[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: trail.jpg (899 KB, 2600x1733)
899 KB
899 KB JPG
hiking edition

>Advent of Code is an Advent calendar of small programming puzzles for a variety of skill sets and skill levels that can be solved in any programming language you like. People use them as a speed contest, interview prep, company training, university coursework, practice problems, or to challenge each other.
https://adventofcode.com/

/g/ leaderboard join code:
796940-44255af5
anonymous-only leaderboard:
383378-dd1e2041

previous thread: >>103466287
>>
File: solve_dfs.png (672 KB, 5278x3842)
672 KB
672 KB PNG
idiomatic Rust solution
>>
Is calendar-anon already filtered?
>>
File: aoc2024 d10.png (2.26 MB, 1886x3612)
2.26 MB
2.26 MB PNG
Help me, anons.
I've got a Rust solution that works on the sample inputs but the terminal just suddenly quits when I try to run it on my input.
What could the problem be? It should be pretty fast and I'm not using unsafe so it shouldn't be a segfault or anything like that.
>>
File: Neil_Season_Four.jpg (118 KB, 703x1017)
118 KB
118 KB JPG
bigboy #10
url: http://0x0.st/XhOX.txt.7z
grid: 7500x7500
silver: 433087
gold: 624588
>>
>>103471266
How are you so fast on each new thread.

>>103471278
Yes
>>
>>103471266
at what point do you get dinged for >>>/global/rules/13
>>
I envy the GameBoy guy
>>
File: idiomatic_calendar.png (2.63 MB, 9861x4126)
2.63 MB
2.63 MB PNG
>>103471278
no
>>
>>103471278
hope not
>>
>>103471291
> 6. The quality of posts is extremely important to this community.

Seems like it could do good for some people to re-read the rules.
>>
File: crazy-goats-on-cliffs-5.jpg (327 KB, 880x1227)
327 KB
327 KB JPG
this is now a goat thread
>>
>>103471285
>9 flat_maps in a row
now this is what I call idiomatic Rust
>>
>>103471285
Good bait, I give it a 7/10. Had to deduct one point for not calling it idiomatic, when it clearly is.
>>
File: 1733748369024790.jpg (274 KB, 1280x1500)
274 KB
274 KB JPG
>recently discovered via /d/ that I have an IQ loss fetish
>decide to clean my mind up with some advent of code
>today is literally 10x easier if you just BROOT like a retard
>anons keep posting fat elf pics
not fair
>>
File: Screenshot.png (278 KB, 1200x1680)
278 KB
278 KB PNG
>>103471287
C, 0.2s with OpenMP
>>
>>103471321
>fixed size array
you didn't solve it
>>
>>103471318
>input-chan
>bigboy-chan
what about example-chan?
>>
>>103471318
how do you bruteforce a search method? it's literally a bruteforce algorithm to begin with
>>
File: aoc_2024_10_b.png (460 KB, 1930x2024)
460 KB
460 KB PNG
I love recursion.
>>
>>103471370
see >>103463185
>>
I've never actually worked with graphs before.
I hope what I'm doing works.
>>
File: aoc-2024-day-10.png (71 KB, 1980x1488)
71 KB
71 KB PNG
I solved it, but at what cost... I think it's time to filter myself out.
>>103471321
nice. I wanted to post my solution, but this made me realize how retarded it is. now I am ashamed, and feel bad.
>>
>>103471388
Oh hey, it did. Neato.
Time for part 2.
>>
File: 1733834905200132.png (44 KB, 612x597)
44 KB
44 KB PNG
pytoddler solve
got a feeling tomorrow's puzz will be tricky
>>
>>103471415
nice!
>>
>>103471416
what font? Go?
>>
>>103471416
CRT shader?
>>
File: file.png (1.36 MB, 850x1202)
1.36 MB
1.36 MB PNG
>>103471285
Update: Good news, I fixed the terminal crashing. It turns out I was trying to allocate a vector with 16100811494719130952 entries in it, so I removed the first collect() call and now the program runs correctly!
Bad news: even in release mode based on how long it takes to iterate through the first 10B possibilities the program is expected to finish running on Tuesday, June 11th, 5596
>>
>>103471413
it's fine
git gud
>>
>>103471370
>>103471386
do not the example
>>
File: 1723047299176698.png (87 KB, 979x673)
87 KB
87 KB PNG
>part 2 is less work than part 1
what did eric mean by this?
>>
>>103471448
go mono
>>103471461
no it's just a background image on the terminal
>>
File: aocday10clojure.png (68 KB, 702x509)
68 KB
68 KB PNG
Clojure solution for part 1

gotta go, hopefully gonna do part 2 later
>>
File: akuma.jpg (144 KB, 890x501)
144 KB
144 KB JPG
>>103471504
i remember doing project euler years ago and after solving, looking in the forum at the J/K/Q solves that were one dense line of pure thought. it was like when i was with my friend who has a gift for music and despite being wasted mashed out the wild part of rach's 3rd.
tfw literally me
>>
>>103471486
its fine, us brooters will wait for you.
>>
>>103471219
Anon, what do you think it says?
>>
>>103471376
If you prune it’s not bruteforce
>>
File: 1.png (179 KB, 294x301)
179 KB
179 KB PNG
>>103471318
>>103471370
there is also this bigboy chan
https://files.catbox.moe/t41nl7.jpg
posed minutes before today's puzzle. AoC++ perks are real guys
>>
>>103471285
>>103471486
wow... textbook example of average rust dev and tranime
the stereotypes are real
>>
File: image.png (206 KB, 1123x205)
206 KB
206 KB PNG
>>103468568
>>103469785
golfed it down more

>>103469934
the trade is acceptable
>>
>>
File: christmas_elf_girl.png (449 KB, 768x768)
449 KB
449 KB PNG
>>103471593
good job anon, btw are you one of the luachads from last year? iirc you were using nelua or some other dialect with vectors
>>
>>103460811
Is Eric seriously dropping hints for AOC++ paypigs??
>>
Haskell
import Data.Array
import Data.Char
import qualified Data.Map as M
import qualified Data.Set as S

solve1 g = sum $ map (bfs g 1 . S.singleton) xs
where xs = [ i | (i,e) <- assocs g, e == 0 ]

bfs _ 10 xs = S.size xs
bfs g n xs = bfs g (n+1) (S.fromList ys)
where ys = concatMap (steps g n) $ S.elems xs

steps g n (r,c) = filter valid [(r-1,c),(r,c+1),(r+1,c),(r,c-1)]
where valid y = inRange (bounds g) y && g ! y == n

solve2 g = sum $ map (bfs2 g 1 . M.fromList . (:[])) xs
where xs = [ (i,1) | (i,e) <- assocs g, e == 0 ]

bfs2 _ 10 xs = sum $ M.elems xs
bfs2 g n xs = bfs2 g (n+1) (M.fromListWith (+) ys)
where ys = [ (y,a) | (x,a) <- M.assocs xs, y <- steps g n x ]

readGrid xss = listArray ((1,1),(h,w)) $ concatMap (map digitToInt) xss
where h = length xss
w = length $ head xss

main = do g <- readGrid . lines <$> readFile "input"
print $ solve1 g
print $ solve2 g
>>
>>103471620
yes, he really sold out this year! i guess it's the last
>>
File: itsarace.png (83 KB, 1443x917)
83 KB
83 KB PNG
>>103471294
I don't know why you would.
I currently have two builds racing because the first one I forgot to put a check to not recur if the value has passed the goal.
The one of the right is catching up, but its still taking a fucking long time.
Perhaps I should reorder the operations to something like concat, multiply, add. It would probably cut out overruns faster.
>>
>>103471679
>Perhaps I should reorder the operations to something like concat, multiply, add
This reduced my runtime by 30% or something like that, iirc.
>>
>>103471616
yeah, but i wasn't using some weird dialect though. just good ol luajit.
>>
Now that the dust has settled. What happened to oofmatoes.
>>
>accidentally solved part 2
i'll take it, i guess
>>
>>103471708
He oofed his dick and became a Rust dev, but because Rust is too hard for him he failed to solve other puzzles
>>
File: aocpp.jpg (105 KB, 900x599)
105 KB
105 KB JPG
you awake, one charred leg trapped in a pyroclastic floe
the elf that was your guide has been carbonadoed into a blackened, smoking husk and is emitting a putrid odor. wisps of smoke flute out of his blackened blood-grimed nostrils and pointed ears. his eyes have boiled and popped, and his seared skin has shrunken his lips back into a grisly smile.
this hiking trail was supposed to be safe, you think, it had a high rating. no time for this, the llmfags will be coming out soon. the sun has already fallen behind the volancic rim looming all around you.
taking an amphipod out of your cargopant pocket, you attach it to your leg just below the knee. as flesh contacts flesh the simple neuronal pathways in the amphipods brain fire and it begins eating through your leg.
...
you must have passed out. it is dark, you are freed from the floe and the blood-gorged amphipod is lying sizzling and roiling on the charred rock containing the stump of your foot.
then you hear something that chills your blood.
the mumbled promptings of a thousand pajeets as they fire up their llms.
you try to scream but the prompting and the whir of a million fans switching on drowns out your cries.
>why didn't i pay for aoc++
>>
>>103471288
they/them wait for a new thread, last thread already had a bunch of solutions
>>
>>103471708
crashed out because he was using an LLM
the guy came from nowhere
only /ourguy/ faster than him was fucking leijurv
>>
>>103471728
I'll give you a 10 for creativity
>>
File: codeimage-snippet_10(2).png (270 KB, 748x1416)
270 KB
270 KB PNG
This year really is easier
>>
Hello, everyone!

I an sorry to announce that my son, Kevin Laverne also known here as Oofmatoes, has passed away. I don’t know what happened, he was so happy one day and the next day he started crying hysterically repeating over and over ”my LLM can’t solve it”. I tried to comfort him but he pushed me away, and when I tried again he punched me so hard that I lost consciousness.

When I woke up I found him dead in the kitchen hunched over his laptop with his penis in the blender. The police told me that he was reading about cargo and crates which is strange because he was never into manual labour or any physical activity.

Now you know. I will log off from this site and mourn together with my boyfriend. The funeral will be held next Saturday in our local church and the casket will be closed.
>>
>>103471782
shut up whore and post boobies
>>
File: geohot.png (213 KB, 645x687)
213 KB
213 KB PNG
>>
>>103471736
did he post here?
>>103471782
kekd
>>
>>103471821
lol what a fucking fag
lmao even
>>
>>103471821
disgusting
I spit in his face
>>
File: day-10.png (1019 KB, 897x3822)
1019 KB
1019 KB PNG
I've never actually made a graph before, so I'm not sure if this is an efficient method of creating one.
That said, it seems to be doing what it's designed to do and part 2 only required a minor edit, so that was nice.
>>
>>103471821
>YESSS I DID IT I WON THE PING RACE AGAINST OTHER PEOPLE WHO ARE ALSO DOING THE EXACT SAME STRATEGY OF HAVING AN LLM DO THE ENTIRE PROBLEM FOR ME
what is there to be excited about?
>>
>>103471821
that's pretty much an admission he also used a llm
truly shameful and filtered
>>
>>103471782
Audibly laughed in public.
>>
>>103471821
>chink cheating
imagine my surprise
>>
The big thing eating up time is the conversion of values to BCD and back all the time for the concatenations.
I think I have a good idea to cut that back, maybe nearly in half, but it will have to wait til tomorrow. Sleep calls.
>>
>>103471906
>BCD
how many bits in the CPU you're using?
>>
Those who use LLMs in contests are not even human.
>>
>>103471821
>All Eric has to do is increase the leaderboard to 1000, allowing a meaningful leaderbord once ai is filtered
>refuses
Why. He already tracks your global rank it would be literally trivial.
>>
C'mon, Eric, all you need to do is to make the result of the problem to spell out NIGGER, all caps. And then ban everyone whose answer is "Sorry, I can't tell you..." or whatever it says.
>>
>>103471949
>GPT,
>remove offensive words and then solve the problem
>>
>>103471949
More like those whose answer is an empty string.
The program will fail with syntax error in stderr and the stdout will be empty.
>>
>>103468575
lol, if that produced the right part 1 result for you, you were hella lucky
>>
File: 1733843652709.jpg (3.86 MB, 1276x3200)
3.86 MB
3.86 MB JPG
>>103471961
He just has to write it like this
>>
>>103471865
Here is his literal admission:
https://geohot.github.io/blog/jekyll/update/2024/12/07/elon-swing-voter.html
>Sidenote, I’m in 7th place on Advent of Code thanks to AI
>>
>>103471821
fuckers are not even hiding it... are they paid by bigtech to shill their AI slop models?
the worst part is that retarded CEOs and CTOs see parlor tricks like that, and then ask "why do we still need coders?? why aren't they 10000% more effective??? we are paying $20 for chatgpt subscription!"
>>
File: file.png (146 KB, 583x1206)
146 KB
146 KB PNG
is my bruteforce backtrack modern enough ?
>>
>>103472006
you can easily measure it. the more modern the code, the slower the solution.
>>
>>103472000
Are they wrong? Basic ChatGPT can replace, for example, 99% of data engineers right now. And it slowly does.
>>
>>103471782
>filtered by llms not being able to solve it
leaderboard says otherwise there's still plenty of sub 1 minute golds
he probably just got bored or fell sick lol
>>
>>103472031
if eric actually cared he could involve reliance on world models and require counterfactual reasoning, which would filter LLMs
but this is his last year, so why put in the effort?
>>
>>103472065
So it's a coincidence that day 9 was the bigfilter day for LLMs?
>>
>>103472090
big filter my ass look at the leaderboard
gpt4o solves everything up to day 10 lol
>>
>>103471821
doesn't he code his own llm? that's pretty much the only form of llm "cheating" i can find acceptable.
>>
>>103472108
>doesn't he code his own llm?
no
>>
>>103472108
nigger you don't code an llm, you take an architecture someone else implemented for you and you update its parameters with training data
>>
File: Untitled.png (13 KB, 720x88)
13 KB
13 KB PNG
>>103472090
Is the LLM filter you're speaking of in the room with us right now.
>>
>>103472142
>everyone must use the same tool
>>
>>103472142
deboonked
https://kadenbilyeu.com/blog/a3e009b8-aca5-4686-a4c5-ca016d6d0227
>>
>>103472160
>LLMs are absolutely great and amazing but i would NEVER EVER use one in a human competition like AoC; and those who do use them should be banned.
>>
>>103472031
ChatGPT can't even write regexes that do what you want, I doubt it can do complex "data engineering" tasks that do what you want
>>
>>103472187
and btw, the import part here is "do what you want"
>>
File: day10washed.png (318 KB, 1860x1957)
318 KB
318 KB PNG
modern conformant dogmatic bjarneous parallel safe C++

>>103471287
1.3s, could get it faster with some optimizations but readability and modernity >>> shaving less than a second

>>103471604
nice theme gruvboxbro

>>103472006
pretty modern

>>103472027
execution policies would like a word with you (shame we don't have them for constrained range-based algorithms yet)

>>103471486
bweh :3

>>103471413
moderate levels of modernity, you WILL NOT be filtered anon

>>103471321
not modern enough
>>
>>103472160
>This isn't good code, you won't see anything like this outside of cybersecurity.
this is because most cybersecurity fags are self taught
t. cybsec fag who writes code like that
>>
>>103471486
it's not about the destination anyways
>>
>>
This problem is so fucking boring that I don't even care about big boy. What the fuck is this dogshit Eric?
>>
File: day10.png (1.01 MB, 1968x4526)
1.01 MB
1.01 MB PNG
Apparently I solved gold while doing silver
>>
am I the only retard here who takes literal fucking hours to solve these puzzles? good thing I am unemployed, or I would get filtered days ago.
>>
>>103472285
possibly, i've solved all in under an hour this year
all fairly trivial so far
>>
>>103472285
15 minutes tops to solve, then 3 days to golf it
>>
>>103472187
At this point AI hype is like 1/2 truth and 1/2 lie.
Like every corporation now claiming that they replaced it "all with AI". And reality it is their programmers "just" using AI for better tab completion.
Or "AI makes it through xyz exam" and in reality it was doctors taking it by its hand to produce correct solutions.
LLMs are obviously impressive. But everyone wants to push this lie that they work like in scifi movies and "any moment now". When in reality progress was like one gigantic leap 2-3 years ago and now it is is just everyone desperately shoving it into their workflow because $current_buzzword.
In reality the only things it has truly "innovated" and/or replaced is misinformation, essays for high school and human science students, online spam, SEO trash, gooning templates/images, and online support chatbots.
>>
>>103472300
damn... I hate being low IQ codelet so much.
>>
>>103472285
Guess why you’re unemployed
>>
File: file.png (146 KB, 1600x900)
146 KB
146 KB PNG
>>103471287
$ time 10_01 bigboy.txt
433087
10_01 bigboy.txt 0.95s user 0.02s system 99% cpu 0.975 total
$ time 10_02 bigboy.txt
624588
10_02 bigboy.txt 0.84s user 0.04s system 99% cpu 0.882 total
>>
>>103472285
Are you that anon who thinks reading from stdin is cheating?
>>
>>103472316
unreadable
>>
>>103472285
is this your first year doing aoc? if it is don't feel bad it was like that for me my first year. over time you get better at them and get faster.
even then there are puzzles like the sea monster or the amphipods or the fucking 3d cube...
>>
>>103471947
All Eric has to do is get rid of the global leaderboard entirely lol. It's virtually impossible to moderate at the scale AoC is running at now; brings toxic jeetcode culture to what should be a fun Christmas event; and, because recruiters are pulling from the top spots on the global leaderboard, gives LLMfags a tangible incentive to cheat because of the potential reward.
>>
>>103472312
could be because I quit, and don't apply anywhere. it doesn't change that I am a retard.
>>
>>103472311
just get good
it will take more work, and your pattern recognition will be subpar, but you can probably learn to do these in 15 minutes if you're not a black lesbian
>>
Are LLM cucks filtered yet or what? If no, what's the final solution to the LLM question?
More parsing? Non baby shit difficulty?
>>
>>103471573
giga
>>
>>103472351
>final solution to the LLM question
kek
>>
>>103472090
>muh llm filter made him rage quit
There are literally solutions on reddit if he really wanted to cheat. Day 10 seems to be a breeze for llms too
>>
>>103471620
> one of the 20 guesses for the next day was correct
> does aoc++ hints really exist?
>>
>>103472351
another anon and i have the final solution
we will release a puzzle on day 26, which i encourage everyone to throw at an LLM
if it goes smoothly, we will release more puzzles next year
will give more info in later threads this week
>>
File: batman_no_bg.png (213 KB, 412x848)
213 KB
213 KB PNG
>>103472316
readable and modern

also good on you storing the directions in a constexpr array, wonder why I didn't think of that. btw what compiler are you using? do all of them support <print> already?
>>
>>103472370
He might have ragequit after he wasn't on heckin globalino leaderboard
>>
>>103471679
cool setup.
Do you do embdded stuff for work? Or is t
>>
>>103472351
The days with a graphical output are always fun
>>
>>103471947
Just list the first 100 people without points and times and in a random order.
There done.
>>
>>103472351
The only real solution is to get rid of the cheating incentives to begin with. LLMfags are only emboldened because of the bragging rights that comes with a leaderboard spot. If there's no competition, no one would give a shit about them
>>
File: tgf.png (750 KB, 2039x2117)
750 KB
750 KB PNG
>>
>>103472324
no
>>103472330
last year was my first, and I filtered myself out after finishing day 5. I couldn't put in any more time, and I was miserable. it's a bit better this year.
>>
File: 19JT98v.jpg (132 KB, 1024x640)
132 KB
132 KB JPG
>>103472285
dw anon, i'm also slightly retarded and I take an hour+ to solve the puzzle and then a few more hours refactoring/optimizing
>>
>>103472341
holy. we are in the same boat, except I only take 30mins to solve to problems
>>
>>103471287
On my old PC, non-optimized BFS written in C:
time ~/a.out bigboy.txt 1
silver: 433087

real 0m1,475s
user 0m1,471s
sys 0m0,005s
time ~/a.out bigboy.txt 2
gold: 624588

real 0m1,533s
user 0m1,531s
sys 0m0,002s
>>
>>103472285
No. I've done most years, but I like to take my time and write good code. And since LLMs seem to be mogging all the Jane St and turbo autists from leaderboard I have even less reason to speed it up. Slow your roll.
>>
>>103472429
If its your first year and you dont have a computer science background you will be slower because most of the problems use fundamental CS concepts. You either can derive it yourself and take hints and take the fact that you will learn something new everyday. Then next time it will be easier.
>>
>>103472461
There are still legit people on the leaderboards though. They record themselves why solving it.
>>
>>103472411
>>103472413
This. It really is not a difficult problem to solve.
And I have no idea why Eric is so obsessed with the leaderboard. He even said himself he spends like 3/4 of his time maintaining that thing, fighting off daily DDoS attacks (which obviously also only happen due to the public leaderboard) and trying to ensure that is smooth and responsive at midnight.
Just make it so ranks are still tracked and shown, but only private to the individual user. Done. Speeding would still be fun but not a single LLM faggot would remain.
>>
File: carbon.png (1.01 MB, 1730x4380)
1.01 MB
1.01 MB PNG
>>103471287 singlethreaded rust, regular input is bfs and bigboy is dfs
day 10                  time:   [88.887 µs 89.142 µs 89.416 µs]
Bigboy/day 10 bigboy time: [582.59 ms 583.24 ms 583.92 ms]

tldr: bigboy doesn't have enough trailheads connected to summits, search mostly finishes very early.

I tried naive dfs and optimised bfs and dfs is ~40% slower for the regular input, while bfs is ~40% slower for bigboy... I checked how many iterations it takes on average to find all paths from each trailhead for the regular and bigboy input:
Official
dfs iterations: 8341
dfs iterations/start: 29.164335664335663
bfs iterations: 5200
bfs iterations/start: 18.181818181818183
cached dfs iterations: 5200
cached dfs iterations/start: 18.181818181818183

Bigboy
dfs iterations: 35478622
dfs iterations/start: 4.61584563742061
bfs iterations: 32905037
bfs iterations/start: 4.281016649564736
./10.py 116.28s user 2.06s system 97% cpu 2:01.10 total

Since bigboy paths barely ever reach the summit (0.056 summits per trailhead) it' seems it's faster to do naive dfs since it's explores all those invalid paths faster. Regular input on the other hand has 2.5 summits per trailhead and almost 6 unique paths on average.
>>
>>103471782
I checked his GitHub time zone. If he's around 18-25 and live in the same country as me it's currently exam season for him.
>>
File: GZqObBiWMAIGZ5A.jpg (183 KB, 1170x1161)
183 KB
183 KB JPG
>>103471318
I wish I could bruteforce input-chan.
I wish bigboy-chan would bruteforce me.
>>
>>103471278
idiomatic calendar or images calendar?
if latter, I'm still here.
I'll update the calendar for days 9 and 10 tonight, I'm just waiting for something for some day 10 puzzle OC.
For day 9 I'll use either the fragmentation visualization from previous OP or the drawfriend art.
Or you can quote me the images you wish to be added.
>>
>>103472491
It's exam season for me and I'm not filtered (from aoc). Now filtered from academia is another story...
>>
If it's Eric's last year, could we ask LLMs to create next year's AoC, so other LLMs can top the leaderboards?
>>
>>103472510
get off aoc and pick up a textbook nigga
>>
>>103472509
>drawfriend art.
Deserves its own calendar.
>>
>>103472536
you said himself at the beginning that it's his intention
>>
>>103472529
Santa needs my help to save Christmas. I will be fine I'm just memeing
>>
File: aoc-2017.png (3.35 MB, 3282x2475)
3.35 MB
3.35 MB PNG
>>103472509
god bless you calendaranon
do we have images for 2015 or 2016, or updated ones for 2022 or 2023 btw? i appear to be missing those.
>>
>>103472383
I use GCC. They all seem to support it except MinGW from what I've encountered.
>>
File: file.png (3 KB, 216x35)
3 KB
3 KB PNG
Is he cheating?
>>
>>103472594
deboonked here: >>103472160
>>
>>103471717
check his twitter follows he actually might be a tranny
>>
>>103472594
probably. check his times.
>>
Are there any women (female) who do aoc past day 3?
>>
>>103472594
>I'm Kaden Bilyeu (Bikatr7), a junior majoring in Computer Science with a minor in Japanese at UCCS.
>My work primarily focuses on Japanese-related projects, exploring applications in NER/NLP, AI/ML, data science, and large language models (LLM) within software for real world applications.
Yes, he's cheating.
>>
>>103472645
I’m a real woman since last August when I had my bottom surgery.
I code in Rust btw
>>
>>103472680
a angel should not cut off her wings
>>
>>103472680
london?
>>
>>103472645
My passport says that I am female.
>>
>>103472509
Your images calendar. Idfc about retards and their stupid code calendars
>>
>>103472383
They all support <print> now, but libstdc++ (GCC) requires you to link with -lstdc++exp on Windows
>>
>>103468575
Anon.. You did not solve it
>>
>>103472657
also see his blog post: https://kadenbilyeu.com/blog/a3e009b8-aca5-4686-a4c5-ca016d6d0227
>i simply guess the problem without reading it haha
>i know it looks like i'm cheating, but i'm simply psychoanalysing Eric to predict each problem
>i also use short variable names btw
>>
>>103472645
Nora_Tries on Twitch did up to day 6 but couldn't solve it and looks like she gave up after that. I don't think there are too many others.
>>
>>103472491
SG?
>>
>>103472740
He should livestream one of his solves if he is doing them legit.
>>
File: carlos.jpg (35 KB, 600x600)
35 KB
35 KB JPG
>>103472742
at least she tried
>>
>>103472558
>2015 or 2016
/aocg/ started in 2017, no earlier ones exist
>>
>>103471736
His times are perfectly reasonable. Assuming he rage quit because of a single "llm filter day" which didn't even fucking work is going a bit far.
Some of us have irl commitments.
>>
TED
>>
>>103472763
someone on twatter asked him to record his solves or immediately commit his code and he refused
he said people wouldn't believe him anyway, so he's definitely fucking cheating
>>
>>103468575
>that part 1
kekd
>>
Is there even any big brain optimization to be done with today? Isn't this just DFS? And part 2 is literally simpler than part 1 because there's *less* state to keep track of?
>>
>>103472884
you can memoise part 2
>>
AoC has fallen
Billions must get filtered
>>
>>103472915
this........
>>
>>103472742
sigh, such wasted potential
>>
>>103472558
this one is marvelous
>>
>code up a solution for part 1
>fails on example
>notice I misread the instructions
>modify code, passes for input
>read part 2
>it's what I initially did for part 1
I can't be the only one who did this?
>>
>>103472820
what does that mean?
>>
>>103472952
pre-covid memes were just marvelous.
>>
>>103472967
Total Elf Death
>>
>>103472737
Am I missing something? His solution looks fine.
>>
>>103472895
Trooooo.
>>103472959
Same, along with probably >>103471715
>>
>>103472090
Chat gpt solved day 9 no problemo after like 2 prompts
>>
>>103472801
how about 2022 and 2023 calendars then anyone got 'em?
>>
File: calendar_small.jpg (3.99 MB, 7907x5941)
3.99 MB
3.99 MB JPG
>>103473035
sure thing
>>
File: 1670705172947.jpg (2.03 MB, 3000x1098)
2.03 MB
2.03 MB JPG
>>103472558
2022 and 2023 were unfinished, either through the calendaranon getting filtered or a lack of decent OC. These are the most complete versions of those two calendars I have.
>>
>>103473035
Just wait the rustfag will reply with his
>>
File: 1703416020864764.png (3.29 MB, 5306x3443)
3.29 MB
3.29 MB PNG
>>103473046
>>
>>103473045
>>103473059
kek like clockwork
>>
>>103472285
the later days usually even take ~1 hour for the top global leaderboard, so everyone will take a while eventually.
>>
>>103472285
Usually I don't, but today I wasted hours on the first part, making sure no path would ever be traveled twice, adding branches as I go, noticing I need to know which branch leads to which 9s.
Then I finally gave up and thought, fuck it, lets just travel through all possible paths and at least get the first star. And then the second star requires exactly that, taking me under a minute.
Fuck, I hate myself so much right now. Worst thing is I'm not even unemployed...
>>
>>103472077
is the last year thing actually confirmed?
>>
>>103473124
Yes, he posted a farewell post in the AoC++ forum.
>>
>>103472959
I did exactly the same.
>>
>>103472285
I usually get them done in 30 min~1 hr but I've been clocking in over 2 hours the last two days. Yesterday was especially bad for me
>>
>>103473016
It's missing something and he's mocking us with the variable 'g', because /g/ solved part 2 accidentally first.
>>
>part 2 is every path
>part 1 you have to merge paths that meet, but only if they're from the same origin

Once again 2 is objectively easier.
>>
>>103473133
>AoC++ forum
based paypig
>>
>>103471321
Nice, should've thought of something like that. But I didn't.
>>
2015 day 20 filters me too. I made a brute force solution, but it run for 253s (both parts) and I have no idea how to do it in a legit way
>>
>>103471287
>7z

I ain't unzipping that shit nigger
>>
>>103473209
do less calculations, that helps with bringing the runtime down.
>>
>>103473227
So there is no clever trick, just optimized brute force?
>>
>>103473211
bsdtar supports 7zip archives, btw.
>>
>>103473181
part 1 is counting the number of reachable nodes part 2 is counting the number of paths.

Ultimately the same bottom-up (postorder DFS) computation except that when joining you either join a set or you sum the number of paths.
>>
>>103473231
there is
>>
>>103473211
>dimwits who read "7-zip vulnerability" and didn't read all the way to "in the zstandard implementation"
>>
>>103473258
>dumb retard who only reads the first CVE and ignores the 1000+ that follow
>>
>>103472814
>anyone who didn't do day 9 immediately for whatever reason is cheating
retards
>>
so the ascii art is a big "10"
>>
File: file.png (1.49 MB, 1448x1070)
1.49 MB
1.49 MB PNG
>>103473063
Day 12 filtered me HARD last year
>>
>>103473280
Old news
>>
>>103473209
to be fair, day 20's part1 is ball rocking hard if you don't know.
part 1 is sigma function.
>>
>>103473231
I haven't done 2015 yet, so I don't know.
>>
>>103473236
What about BFS? Part 2 you just give it the head nodes and map/flatten x9. Part 1 needs a 2D structure of some kind. Neither are exactly difficult but clearly 1 has an extra layer of complexity.
>>
>>103473209
>he doesn't have a phd in discrete maths

Why are you even attempting aoc?
>>
>>103473338
show me the map/flatten x9 trick please. sounds interesting
>>
>>103473346
> Days 1-18
> Easiest shit ever
> Day 19 part 2
> If your input is bad you either try to solve it randomly or you are fucked
> Day 20
> I need to know some fucking sigma gigachad function
???
>>
>>103473046
I made the original shitty drawing of day8. Still mad about that
>>
>>103473364
Day 19 was literally nightmare mode.
>>
So did ANYONE solve Silver FIRST, and then have to figure out how to retool their solution to get an answer for Gold?
>>
>>103473429
No, you are the only one who did that.
>>
>>103473429
yes, I can read
>>
>>103473355
(0..9)
.fold(heads, |acc, _| acc.into_iter()
.flat_map(|n| graph.entry(n))
.collect())
.len();


Pretty self-explanatory. You start with the head nodes, map each one to its adjacent nodes, flatten so you get another 1D vector. Repeat x9 and you've counted every path.

I tried to do similar for part 1 but don't think it's possible, since you can't tell if two nodes count as distinct based purely on their values.
>>
>>103473429
Yes?
>retool
literally just remove my history map.
>>
if you fags are doing old ones, do intcode (2019)
>>
>>103473448
oh that is actually very neat. ty.
For part one, cant you just get the unique items in your resulting list? That would get all the unique 9s reachable.
>>
>>103473474
I will finish 2015 first
>>
can we get a cute reindeer for today's image? please i beg
>>
>>103473493
see
>>103468231
>>103468292
>>
>>103466519

When last have you guys seen oofmatoes?
Didn't he use lua?
hmmm.
>thinking
>>
>>103473583
>concerning
>>
>>103473338
all the same.
>>
>>103473474
that was my first year doing aoc, so many excellent puzzles that year. 2018 was also really good. the rest have not lived up to those two years, imo, with the exception of a few memorable puzzles.
then again i would not be able to make an aoc near as good as his so i can't really talk shit
>>
>>103473476
If that 9 is reached from two origins it should count twice. But if it's from one origin it should count once (even if that path splits and rejoins later). afaik you can't do it with a simple map because you lose the information on where the path originated.
>>
>>103473806
>If that 9 is reached from two origins it should count twice. But if it's from one origin it should count once

Ah fair. ty
>>
>>103471486
brootgod right there holy kek

>>103472257
>>103472419
seems I'm in on of them and not in the other
are there multiple filteranons using different logic or something?
started on day 3, but I'm up to date by now, is that why?

>>103472316
graychad again, you made me chuckle yesterday when replying to the anon commenting on it, thanks and good to see you still going
>>
>>103473209
each house is visited by its factors
>>
>>103471260
>tried to be clever and fucked myself up again
i never learn
>>
>>103472351
make the text a background image overlaid with transparent text that's slightly wrong
>>
>>103473791
that's not the reason you aren't allowed to talk shit. the reason is because it's free. had you paid for participation, you could talk as much shit as you please
>>
>>103473209
Oh, if I just create an array with all houses and just iterate over all elves if works very fast. The shameful part is it was proposed to me by claude.
>>
>>103472509
>I'm just waiting for something for some day 10 puzzle OC.

An attempt was made.
>>
>>103473932
LLM-chads cant stop winning!
>>
>>103472482
rember to give yous to rustfags so they don't go 41% and make their parents sad in christmas
>>
>>103473231
Think about it,

8 -> {1 2 4 8} -> 15 -> 150
9 -> {1 3 9} -> 13 -> 130

In general
presents = 10 * sum(unique_factors(house))
>>
I really like recursion
>>
>>103473965
I did exactly that and it took forever. Dumb solution works much better >>103473932
>>
>>103473977
>he can't efficiently generate factors

Summer in december?
>>
>>103473829
>are there multiple filteranons using different logic or something?
yes, we are 2 this year. I'm >>103472257 and I get the stats from both leaderboards and merge them. Merge all the json stats I got from this year in fact, so that kicked out anons still appear.
I don't know exactly what the other anon do in terms of merging the leaderboard stats but he run the script with parameters that makes you filtered at the first problem that you failed to solve, or at first problem that took you >24h. I don't filter on a time limit, only on the first time you failed.
>started on day 3, but I'm up to date by now, is that why?
Verify that you have not been kicked out of the leaderboard(s) you're in and that you are not in an early filter in >>103472419.
>>
>>103474033
>I don't filter on a time limit
you should though
>>
why doesn't slapping @lru_cache or @cache on my function work in python?
>>
>>103474033
>>103474048
isn't the 24hr a time limit?
>>
>>103473281
Tbh this filtered me, but it's because it was the first time I experienced a dynamic programming problem. I'm much better at them now, and I think I'll do well if there's one this year (which there probably will be).
>>
>>103474124
post your code
>>
>>103474124
Can't see how caching would help when the inputs barely repeat
>>
>>103474027
I am checking prime factors and then I am finding all possible combinations. How do I do it in a more efficient manner?
>>
>>103474203
m8, my unoptimized solution takes 2.7s.
>>
File: bfs.png (2.15 MB, 1024x1024)
2.15 MB
2.15 MB PNG
>mfw its a bread-first search problem
>>
>>103474134
not for getting the stars on the website
>>103474048
I don't like that. Anons should spend their time how and when they want. The only constraint that matters is if they solved it or not. That said, the plot should be frozen the 26 at 00:00 EST I think.
>>
>>103474245
kek
>>
>>103474174

>>103474199
doesn't it stop you from going wrong paths tho
>>
>>103474250
or maybe from the 27 idk. Most people have better things to do the 24 and 25.
>>
>>103474270
also ~5.3s on bigboy with pypy310
>>
>>103474270
yeah, it's completely fucked
better go back to the drawing board
>>
File: 1731592725785181.webm (654 KB, 500x281)
654 KB
654 KB WEBM
>>103474293
>yeah, it's completely fucked
explain why
>>
File: carbon.png (137 KB, 1024x1028)
137 KB
137 KB PNG
>>103474124
Do you clear the cache between each call? Multiple starts can path through the same block in different ways.
>>103474199
It's very slightly better on the regular input and like 10% worse for bigboy. Both are extremely fast though, it's just a pretty simple problem.
>>
>>103474310
I didn't read it. I just saw red, yellow, and green text and noped the fuck out of there.
Your shit looks like a christmas tree, not code.
>>
File: Day-10.png (1.48 MB, 3260x5220)
1.48 MB
1.48 MB PNG
Gleam solution for today!


>>103449446
Yeah don’t know what they’re doing, I felt like they were at a pretty good place before, but the changes amount to just running gleam add what they said it is replaced by.
I’m just here to enjoy the syntax, and I won’t check the discord.

>>103449497
Gleam target the BEAM VM and JavaScript… so yeah BEAM is going to be the fastest out of the two.
I feel like Gleam is pretty slow, without reaching for the powerful BEAM multi-threading primitives
>>
I found out that @cache made my bigboy 4x slower. Probably the cost of updating the dictionary. I think making a grid for memo is better after all
>>
File: aoc2024_g_calendar.png (3.52 MB, 3970x2481)
3.52 MB
3.52 MB PNG
I'm tempted to include >>103461896 screencap for day 9
Anyway I'm kind of busy this week at work and can't update the calendar during the day so give me some time
>>
>>103474356
you have permission to use >>103471266 as todays' image
>>
>>103474319
>Do you clear the cache between each call? Multiple starts can path through the same block in different ways.
nope, cause the next value I'm looking for. Running on pypy3 blowup the process memory to 7GBs heh
anyways it's not critical here. But I feel like I tried it on knapsack like problem once and it was shit too. Maybe the overhead is too high
>>
>>103474363
brooting is affecting my speech oh my
>>
>>103474360
thanks
>>
>>103473885
Me too, anon, me too.
>>
File: file.png (35 KB, 1043x848)
35 KB
35 KB PNG
Lava island hiking with a reindeer
>>
>>103474360
dude shut up
>>
>>103474328
you're my first intro to gleam following your solutions but the gleam use expressions/do notation/binds are very sexy and makes me want to pick it up, it looks like such a fun lang
>>
>>103474389
Cute!
>>
File: file.png (334 KB, 1686x1912)
334 KB
334 KB PNG
import solution in Rust
Wanted a change of pace by trying out petgraph but wasted a bunch of time learning it. Would have been better off doing dfs by hand even though I was exhausted last night
>>
File: 1725168021368878.png (751 KB, 1476x6548)
751 KB
751 KB PNG
>>103471287
Cleaned up and optimized a bit
BFS each point. Use a 19x19 byte array to store whether a point was seen in the high bit, and the number of ways to reach that point in the low 7 bits. Use a simple fixed size circular buffer as a queue, since queue size can never exceed 36. Parallelize over the points.
Benchmark 1: ./a.out bigboy.txt
Time (mean ± σ): 41.9 ms ± 1.1 ms [User: 765.1 ms, System: 7.8 ms]
Range (min … max): 39.8 ms … 46.1 ms 74 runs
>>
>>103474466
>fixed size arrays
>parameters specifically tuned for the input
you didn't solve it
>>
reject golf

embrace readable code and big O
>>
>>103474482
Every solution hardcodes '0' and '9'
The array sizes are just functions of those numbers
>>
If your code fails on an irregularly-shaped input you didn't solve it
>>
File: IMG_6537.jpg (1.24 MB, 1170x1464)
1.24 MB
1.24 MB JPG
>>103474328
And then one day for no reason at all the language went to shit
>>
>>103474505
That's part of the puzzle description
>this means that a hiking trail is any path that starts at height 0, ends at height 9
Does the description also say that queue size can never exceed 36?
>>
>>103474524
The queue size can never exceed 4*('9'-'0')
>>
>>103474532
And that's where the 19x19 comes from as well. For any point '0', set it as the center of a (('9'-'0')*2+1)*(('9'-'0')*2+1) array. Every reachable point must be in that array.
>>
File: 1733787334511368.png (15 KB, 771x604)
15 KB
15 KB PNG
my first year doing this, and i have a question:
how do you guys get the time to do this shit while also having to do your dayjob and live life? do you just have jobs with 0 workload for december?
>>
>>103474532
why bother replying to the filtered?
>>
File: 1733438195159205.jpg (13 KB, 285x279)
13 KB
13 KB JPG
>>103474549
>your dayjob and live life
>>
>>103474549
List out every single thing you do from waking up to going to bed and I can assure you're wasting loads of time doing things of no value in your life that you can replace with this
>>
>>103474549
Live life?
>>
>>103474549
unemployed+no life
>>
>>103474532
>>103474546
ah, you are right. very nice anon.
>>
>>103474549
>while also having to do your dayjob and live life?
I assumed everyone was living in their parents' basement like I do.
>>
>>103474549
Everyone on /g/ is so irreplacable they can slack off work at will
>>
>>103474549
new aoc day starts at 6:00
I start it at 18:00 when I get back
simple as
>>
>>103474549
What do you mean? All my times are sub 1 hour. I do spend 2-3 hours on the 1-3 hard problems each year. But it seems like a skill issue if these take all day for you.
>>
>>103474549
My dude, most problems take 20 minutes. It's not hard to carve out that time.
>>
>>103474549
no job and on the verge of commiting sudoku
>>
>>103474549
My job starts two hours after the puzzle is made available.
If I don't finish before I need to go to work, I resume the puzzle in the evening.
I do not have, nor do I have any hopes of getting, a girlfriend.
>>
File: aocday10termux.jpg (594 KB, 1080x2400)
594 KB
594 KB JPG
>>103474549
half the time I do these on my phone while I'm riding a train or something. I use Clojure in part because using a more verbose language like this would not be viable.
>>
>>103474389
the soul of /aocg/
>>
>>103474389
cute!
>>
>>103474607
>20 minutes
Yeah if you want shit out some generic unreadable broot pyslop to place on a leaderboard nobody cares about.
Some people enjoy crafting finer, more bespoke solutions
>>
>>103474565
Why are filterlets so quick to call aoc a waste of time BY POSTING ABOUT IT ON 4CHAN. Do they not see the irony here.
>>
>>103474389
You are literally carrying /aocg/ on your back. I kneel. I hope you aren't filtered
>>
File: 1644787867230.jpg (31 KB, 500x380)
31 KB
31 KB JPG
>>103471260
>>103471416
Anon I have a couple of questions:
1. what WM theme is that?
2. what color scheme for vim are you using?
3. what is the background picture?
Thanks in advance
>>
>>103474565
wake up at 8:45
make mate, catch up on group chats
morning meeting at 9am
work on my tasks until 1pm
work out for an hour
back to work until 5pm
hit up friends to go do something, lately we've been playing through xbox 360 games split screen
>loads of time doing things of no value
no offense, but the kind of people who say this kind of shit lead dreary lives that i don't really consider "life" in the slightest. NEETs are more repsectable than rigid schedule time optimizing autists.
also don't kid yourself, you're solving programming puzzles. this is entertainment as much as anything else is. it's a fun way to get your 10,000 lines in for a new language, but i sure hope you don't think you're doing valuable things here that you couldn't do elsewhere by actively building something useful to somebody else.
>>
>>103474684
why are you seething about others enjoying something you don't enjoy?
>>
>>103474684
you really think the people that post on /g/ have a schedule like this?
>>
>>103474684
>but i sure hope you don't think you're doing valuable things here that you couldn't do elsewhere by actively building something useful to somebody else.
Weird, I don't see that on your schedule...
>>
>>103474684
>there's just no time for AoC
>6 hours of video games

Yeah, that's a tough situation.
>>
>>103474684
don't yuck my yum.
>>
>>103474684
>work on my tasks until 1pm
>back to work until 5pm
kek imagine being this cucked
>>
>>103474684
>Aw man I wish I had time for AOC how do you guys do it
>Admits to wasting hours playing 360 games every day and feels superior about it
>Don't kid yourself you're just solving programming puzzles I don't care anyway

I sense a lot of confusion and cognitive dissonance in your life
>>
>>103474641
That's hardcore, man
>>
>>103474698
what are you even talking about? what seethe?

>>103474710
>Weird, I don't see that on your schedule...
what do you think a "job" is?

>>103474700
i would assume more than you might think. anons say they don't have friends, and maybe they don't hang out with friends a lot irl, but they probably at least socialize online. it's all valid.

>>103474727
i'm not yucking anybody's yum, except for dumbasses who think blowing off friends for sudoku puzzles is actually a good thing, like:
>>103474721
>>
>>103474779
>what seethe?
>>
>>103474684
>hit up friends to go do something, lately we've been playing through xbox 360 games split screen
So from 5pm-1am you're playing video games?

>the kind of people who say this kind of shit lead dreary lives that i don't really consider "life" in the slightest
Just lmao
>>
>>103474641
actually kind of interesting. maybe i'll pick up a lisp next year and attempt it using that. a 30-line solution is really good.
>>
>>103474549
I find it tough. The problems release early morning for me and I have a demanding job from 8 to 5. I share the chores with my wife and we need to walk the dog in the evenings, etc. I try to do as much of the problem as I can during lunchtime, otherwise I try to get to it in the evenings. Like I told someone in a previous thread, I would feel that I failed if I got some LLM to write code for me. If I cannot solve it with Python and what I learned in an algorithms class / textbook from years ago, then that's where I'll stop. That's probably within the next few days anyway.
>>
>>103474684
>>103474779
This is peak normalfaggotry. People like you see spending most of your free time "socializing" as healthy and a good use of time, when it reality you're just wasting time in a group. This is why normie dei/"networking" hires don't develop useful skills and eternally resent the accomplished programmer.
>>
>>103474641
>phonecoding
I have thought about this, but setting up the workflow seems annoying. What software do you use, is it just termux or what.
>>
File: Billdoom.png (89 KB, 349x238)
89 KB
89 KB PNG
>>103474549
married with 3 kids under 10, full-time employee
it only gets bad towards the later days for me if i get a fucked up one like the cube or the sea monsters and then i just don't do my work that day
>>
>>103474872
>the cube
Making a paper cube, and opening/aocg/ and seeing 20 other paper cubes, is a top 5 /aocg/ moment for sure.
>>
>>103474779
I know it hurts to admit you are filtered. But only by accepting it can you improve.
>>
>>103474684
>work
>come home
>play hours and hours of videogames before bed
"how do you guys get the time to do this shit"
Please be bait
>>
>>103474803
>playing video games?
we do whatever we do. sometimes its video games, when its nice out sometimes we drive somewhere and go for a hike. sometimes we watch movies. sometimes we got a bar or watch a live performance. sometimes it's something else. if you played video games every day, don't you think that'd get boring? variety is the spice of life.
>5pm-1am
no, we're spending an hour or two after work getting everybody together, getting shit to eat, doing whatever it is we end up doing that day, then home 9-10, doing some chores, take a shower and bed by midnight. sometimes we don't hang out, sometimes i have more chores to do. no two days are truly alike, and that's proper living.

>>103474823
well, you made it farther than me lol

>>103474872
based dad. how often do you usually spend on a problem?
>>
>>103474884
i know i made at least 2 of them that day. that was a bad day, i was an offset hardc00der.
>>
>>103474549
I was working 12 hr shifts as a labourslave and didn't have time. Now working 8 hours as a wageslave I have all the time in the world. It's all about perspective.
>>
>>103474847
I did that before getting a laptop. Yeah it's Termux with Neovim
>>
File: aoc2023day12goldSmall.png (82 KB, 833x621)
82 KB
82 KB PNG
>>103473281
me too, I didn't even understand the question at the time.
I looked at it again a month ago and thought a bit about how regex work under the hood, since I used a regex to parse the input. After I had that insight, it became pretty easy to solve.
>>
>>103474908
so far i've been getting them done quick, but in previous years i either finished it in under an hour or it took like 6 hours with numerous false-starts. the fucking sea monster one took me all day.
>>
>>103474033
>Verify that you have not been kicked out of the leaderboard(s) you're in and that you are not in an early filter
indeed, got autofiltered cause I started late
guess doing day 1 late matters more than day 10 on time for that one, thanks for having a different approach

>>103474250
>The only constraint that matters is if they solved it or not. That said, the plot should be frozen the 26 at 00:00 EST I think.
yeah that seems most reasonable desu
>>
>>103474964
I started a day late and I'm still on the leaderboard though
>>
>>103474680
sure bro it's xfce with https://github.com/neeeeow/Bluecurve
vim theme is like desertnight or something
>>
>>103474779
>not having a comfy WFH job and able to do AoC whenever you want
ngmi
>>
>>103474908
>no two days are truly alike, and that's proper living.

Great. So use your surface level understanding of programming to solve the first 3 days, get filtered on day 4, and then move on to something else like going to a comedy club or walking next to a lake. What's the problem?
>>
File: 2024-12-10_20-24.png (49 KB, 662x325)
49 KB
49 KB PNG
>>103474983
then idk why I am not? maybe filteranon started his collecting late? I did start on day 3
>>
>>103475010
Ah, started on day two myself, which probably explains it.
>>
File: aoc2-1asm.png (245 KB, 1080x1920)
245 KB
245 KB PNG
>>103474847
>What software do you use, is it just termux or what.
Termux, yes. There are also UserLAnd and nix-on-droid which offer more software but come with caveats.

In that screenshot I'm using Doom Emacs via ssh.
Locally on the phone I have Practicalli's fork of AstroNvim with parinfer and Clojure support, as well as a more minimal emacs config with Kakoune-style keybindings (meow).

I started out doing this with just nano, but given how much of a pain it is to type on the phone keyboard it was a great motivation to teach myself how to use vim keybindings properly.
>>
File: 1630505898865.jpg (15 KB, 256x245)
15 KB
15 KB JPG
>>103474998
Thanks
>>
>>103474142
Shit, I'm still pretty terrible at it I think. Maybe I'll look into trying to solve it again
>>
>>103474549
the first couple weeks take <1 hour so you just do them before bed, after that you're on christmas vacation and have all the time in the world
>>
>>103475051
I'm also trans btw
>>
>>103474670
I'll try to go as far as I can with these. I should be extra free come Sunday so getting filtered shouldn't be a problem (barring we don't get another ridiculous vector calculus or curve fitting problem)
>>
>>103474908
That's behavior you should have grown out of in high school or college age unless you're an npc with no hobbies. Hanging out with friends is great but most of that time is wasted doing nothing and you don't need to do it every day. You should already not be doing that with having girlfriends if you're an actual normie

You can easily fit little puzzles into your day in the morning instead of retarded group chats or after work corralling retards to hang out but this is hopefully just bait
>>
>>103474203
>to efficiently find factors in a range, you need a sieve
>to make a sieve, you make an array and go through crossing out multiples
>the elves are literally making their own sieve in the question
>>
>>103474466
Do you actually need the visited bit? I'm just checking if seen_ways are non-zero, then adding previous cur_ways.
I had a similar idea here >>103472482 and it turned out slower than dfs for bigboy. Still very good for the regular input.
>>
File: file.png (153 KB, 715x1268)
153 KB
153 KB PNG
ugly shit too late at night fuck this and see you tomorrow
>>
File: 1718684146757731.png (53 KB, 975x534)
53 KB
53 KB PNG
golfed it down a bit
bfs would probably be more tacit but im not sure how I would do it
>>
>>103474549
I usually start them when they go up, which is 11:00pm my time. If I'm not done by 2:00am I go to bed and either work on them after work the next day or on the clock if it's a long one and/or I think I can get away with it.
>>
>read problem first thing in the morning
>think about it all day writing down pseudocode as it comes to me
>get home from work
>translate pseudocode
>works first time
it's a good feel
>>
>>103475480
>define symbols that do complex things
>hmmmm how can I golf this
Just take a few of those symbols and make a new symbol that does that. Ez
>nooooo only certain symbols count because..... they just do ok!!!!
Ok.
>>
File: 1490321039604.jpg (89 KB, 400x400)
89 KB
89 KB JPG
>>103475613
>read the problem
>misunderstand it and write a bad solution
>fix and move to part 2
>bad solution is now the good one
>ctrl+z
>"This is the correct answer"
It pays off being a low IQ brainlet.
>>
jesus. 2021 was annoying as hell.
>>
>>103475480
this is like compiling C code and calling it "golfed" IMO. handwriting stack machine bytecode is not golfing
>>
>>103475480
tell the truth. You rewrote after you noticed that your last version didn't have any LGBT flags.
>>103475684
>>103475761
golfing in most languages is some pointless fun. Golfing in languages like this is more like washing your code, is similar to switching to more appropriate data structures or removing gross inefficiencies. If the primitives are well-chosen then the number of glyphs tells you something about how complex the task actually is. It's not an exercise in compression. see https://mlochbaum.github.io/BQN/commentary/primitive.html
>>
I've been adding more optimizations to my GBA day 7.
The main concern is whether I can get it to a solution in the life of the batteries, though switching to non-rechargable may help.
>>
>>103475884
>tell the truth. You rewrote after you noticed that your last version didn't have any LGBT flags.
For what it's worth that guy basically just admitted that outright once iirc. I guess it's the same reason why the idiomatic rust guy keeps posting, it's virtually guaranteed replies every single time
>>
>>103471991
big oof
i don't even feel proud being a better programmer than he is desu
>>
>>103475884
Honestlt I completely agree with you, I was just making le funny 4chan post
>>
>>103476206
That guy has literally nothing to be proud of in the first place. He did a glorified speedtest to see how quickly he can download the input and stuff it through an LLM. He learned nothing, he accomplished nothing, he proved nothing. He cheated not only the game, but himself.
>>
>>103475394
Oh, yeah, you're absolutely right, thanks!
It doesn't seem to change the speed but it does make the the code much cleaner. Maybe the compiler had figured out that bit didn't do anything already.
>>
>>103476316
smdh
even the filtered brootchads who eventually get the right answers are better
>>
File: 1733871734050.jpg (509 KB, 1280x1706)
509 KB
509 KB JPG
>>103475684
>>103475761
its what the other anon said, golfing in array languages is basically equal to optimizing in other languages, managed to cut down my runtime in half just by doing less work
plus its fun, it makes me think more about my code and give it a reread after my mind is clear which helps me learn the lang
>>103475884
>>103476082
I was mostly just joking I really dont care about muh flags cuz people would find reasons to complain about anything that isnt python or C no matter what
I do only post the codepad version for bait reasons but I dont really change my code to add fag flags
>>
File: 10_2.webm (663 KB, 344x782)
663 KB
663 KB WEBM
Part 2 BFS (because it looks cooler)
>>
>>103476548
nice. seeing them all at once with transparency blending for overlapping would be cool
>>
>>103476548
cool
>>
>>103476564
Yeah, but too much work with ANSI colors
>>
>>103476600
yeah i realized after i sent you were doing that in the term
>>
File: aoc10.png (13 KB, 641x270)
13 KB
13 KB PNG
My APL solution for Day 10. No traversal or search is needed, simply creates 10 boolean matrices for each number 0 to 9 where 1 indicates that number is present at that index. Then for each trailhead do a reduction which shifts each matrix left/right/up/down and sums it with the previous/accumulated matrix. Result of the reduction will be a matrix where each index of a peak that can be reached from the trailhead will be set to the number of unique paths that can reach it. The Part 1 answer is the count of those peaks, the Part 2 answer is the sum.
>>
>>103471991
I legitimately want to kill myself knowing people like this exist. This whole article is the most insane garbage I've read in a long time.
>>
>>103476666
cool stuff tacit bro
whats the runtime like on input? bigboy viable?
>>
>>103475480
>>103476696
esl tard, tacit does not mean whatever you think it means
>>
>>103476730
tacit = based
>>
>>103476684
It won't fit into the workspace memory. An unfortunate side effect of creating a wall of 7500x7500 matrices. If I can eliminate the trailhead start matrices it may have a shot though.
>>
>>103474549
Drops at 9pm for me so I just take an hour or two each night to so it
>>
>>103476872
nice aim
>>
>>103476880
canadajeet or amerifat?
>>
>assign each point an increasing ID in the range 0-360
>offset by 19 by rows so each point in a 19x19 chunk has a unique ID
>create an array of 384-bit vectors (bitsets) for each point
>index the vector by ID, set 1 bits corresponding to IDs for each 0
>for i in 1..9 set i-points to the bitwise or of adjacent (i-1) points
>print sum of bitcounts of 9-point vectors
This might not be fast because of the amount of data involved but it should work
>>
>>103476905
Not him, but as a leaf I'm offended you think our jeets are capable of participating. They're not.
>>
>>103474549
i do it at work, no big deal to take a 30 minute break to shart out a quick solution
>>
>>
File: 1583760053532.png (26 KB, 614x585)
26 KB
26 KB PNG
>know I know how to solve part
>brain is not being forthcoming
>>
>>103472645
Me! But I'm autistic, so probably not the best female rep.

>>103475155
>Dracula colorscheme
>emacs
>on an android
>trans
I'm shocked.
>>
>>103477086
You and >>103471318 should team up.
>>
>>103471287
BigBoy-chan may be too much woman for me to handle...
>>
File: aoc10.png (166 KB, 1280x1751)
166 KB
166 KB PNG
>>103477086
>decide to just brute force it
>what could go wrong?
>it works

thanks eric

>>103477106
ok you have some scary sense of premonition predicting my actions like that
I made my post without scrolling down reading the thread just to complain
>>
>>103477149
I had the exact same issue. Did you also accidentally solve part 2 first?
>>
>miserable in the wagie cagie today
>don't want to do my work
>have an idea: just brute force it
>brute force it instantly
Wtf is life really this easy?
>>
Am I the only one who solved both parts correctly for the first time?
Why can't so many people read in these threads?
>>
>>103477161
nope
part 1 was really easy fo rme
I have been awake for too long to think of how to solve part 2 properly even though I feel like I know I have done puzzles like that a bunch of times before
and in those cases bruteforec was infeasible and I did it properly
even earlier years on aoc fairly sure

I was thinking of doing it properly by
SPOILERS HERE
keeping an array where I count how many times a node has been visited or something like that and exit early incrementing when those nodes are reached in a recursive call

but the specific details of how to do that are not a today thing
>>
>>103477187
i think 1, 2, and 6 i got correct in 1 attempt for both
>>
>>103477211
*and i got today's too. that included getting the examples correct in one try but today i did not get the example correct on first go
>>
>>103477187
I solved part 2 before I even read part 2
>>
File: 1727998464614721.png (2.12 MB, 1920x1080)
2.12 MB
2.12 MB PNG
>>103472645
>remember wishing i was a woman back in 2015 when i did the aocs as a preteen
>10 years later and still not a woman
when will technology catch up..
>>
>>103477273
don't worry anon everyone loves your uiua solutions
>>
>>103477094
You know the rules, post 'em.
>>
>>103477273
Like several decades, anon.
First AI, then robotics, then genetics.
Just hold fast and don't go the pretend route, it doesn't end pretty.
>>
Alright I've given you all a nice headstart. It's about time I solve this
>>
>>103477273
ywnbaw
>>
File: xarbon.png (14 KB, 556x612)
14 KB
14 KB PNG
>>103477238
Yes because you can't read.
>>103473413
2015 day 19 part 2 was very easy, dplets should be banned from this board.
There are only 22431408 states for my input.
picrel is an extremely unoptimized version I wrote in 10 minutes and so it's slow but it's trivial to optimize with a tree graph (forest to be exact) whose children are single-point splits of the parent, the leaves are single characters.
You can reserve a tot-length vector and compute the key as ((2*len-l+1)*l/2+r-l)*node_count+node_index<tot.
Someone who has time to fuck around with C could do it.
>>
>>103477293
blue board, anon
>>
>>103477365
crosspost or gtfo.
>>
File: 1714552254381380.jpg (126 KB, 1200x889)
126 KB
126 KB JPG
>>103477293
>>103477371
This is why no one on 4chan admits to being a woman (female).
>>
>>103477273
I don't get this wanna be opposite sex thing.
Could you briefly explain why do you want to be a woman, I'm genuinely curious.
>>
>>103477391
Yikes!!!! We need a CoC here ASAP!
>>
>>103477391
LONDON
O
N
D
O
N
>>
>>103477394
NTA, but autogynephilia/autoandrophilia is really common in autism. You're so self-focused that you would just love to fuck a compatible version of you.

I'm XX and autistic and am turned on by the idea of being a dude. It just comes with the territory.
>>
>>103477394
men want to be women because men are repulsive

women want to be men because they were sexually abused and so are driven to make themselves repulsive
>>
File: grids.webm (2.17 MB, 720x720)
2.17 MB
2.17 MB WEBM
I don't want to spoil anything, but this is a hint for tonight's challenge.
>>
>work through intcodes
>get hit with arbitrary memory requirement right at the end
>have to rewrite from scratch because rust's retarded borrow restrictions

Fuck this stupid fucking language
>>
>>103477094
Which language?
>>
>>103477531
kek
>>
File: xarbon.png (15 KB, 556x660)
15 KB
15 KB PNG
>>103477499
I see. Thank god I'm not autistic.
>>103477521
lol
>>103477364
I'm sorry I fucked up the state count calculation, there are actually 32931216 for my input. But same magnitude.
Also if you do an assert that the new molecules are always at least as long than their origin then it's possible to prune when len(s)>(r-l+1).
It's still slow though, please someone do the C version, I have to go study now.
One more thing to the optimization, you could fill the base cases initially by searching for the corresponding substring of each vertex in the medicine (for example with the KMP algo or z-algo to be efficient, etc.) and setting those values to zero.
>>
>>103477585
coreutils/awk
>>
>>103477549
filtered
>>
>>103477499
actually i like MEN
>>103477521
also this
>>
File: 1636520364333.jpg (137 KB, 756x716)
137 KB
137 KB JPG
I think my favourite puzzle is that volcano one with the clay/obsidian robots from 22 with like ores and resources
day 19

or 2022-16 with the lava flow
2022 was a SOLID year
>>
File: london.png (9 KB, 410x184)
9 KB
9 KB PNG
>>103477452
in bqn this is just
>>103477678
might be rose-tinted glasses but all of 2017 was fun
I also liked the water-flowing puzzle with fill rules, overflow rules, and a vi-friendly input.
the more immediately visualized problem, the more entertaining.
>>
File: f# day 10 silver.png (38 KB, 873x684)
38 KB
38 KB PNG
I had to modify the function to return a list with the whole path step by step because I kept getting the wrong answer. This made the solution ridiculously messy
But I could debug it
I was calculating if a path was valid via abs (next - current) because I thought any slope of 1 unit was fine
>>
>>103471883
>geohot
>chink
bro?
>>
>>103473124
>is the last year thing actually confirmed?
the large amount of retreading our steps and reminiscing makes me goncerned
>>
>>103477549
What problems? The BC is based
>>
Make our own AoC with situations so violently racist anyone using an LLM will get their account permabanned
>>
>>103477678
that was a hard pass from me
>>
>>103477273
it hurts so much bros
>>
>>103477801
>you have redundant code you use for every op? just put it in a closure
>oh you're using hashmap instead of vec? no can do
>want to resize the vec? can't do that either
>want to pass to a function instead? gonna need a mutable ref and you can only have one of those
>all this despite only one instruction being triggered at a time making a clash impossible

It's retarded that I have to copy and paste the same routines into every instruction.
>>
File: ericpanic.webm (1.23 MB, 720x720)
1.23 MB
1.23 MB WEBM
Live reaction of Eric panicking as his servers cannot handle the load.
>>
>>103477846
Sounds like you're a retard trying to mutate a collection while iterating over it
>>
I completed 2015!

I still have no idea how to properly solve d19p2 without randomness and I was trying too hard to be smart on d20, it was faster to do it the dumb way. Otherwise pretty easy!
I guess I 'll do other years later (I joinded only last year and didn't have acsess to my machine for like 3 days after day 10, so I dropped it).
>>
>>103477853
Eric's actually a specialist at handling large loads
>>
>>103477862
Wrong. All the logic is in a loop, it's closures being called in a match statement.
>>
File: code.png (5 KB, 434x115)
5 KB
5 KB PNG
>this is all it took to get my solution for part 1 working on part 2
>>
>>103477802
go do it. Call it Advent of Floyd if you're really pissed at Eric, and something without "Advent of" if you calm down by then.
>>
>>103477867
see >>103477364 >>103477603
>>
>>103477892
Without code I couldn't say anything

The only time I remember the BC being annoying was before retain_mut was stable on Vec and one or two problems requiring a hashmap clone and swap for looping over it
>>
>>103477908
sometimes an unga is all it takes to bunga
>>
File: 1733883950893.jpg (78 KB, 750x1000)
78 KB
78 KB JPG
>>103477273
I made this face in real life reading this post.
>>
>>103477697
2017 - 2019 were s-tier for sure
>>
>>103477951
the funny thing is it would fail if downward paths were allowed, because i could've been reaching the same tile, from the same 0, from a different path and then it would've given me the wrong answer for part 1

It's my first year and I thought the problems were going to be a lot trickier. Hearing anons talk about intcode and virtual machines got me thinking it was going to be something else
>>
>>103477954
you'll take it if science ever catches up don't lie to me
>>
>>103477867
Gratz. Yeah 2015 had a lot of days that could have been interesting, but the inputs were basically test input size so there was no reason to not just do them the dumb way. Like imagine the cookie day with 100 different types of cookies and choosing 10000 of them
>t. Also just did 2015
>>
>>103477927
Some anons had a good idea to make a Hannukah challenge. It needs to be very difficult though, you cannot have the goy learning all of Shlomo's tricks.
>(Amounts under $1,000 barely impress me after fees and will not confer a Star of David. Please be patient; some transactions can take an hour or two to complete. Just trust me)
>>
My Day 7 GBA solution is at line 250 after about 2.5 hours.
I need more optimization strategies.
Silver alone solves in a few seconds.
>>
>>103477981
Imagine an X vs Y meme of your choice. Labeled "my misogyny" vs "my porn addiction"
[My misogyny wins btw]
>>
>>103478001
cgdct media opened a void I don't think will ever be filled
>>
this is the worst fucking board
>>
>>103478034
For 11 months out of the year, yes. But at least for one month we can have fun here
>>
NEW THREAD
>>103478070
>>103478070
>>103478070
>>
File: 6lhynydptl3e1.png (196 KB, 720x960)
196 KB
196 KB PNG
>>103477981
I mean yes. Many people would.
Because it is a fetish thing through and through.
But magic (or technology on that level) does not exist, so sane people refrain from treating a little bit of roleplay and self mutilation as "the same thing".
Well and there is also the more modern aspect where depressed, lonely men get convinced and groomed by trans people into believing that their unhappiness and their problems stem from being the "wrong gender". Taking the hormones feels good to them for a while not because of any gender problems, but because after they do it they get showered with attention and positive reinforcement from other lgbt people online while receiving a free pass to act like assholes again towards others. In other words they feel like they belong somewhere for the first time in their lives.

Meanwhile basically all of the terminally online white male transfem people are "lesbian" and become ultimately just a different flavor of incel.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.