3 ms·

It's only when HN discuss a topic in which you are an expert in that you realise 90% of the comments are absolute nonsense. It happens every time there is a bi

It's only when HN discuss a topic in which you are an expert in that you realise 90% of the comments are absolute nonsense. It happens every time there is a biology topic as well.

grenoire 9y ago
Same with economics; when someone says in the comments for a rant piece that homo economicus is a core concept of all modern economics because that's what they learned in that one 101 class they took... Grr...
- dsacco 9y ago
  It must be quite fun for you to read so many armchair financiers here reconstructing modern economics from first principles on the one hand, while throwing in gems like “economics is not a science” because “not all participants are rational actors as usually assumed” on the other :)
  - grenoire 9y ago
    That's very apt. Except it stops being fun pretty quickly.
kristjankalm 9y ago
The Murray Gell-Mann Amnesia Effect
- icc97 9y ago
  That's an excellent talk from Michael Chrichton [0]. Plus this was another rather relevant piece to some of the comments too: > Often, the article is so wrong it actually presents the story backward-reversing cause and effect. I call these the “wet streets cause rain” stories. [0]: https://web.archive.org/web/20141025035056/http://www.thegreatideas.org/aww/TGIO332.pdf https://web.archive.org/web/20141025035056/http://www.thegre...
dsacco 9y ago
> It happens every time there is a biology topic as well. Let’s add cryptography and quantum complexity theory while we’re at it :)
lorenzhs 9y ago
It happens with core computer science topics as well. The misconceptions around Big O notation, for example, are astonishing. O(f(n)) is the set of functions that asymptotically grow at most as fast as f(n). It's not limited to measuring running time in the RAM model, or somehow captures the variable n, or anything like that. You can use it to measure memory use, or communication volume, or anything else you like. It can be defined in multiple variables, like number of nodes and edges, or input size and approximation factor, or whatever else you might need. And all of n², n²+n log(n), √n log(n), 42, and 1/n are in O(n²) because O isn't Θ, ffs.</rant>

It's only when HN discuss a topic in which you are an expert in that you realise 90% of the comments are absolute nonsense. It happens every time there is a biology topic as well.

grenoire 9y ago

Same with economics; when someone says in the comments for a rant piece that homo economicus is a core concept of all modern economics because that's what they learned in that one 101 class they took... Grr...

dsacco 9y ago
It must be quite fun for you to read so many armchair financiers here reconstructing modern economics from first principles on the one hand, while throwing in gems like “economics is not a science” because “not all participants are rational actors as usually assumed” on the other :)
- grenoire 9y ago
  That's very apt. Except it stops being fun pretty quickly.

kristjankalm 9y ago

The Murray Gell-Mann Amnesia Effect

icc97 9y ago
That's an excellent talk from Michael Chrichton [0]. Plus this was another rather relevant piece to some of the comments too: > Often, the article is so wrong it actually presents the story backward-reversing cause and effect. I call these the “wet streets cause rain” stories. [0]: https://web.archive.org/web/20141025035056/http://www.thegreatideas.org/aww/TGIO332.pdf https://web.archive.org/web/20141025035056/http://www.thegre...

dsacco 9y ago

> It happens every time there is a biology topic as well. Let’s add cryptography and quantum complexity theory while we’re at it :)

lorenzhs 9y ago

It happens with core computer science topics as well. The misconceptions around Big O notation, for example, are astonishing. O(f(n)) is the set of functions that asymptotically grow at most as fast as f(n). It's not limited to measuring running time in the RAM model, or somehow captures the variable n, or anything like that. You can use it to measure memory use, or communication volume, or anything else you like. It can be defined in multiple variables, like number of nodes and edges, or input size and approximation factor, or whatever else you might need. And all of n², n²+n log(n), √n log(n), 42, and 1/n are in O(n²) because O isn't Θ, ffs.</rant>