sophie alpert
@sophiebits
Followers
63K
Following
130K
Media
630
Statuses
18K
logging lets you travel through time. debuggers let you travel through space.
16
77
536
new from me: everyone is wrong about that slack flowchart. systems are complicated if you make them complicated but you can also make them simple
2
7
59
RT @sophiebits: @lcamtuf great news, I just tested and this code already works perfectly today:. long *x = (long *)malloc(sizeof("Enough me….
0
6
0
if the p99 of a nonnegative data set is 5x the mean, you know that at least 1% * 5 = 5% of the sum comes from the top 1% of data points.
2
0
6
this is not true!!. even if you don’t know how to read an analog clock you know that it has two hands and the positions of the hands somehow indicate what the time is, which is all the requisite knowledge for this joke.
You know what’s crazy? There’s an entire generation that wouldn’t get this joke because they don’t know how to read an analog clock.
1
0
11
RT @pitdesi: Sad news: .as part of the T3 renovation project at SFO, we will be losing THE BEST moving walkway on the planet. get your boun….
0
90
0
gpt o1 scores 13.9/15 on the AIME but I got 14/15 in high school so I’m still smarter than a 5th grader.
0
0
19
A shark could swim faster than me, but I could probably run faster than a shark. So in a triathlon, it would all come down to who is the better cyclist.
0
0
3
what do I actually want?. is my only hope to use the mean instead.
2
1
1
knowing any particular percentiles for the parts tells me basically nothing about the percentiles for the sum!. I can of course like… filter to samples where the total time is >=p89, <=p91 but that’s atrocious and feels like it might be exceedingly unstable.
1
1
1
hey stats nerds:. let’s say I’m measuring some repeated process with total_time = part1_time + part2_time + part3_time. if I want to understand p90 total_time better, my instinct is to look at p90 of each part. but that doesn’t do what I want at all!.
3
1
9
is there any CSS I can add so that my first div is the same width as my second (i.e., so that it's only as wide as the widest wrapped line).
18
1
24
RT @Scaevolus: > _the_ portable way to get the current unix epoch in posix awk is srand(srand())
0
28
0
what game is this text overlay from?.
6
0
17
Seems like most popular DBMSes/caches aren't designed to support these operations, though let me know if I'm missing something!.
1
0
2