Hacker News Reader

Comment by bilsbie | Hacker News Reader

[flagged]

How do people who don't want anybody to "trust the experts" plan to live in a world without experts?

To be a bit less snarky, basic education on philosophy and science should make it clear what are the disclaimers implicitly assumed when trusting the experts.

A serious answer to your question, something I learned from interacting with a couple especially open flat-earthers once: they want everyone to "trust their senses" and "do their own research". In one sense you could say they want to destroy science as an institution, but in another you could say this kind of behavior is closer to the Platonic ideal of science: replicable experimental results that reiterate useful theories can be passed between people rather than just handing down insights from on high.

Their attitude is certainly encomiable, and yet one cannot help but wonder how they could reach such spectacularly wrong conclusions when, in 2023, it is not so difficult to verify (in multiple ways) that the Earth is round (thanks to long distance traveling, instant communication, and some geometry).

The crux is that human knowledge has progressed so much that "trust your senses" is simply not feasible anymore, as the majority of modern scientific discoveries cannot be verified with the resources available to normal people, and one would need years of study to actually "do their own research" in a productive manner (as I was discussing with the sibling comment, who is convinced that a few days of study are enough to judge neurology research).

People who are labeled as experts and who talk the loudest in society aren’t always capable of independent critical thought, so a lot of what they say is useless.

Nothing is stopping you from independently investigating a subject. You don’t need to know about phylogenetic trees and be a bio wizard to understand how neurons work and learning how neurons work would take you a day or two max. Pick a narrow but not too narrow scope, research it and make your own conclusions. I don’t know why anyone would ever blindly follow “experts” other than sheer laziness

It is always good to be skeptical of people who talk the loudest, regardless of their role in society.

> learning how neurons work would take you a day or two max.

I seriously question your definition of expertise. Harvard offers a bachelor's degree focused entirely on neuroscience, with at least two separate exams focused only on neurons. And that is barely enough to understand current research in neuroscience, since it usually builds up on literature that is too recent for a bachelor's.

If you took your high school and college education even somewhat seriously, you would have already learned how an action potential travels through a neuron, and you’d have a good grasp of science fundamentals to read any research paper. I learned it in my high school biology class along with a host of other things. Also, who says your so called “experts” know those details either? I’ve met a physician (recent new grad) had no idea what mitosis was, something I’ve known since I was like 12 and I develop software for a living. We’re supposed to trust what these people say? Nah

Got it: neuroscientists don't know about action potentials, and high school education is enough to appraise any research paper.

Between one Jira issue and the next, try to read some actual neuroscience preprints and come up with a few reasons why that piece of research is good, and a few why it is bad.

Take this one [1], for example (Cerebrovascular disease drives Alzheimer plasma biomarker concentrations in adults with Down syndrome):

> Main Outcomes and Measures: We examined the bivariate relationships of WMH, Aβ42/Aβ40, p-tau217, and GFAP with age-residualized NfL across AD diagnostic groups.

Are these biomarkers specific enough? Did they miss any? Why did they limit the investigation to bivariate relationships?

> We [...] examined whether 1) GFAP mediates the relationship between WMH volume and p-tau217 concentration, 2) whether p-tau217 concentration mediates the relationship between WMH volume and NfL concentration, and 3) whether p-tau217 concentration mediates the relationship between GFAP and NfL concentration.

Why did they test these three hypotheses? Did they miss anything interesting? Why did they choose that specific method for mediation analysis? What limitations does it have? Are there alternatives?

> Two specific percentile thresholds were computed: [...]. These thresholds initialized a Gaussian mixture model (GMM) and expectation-maximization algorithm within the white matter segment of the FLAIR images, using two components to represent hyperintense and non-hyperintense voxels.

What do you think of their method to quantify white matter hyper-intensity from the MRI scans? What percentiles did they use, and how sensitive is the analysis to these choices? Is a gaussian mixture model appropriate?

If you were a peer reviewer, would you think that this paper is ready to be published? What feedback would you give to the authors? What is the significance of these result, and what future research do they support?

See you in a one day or two max, cowboy.

[1] https://www.medrxiv.org/content/10.1101/2023.11.28.23298693v...

Big words don't scare me. If I cared about research related to people with Down Syndrome, then I’d read it and understand it thoroughly, but I don’t. Critical thinking skills are largely innate, it doesn’t matter if you have a PhD or high school degree. Credentials say nothing about your intellectual curiosity or your motivation to learn new things

https://www.youtube.com/watch?v=MkqQIY7J0fQ

On the contrary, critical thinking can be developed, but it will not save you from ignorance. It takes much longer than two days of reading to be able to thoroughly understand research like the one above, I really suggest you give it a try.

Not sure why you started rambling about credentials now, but I broadly agree with your point. However, again, all I have been arguing since the beginning is that it takes a lot of time and effort to reach expertise in a topic. I really don't see what is so controversial about this. Or do you really think that an average person can learn software engineering and criticize your work in only two days?

For the typical reader of HN, phylogenetic trees would be far easier to figure out in a few days than how neurons work, especially as "how neurons work" isn't even established.

You know what I meant by “how a neuron works” so stop nitpicking, and speak for yourself about not being able to learn it in a day or two

Yes, I'm sure you could know exactly how the well-studied synaptophysin, synaptobrevin, neuroligin and PSD95 cooperate to control neurons in a day or two. Particularly in the context of schizophrenia.

Easy as pie.

Edit: Here's a freely available relatively recent article on just one aspect of neurons, the synaptic proteome, which contains about 1,466 proteins [0].

I will check back in a few days once you've figured out how they all work together.

[0] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7776996/

If I was interested in this paper, I would read it and understand it like I do with all the bio papers I read…all research papers do is narrow in on something highly specific, they aren’t hard to understand once you have the vocabulary down. It’s like reading someone else’s code. And I highly doubt your so called trusted “experts” like Fauci understand any of it

Understanding what a paper says is considerably easier that understanding its limitations. See the questions I asked in the sibling thread for a taste of the difference.

Or think in terms of code: it is easy to understand what some code does, but to understand why it is like that you need to know a lot more about the context surrounding the problem.

And really? Synaptic vesicles? Proteomes? I learned this shit in high school

Mostly obfuscate and lie.

https://en.m.wikipedia.org/wiki/Replication_crisis

It’s bad. It’s so bad we’re at the “coin flip is about as reliable” stage of science. In some soft sciences almost 75% of studies cannot be replicated. AKA are false.

We can’t tell if salt raises blood pressure in humans but there are legions of people here that will state with all the conviction of a Jesuit priest that climate models can predict climate 100 years into the future - all without evidence, or replication, or open data, or falsification, and if you don’t agree with the consensus you will never be published. Literally censoring contrary views. That’s some serious “just shut up and believe me!”.

The certainty that I see based on very uncertain experiments is seriously suspect. I can only hope the structure of scientific revolutions puts this down in the next generation, this one seems hopelessly lost.

But yeah, that’s “science”. And if you don’t believe that just wait for the non scientific personal attacks that are replies to this comment.

"Trust the experts" is always "trust my experts". If you ask them to trust other people with the same (or similar) degree, but a different opinion, they don't. It's just an appeal to authority rather than a real philosophy.

Honestly? These kinds of stories should enforce the general idea. Trust has to be seen as a long term game, though. And requires well run checks to keep things trustworthy.

So, if you view trust as only a short term, "the answer is always right," then that will fail. But you should not have a sole source here.

And this style of system is no different from authority anywhere else. You should be able to trust your local religious/community leader. But you should also not have one that is absolute authority on their own authority.

Good question. Blind faith is never a good idea. However, one ought to ask, what is the alternative?

This approach seems to work rather fantastically:

https://en.m.wikipedia.org/wiki/Illusory_truth_effect

Just keep a constant propaganda campaign going of how flawless "science" is, and people will eventually believe it to be necessarily true.

Reply on news.ycombinator.com