics either, preferring to constantly talk about and argue about specific diets, keto vs. vegan vs. whatever. And science mostly has that question covered, as far as we can tell right now.

For a more in-depth take on this from someone who actually is an expert, I recommend videos by Mike Israetel:

https://www.youtube.com/watch?v=7FM9VAH_EXw&list=PLyqKj7LwU2...

https://www.youtube.com/watch?v=TYeZVfPxwKM

There’s probably a grey area for social sciences like economics.

I also tend to think there’s real nutritional science but it’s being drowned out by all the snake oil.

Snake-oil and moralizing. Dietetics suffers from the same issue that mental health does: there's an underlying stigma that causes people to inflict a moral judgement on the ill; fat people - or so the old line of thinking goes - are immoral. A shocking number of people, including many doctors and scientists, view diabetes as a "fat people disease," and that diabetics are in effect being punished for their gluttony. This isn't always a conscious bias, but it has very negative effects.

It reminds me of the studies they've done in Canada and the US about how natives (American Indians) die a lot of the time in the ER because they're immediately written off as alcoholics or drug-addicts, and so their complaints are evaluated through the lens of "this person just wants opiates."

It’s been said that iteratively developing a software based solution is like testing a hypothesis and seeing what works

Where did this definition come from? Did you just make it up? It excludes all of mathematics and most of astronomy, for example.

Well on some contexts maths is apart from science, see e.g. the STEM acronym :)

People generally consider, although this is not universally agreed, that math isn't science. To me the distinction matters, although it's fairly subtle.

This is underscored by frequently cited articles, like: "The Unreasonable Effectiveness of Mathematics in the Natural Sciences", in which math is treated as separate, but intimately related, to natural philosophy.

I think you can do math without science, but you can't do science without math.

Would you have a physicist would self-taught medicine as your doctor? Of course not.

While a scientific or engineering background might demonstrate capability to understand another scientific field, that doesn’t replace the near decade of education people in that field have.

I would, compared to 8 out of 10 doctors I’ve met.

A dedicated physicist who self-taught medicine would allow hoofs to be zebras, not horses, some % of the time.

False equivalency

Doctors need practice in the hospital and do an excruciating amount of rounds surrounded by senior experts.

Nutritionists? No.

Given a nutritionist who has studied the books and gotten a degree and a physicist who has studied the books and hasn't gotten the degree I would be hard pressed to choose the nutritionist

I am someone with a computational physics background who is coupled with someone who is training to be a surgeon. We lived together during much of her medical school, and thus, are constantly around medical professionals.

I would absolutely trust a physicist (who has done actual empirical lab work) who has deeply studied a problem in the nutrition space over someone conventionally trained in nutrition.

You may be stunned to see how little numerical sense so many in the medical field have. And as someone who grows medicinal plants and follows that literature closely, I read tons of papers in that space. It is full of experimental design flaws and dubious correlations.

I have worked on a medical device that went through clinical trials.

They have a point. It's easy to come with your high IQ and miss some obscure subtle data that people in the field learned the hard way. Sure the man is capable of understanding it quick and even do more.. but it's too easy to get carried in your own newcomer theory.

https://xkcd.com/793/

What makes you think you're right about any of that? Do you have any evidence to back up your (crazy) theory that an entire field of study isn't real science?

Cause from where I'm sitting, nutritional scientists do plenty of real science - studying a very complex problem that poses a lot of challenges to experiments, and still coming up with a lot of great results that teach us tons about nutrition.

Most people that criticize the field do so because they read a bunch of articles. Or because they're in the world of hard sciences, and are simply arrogant enough to think they know better than everyone else, even in fields they've never studied.

Let’s pretend all of that is true. How does a background in the “hard sciences” with experiments involving materials that can’t make their own decisions prepare you for a field with experiments where you have to account for the fact that you can’t imprison your test subjects for 6 months, and that when they are on their own may not follow the experimental protocols (or even lie about that fact)?

Gary Taubes has written multiple scholarly books on the subject. He has won the Science in Society Journalism Award of the National Association of Science Writers three times. I think a training in one scientific field does qualify a person to go spend their time digging into other fields. How well they do is a matter of personal ability and effort and one would have to read their output and judge for oneself. I have read many of Taubes's book and many of the papers cited in his books and have concluded that he is a well qualified expert in the field.

Not true. The fact that he was able to get through two rigorous programs at prestigious universities with low acceptance rates automatically gives him a certain amount of credence.

Well, if it came to nutritional or health advice, would you give Trump more credence, over some random person from a state school? After all, Trump is a Wharton graduate.

I guess there are considerations to be made for nepo babies you know paid their way in but that’s not most people.

Exactly. I don't think his guess is any better than any else given the poor track record of the dieting industry overall. He made a lucrative career by latching on to a theory, that like virtually all diet fads, falls short when applied to real people in a controlled setting.

No one has any idea which diet (if any) is best, or the connection between insulin and obesity, or if fats are good or not, etc. All hunches. Carbs have been falling out of favor for a while now but obesity getting worse.

Idk Ive been a keto acolyte for a long time and anyone who Ive ever known that did the diet was wildly successful on it.

So his claims really dont seem to fall short in my own experience and the experience of people I know.

The problem with anecdotes is that they're not science. And as there is wide variety in humans this is very much a case of ymmv.

Anecdotally, my wife went onto a low-carb diet 18 months ago. She likes low carb food, and isn't a huge fan of bread, potatoes etc anyway, so it was easy. She's not total though, still eats a rusk at breakfast, and has a (home made) pizza once a week. She also started exercising a lot more, and found a sustainable rhythm for that. She's lost 50 lbs. (Around 25kg).

Interestingly, her blood sugar levels have not dropped significantly, but they did drop a bit.

I sporadically do low-carb, and always lose weight doing so. I like bread, potato's, rice though so my diet is much less obsessive. My results are more fluctuating but slowly trending down. I'm in no rush, but eould like to drop around 15 lbs this year. I lose about 1lb per week if I'm "fairly strict" - exercise levels don't change (I exercise a reasonable amount.)

Incidentally, I've found that just slowly reducing quantity has helped as well. Eating from small bowls and plates has helped me.

None of this is data. And I still "cheat" (I'll happily eat normal in social situations, and often on weekends etc) but I seem to have found a pattern I can maintain long term. It's a "diet" in the sense there's good and bad, but its also "lifestyle" in the sense that I try to be "moderate" with the bad.

Science isn't the memorization of facts. To graduate from Harvard, you must actually understand how to do science at a high level. Many of those skills transfer between subjects. Having a physics degree might not make you a great biologist, but it's also not nothing.

Reply on news.ycombinator.com