Pup Biru

the fact that nothing happened doesn’t mean it was overblown. the fact that nothing happened means all the engineering effort and sleepless nights that went into the problem were well-spent

prevention that succeeds so completely to prevent a disaster that you don’t even notice is just a job very well done: not evidence that the disaster wasn’t real or very dangerous

stringify is used to serialize an object good enough in the large majority of cases because we have far more poorly implemented hacks to work around

i agree but hate to add the caveat that it (well, typescript) these days is by far the language that i’m most productive in, with the fewest bugs

it took me a long time to admit that to myself

it’s still a human rights violation

if they’d have called it “related reviews” it might have gone down okay

and gone down well if reviews about the publisher/developer/etc were shared between games automatically… even if they were separated

it’d be annoying for like 2 sec giving up better auto complete when i’m writing code (keywords: when IM writing code), or “intuitively find and replace non-identical things based on what i just did”

and then id get over it because i could buy RAM again

see you’re misinterpreting the reason they got away with it though… it looks like this whole oopsie doopsie my widdwe AI did a crimes sowwie but it’s actually just… and you’ll never guess it… ready for it? the law doesn’t apply to the rich…

I’m saying if there’s 1000x more lines of code than necessary, it’s unmaintainable slop. I could understand 2x or even 10x the lines of code just being a difference in architecture, but even 100x is getting crazy.

and i’m saying who’s to say what’s necessary with the linux kernel? it’s a one of a kind project: the largest software undertaking by humanity, ever

there probably are 100x more LOC than technically necessary… heck there are… how many LOC in the linux kernel? millions? there’s n-1 more LOC than necessary because it’s C and can be a one-liner (hell even python can be a one liner if you turn it into something resembling brainfuck more than python)

what are you comparing 2x, 10x, 100x to? bsd? windows? neither are particularly good comparisons because linux is so much more capable (in many respects) than either, and solves different problems of scale that either too

linux has no point of comparison for a simple LOC metric… the only thing we can judge is the output itself: its efficient to run, velocity of new features is great, bug count is low for a project of its size, and tech debt gets paid down

again, i think given those factors it wouldn’t matter if linux were 100x the LOC that it actually is: it is the benchmark that other things are measured against

LOC has always been, and will remain a terrible metric for anything except number of newline characters in a source repo

and if a display driver can’t be written in 1 line that doesn’t make the thousands of lines unmaintainable slop

some things just require more code

i’d argue if the linux kernel were billions of LOC, then that’s just what’s necessary

and necessity is probably relative at that point; many of your LOC are likely technically unnecessary abstractions to help with human understanding

LOC doesn’t inherently make things unmaintainable slop, as long as you have strong abstractions (which Linux certainly does)

one could argue that most Linux distros have billions to trillions of LOC and their abstraction is packages, which are kinda a level of abstraction higher than libraries, which are just a level of abstraction higher than a folder of util functions

things being “not my code” doesn’t mean it’s not part of your system

if it’s a partnership with microsoft, imo it’s an awful idea… to me it reads as:

everyone knows linux is secure AF… way more secure than windows in many contexts, so it sets you up in the “wait this is BS” mindset

the logical next step that you make, imo, is that the argument that closed source is more secure than open source in other contexts should be viewed with similar scepticism: office/libreoffice, onedrive/nextcloud, etc

(which is true: closed source software is often far more insecure than open source software)

as if we needed any more reasons to demand fibre

do NOT make mistakes

DROP *

much more preferable to

GRANT *

that is absolutely NOT the way that works in any functional environment

your ability to deploy quickly comes directly from your automation, and those automation tools - NOT developers - have the secrets in them

deploy to prod multiple times per day isn’t some win by itself… the ability for large teams (not 1 fuckwit and a goon squad of agents) to deploy without breaking things and in ways that are safe is the win here… anyone can deploy to prod multiple times per day… but anyone isn’t netflix (the originators of the “multiple times per day” line) with the uptime they achieve over years while doing it

if you talk AND type you can burn twice the tokens and get the same result! it’s a win win win

along with most modern languages… it’s the way we deal with async when you don’t want callback hell. it’s just a complex problem domain

like… what… JITs are complex so that’s a problem for V8 specifically?

they’re talking about inheritance as a suggested new way to pay for your copilot subscription

it’s like how the linux kernel isn’t semver: the number after the decimal point has a maximum of .19 and then the first number increments

except that one time

kernel version 4.20 is the only .20 kernel version

tmux > gnome

a great illustration of the dunning-kruger effect