I really like the design of this HTTP library's API and will probably try it out in personal projects in the future: https://gitlab.com/honeyryderchuck/httpx#httpx-a-ruby-http-library-for-tomorrow-and-beyond … (You get parallel requests "for free" without having to muck up your code a lot or implement promises/event loops/etc.) h/t HN
-
-
Incidentally, this memo is one of the most salient examples I know of something I have seen many times in life: Some people are called crazy for getting the next big thing about 40% right while the consensus estimate of the rest of the world is less than 1% right.
Show this thread -
They're called crazy when they make their call and, when the future rolls in, people say "Anybody could have predicted the future, which was obvious, except for this person who got it 60% wrong."
Show this thread -
After you start looking for this pattern you'll see it a lot, and I think it should inform: a) how to be an instrumentally successful "crazy person" b) interpreting the work product of crazy people c) one's humility on estimates of future d) praise awarded to critics.
Show this thread -
Being instrumentally effective while being crazy: a) Operationalize crazy beliefs and win with them. Winning reduces the sting of being made fun of. b) Pick what subset of beliefs to share; establish a track record of being right on the lower-risk portion of your craziness.
Show this thread -
Interpreting the work product of crazy people: a) Try to rigorously evaluate arguments, particularly post-hoc, and determine "Was somebody right by accident or did they see something the rest of the world didn't?" b) Know a few more crazy people than seems useful. c) Find the 40
Show this thread -
c) One's humility on estimates of future The degree of success expertise has versus craziness in any one instance is formidably huge and the track record of success expertise has against all craziness is extremely low. Weight estimates (your own and others') accordingly.
Show this thread -
d) Praise awarded to critics The value created by, and therefore the correct amount of social esteem awarded to, successfully calling the results of an experiment after the experiment has been concluded is, to a first approximation, zero.
Show this thread
End of conversation
New conversation -
-
-
random nitpick — it should be “cf.” (no period after the c), short for “confer,” meaning “compare” in Latin. I used to think it was just some fancy way to say “see [reference]” but it really isn’t
-
Huh. Seems like I have *quickly greps* 371 incorrect usages of that in my writing. *sigh* Thanks; the more you know. (If it were a true nitpick I'd say "In this house we're linguistic descriptivists!" but this feels like "not errored enough into correctness yet.")
End of conversation
New conversation -
-
-
Really interesting memo. He seems to argue that: 1. Higher level APIs allow much more power than lower level APIs (which used to be contentious but is now more widely accepted) 2. Something about "expansion of notation" which I don't quite understand
-
I think #2 claims that being able to represent a wider variety of data within the same structure is more powerful He uses "strings" as an example of the most generic interface, though in today's world we might replace that with "json", which is common among scripting languages
- 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.