Hacker News new | past | comments | ask | show | jobs | submit login

> … despite the fact that many materials — including graphene, frogs and pliers — can exhibit similar magnetic behaviour.

This is a dishonest argument. Frogs and pliers don’t levitate on a typical magnet in ambient conditions, so you can’t say the behavior is ‘similar’ just because ‘everything levitates in a strong enough magnetic field’. If your measuring stick can’t distinguish these two significantly different behaviors, it’s a useless measuring stick.

Also, more generally, the logic of science is statistics, not Boolean logic. We all know that

[(p —> q) and q]

doesn’t logically imply p, but still observing q makes p more likely when thinking probabilistically.

Overall I found the article to be very dismissive of weak evidence and also shortsighted when considering the potential applications that we can’t think of right now (even if those are not the typical hoped-for applications like efficient power transmission).




The logic of science is logic. Statistics is used to provide confidence intervals.


Do you disagree with ET Jaynes then?


You'll have to be more specific: what are you asking me whether I disagree with?


I'm pretty sure the parent comment just dunked on you by demonstrating a deep well-read understanding of the underpinnings of both logic and statistics.

https://en.wikipedia.org/wiki/Edwin_Thompson_Jaynes

"Jaynes strongly promoted the interpretation of probability theory as an extension of logic. "


You're commenting on the wrong site (or just ignoring the rules & spirit of discussion) if you thought it necessary to tell someone they got "dunked on".


>> I'm pretty sure the parent comment just dunked on you

Oh, sorry, I didn't realise this was a dick-waving competition.


ET Jaynes has a book "Probability Theory: The Logic of Science". It's a nice book, and I was wondering if you had any thoughts on it.


I haven't read the book. What does it say?

On the matter of probability and science, I like how Karl Popper put it, although I can't find the text now so I must report it from memory. His point was that scientific hypothesis formation is an instance of inductive generalisation while probabilistic inference is a form of abductive reasoning, and so using probability to support an inductively derived hypothesis is basically supporting a guess, with another guess.

Statistics of course is not the same as probability. Personally I think statistics is a bunch of hooey.


Confidence intervals are not an appropriate tool for something like this. This isn’t an effect size of dubious magnitude. It’s a simple claim that something either exists or it does not.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: