Astro Hacker News - The Legacy of Daniel Kahneman: A Personal View (2025)

nerdralph |next [-]

I read "Thinking, Fast and Slow" and some of the other references in the article. I found Kahneman's arguments persuasive, however the article makes me re-evaluate those conclusions.

When asked what is more probable, I think in terms of statistical probabilities. However the article makes an interesting argument that most people don't define the term, "more probable" the same way. I'm not convinced Kahneman was wrong, but I do see how simple changes in the wording of a question can lead to a material difference in answers. I also see that my own interpretation regarding the "correct" meaning of words aligned with Kahneman, and contributed to my general agreement with his conclusions.

Tomte |root |parent |next [-]

Kahnemann had the intellectual honesty to accept that large parts of his book are flawed, and he called on psychologists to clean up their act by doing a systematic multiple reproduction study program:

https://www.nature.com/news/polopoly_fs/7.6716.1349271308!/s...

lemonberry |root |parent |next [-]

"Kahnemann had the intellectual honesty"

I once heard an interviewer ask him if Kahneman was still susceptible to cognitive biases after reading the book. He said something to the effect of "absolutely, they're tough to escape". I really appreciated that. People that recognize and acknowledge the fallibility of their own minds are a breath of fresh air.

okintheory |root |parent [-]

I don't think that's a great example. If Kahneman claimed not to be susceptible, it would have greatly undermined his claims about the universality of these phenomena: many other people would presumably also not be susceptible.

lemonberry |root |parent [-]

If I remember correctly I took the interviewer's question to mean "now that you're aware of these cognitive biases are you still affected by them?" not "do you experience cognitive biases?". I don't see the first question at odds with the universality claim. The latter would be.

BeetleB |root |parent |previous [-]

Not "large parts". Just a few chapters.

voidhorse |root |parent |previous [-]

Probabilities are a philosophical rat's nest of sorts. When it comes to statistics, it's generally agreed that we're working with a frequentist interpretation of the meaning of probabilities, but you are right that a person with no prior background could well have a completely different understanding here (subjectivist probability, degrees of belief).

I also think stating presuppositions and limitations around observation and prior knowledge is monumentally important as soon as you begin talking in terms of probabilities, if you really want your statements to be clear, but most people don't do this. There are some ways in which I think the casual use of probabilities can actually be more harmful than encouraging a simple binary boolean dichotomy of "I know" or "I don't know" and need more information.

nerdralph |root |parent [-]

Isn't "I know" just a subjective threshold for the probability of being true? A layman may put that probability at 90%, while I scientist may put the probability at 99.999% before saying, "I know".

kayo_20211030 |next |previous [-]

I like the gracious way that Gigerenzer ends the piece, acknowledging Kahneman's courage in continuing to engage both civilly and respectfully with his critics - separating the argument and the person.

It reflects well on both men.

svnt |next |previous [-]

Gigerenzer is a treasure and there is a reason he is more accurate and less popular: statistics done right does not often compose compelling narratives, and narrative is what sells books.

jancsika |previous [-]

> Intelligent listeners then correctly infer what the doctor recommends and act accordingly.

I feel like if the author were really committed to this position, they would have stated it in the other direction. I.e., 6% to 20% of these intelligent listeners incorrectly infer that the doctor is conveying information through the framing of the question.

Something like the academic version of dogfooding. :)