advent-of-code

[magic oxide]

i have the urge to use a monad

it would be wholly unnecessary

but i would be using a monad

[left yew]

I mean what! 30 seconds!!!

how is this possible?

[magic oxide]

p1 was a fairly routine "simulate the grid" problem

possibly llm tho

[left yew]

that guy must be close to making actual jarvis 😭

but seriously none of the top part 1 guys came on the part 2 leaderboard

[magic oxide]

because p2 wasn't a fairly routine "simulate the grid" problem

[shrewd matrix]

it's kinda plausible that someone just had a p1 implementation lying around tbh

[magic oxide]

yeah

the only kinda weird thing was that the movement inputs were on multiple lines for some reason

[cyan sleet]

@magic oxide we're getting more $

[magic oxide]

it's joever

tho https://docs.raku.org/type/Complex exists

maybe it's not joever

[hollow belfry]

what is AoC Gate Guardian role?

[peak dock]

they just create the spoilers channel for each day when the global lb is filled

i finally solved 2023 d21 p2 😭

my god

idk what the expected solution is but mine is terrible

huh okay the solutions i see on reddit are almost equally terrible

[cyan sleet]

basically if you ||get the number for 1, 2, 3, ... blocks out there is a closed formula that describes it||

I did a ||difference table|| approach for fun to get the ||coefficients|| of the ||polynomial||

but yeah, the big thing that made that feasible are the ||"highways" going vertically and horizontally||

oh I remember why I did my approach

it was in reference to day 9 that year

as in "hey we can actually use that thing for this task"

https://github.com/algmyr/aoc-2023/blob/main/day21/Main.java#L68-L69

[marsh current]

day21/Main.java lines 68 to 69

// This is inefficient, but let's do a difference table
// extrapolation like in day 9 for fun.```

[magic oxide]

oh that day

i used ||gregory newton for the interpolation i believe||

[peak dock]

|| my approach was on a per-position basis. for each position, i bruteforced the min dists to that position in the shifted copies upto like 10 layers of copies. for the innermost 2-3 layers (round up to 5) its unpredictable but outwards from that the min dists to that point increase in multiples of the grid size, so theres closed form expressions to cover the next 202295 layers ||

[cyan sleet]

||sample at 131*i + 65 steps and it's a nice polynomial||

[peak dock]

||idk i dont trust interpolation for integer stuff||

[cyan sleet]

why?

the way I ||extrapolated|| it's guaranteed to ||be integers anyway||

[peak dock]

hrm

i guess i did kinda think of something like that but discarded it

[cyan sleet]

I guess you're familiar with difference tables from day 9, stuff like this

 1  4  9 16 25 36
   3  5  7  9 11
     2  2  2  2
       0  0  0

you reach a constant/zero row and you can then continue filling things in

extrapolating the sequence

this reaching a row of zeroes happens for exactly sequences that are polynomials

there are other fun schemes, like ones to extract fibonacci-like recurrences, which is much more powerful than this

e.g. the berlekamp-massey algorithm

[lime fulcrum]

for me day 6 9 and 15 get an extra file for part 2

[cyan sleet]

part 2 today is just rude

[strange zinc]

looked at it and am just ignoring him now.

[cyan sleet]

I had fun doing part 1 with only regex

[strange zinc]

excuse what

[cyan sleet]

well, regex and transposing the grid

but close enough

[magic oxide]

every single language from now on better have a hashmap and the complex type

[cyan sleet]

fun fact, raku has its own regex syntax

[strange zinc]

using vlang and they dont have tuples... i wasnt too happy, especially cuz I like keeping coords in tuples for hashmap keys

[cyan sleet]

ugh, now I have to make a proper solution

I can't regex my way out of this I think

[rigid fulcrum]

are there aoc solutions posted anywhere?

[unique osprey]

#1047673173447020564

[zinc kindle]

1266/16657. Another day where I decided to sleep on it and come back in the morning. But I still solved it eventually.

[unique osprey]

I was half expecting the bot to push boxes into a Christmas tree today

[rancid creek]

The tree's are in the boxes. 🙂

[scarlet locust]

today is the most difficult part 2 for me so far... pray for me 🙁

[warm gazelle]

woah what is this character

[shrewd matrix]

.

w/ some text shadow

[scarlet locust]

omg it's a shooting star

make a wish

[scarlet locust]

i don't know how, but i somehow managed to write working code. only problem is it takes 10 minutes to compute LMFAO

i don't know how i managed that

[humble copper]

was so stuck on part 2 for today’s challenges because my robot got inside a box and forcefully ripped it in half

[left yew]

Inspect element

[mossy basin]

might do today in nim, why not

[shut trellis]

<@&518565788744024082> Good morning! Day 16 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/16. Good luck!

[long egret]

bro

if there's one more grid based puzzle im gonna lose it

[hexed axle]

You're in a hexagon tiling space with a quandary at hand

[devout tusk]

WHAT ARE THESE EDGE CASES

[hexed axle]

https://discord.com/channels/267624335836053506/1318083671781019670

[magic oxide]

i kinda don't feel like doing this rn

would rather eat food

[pine tiger]

wait what edge cases are there

alright not too bad

esp given that i'm on a plane

[hexed axle]

Smart. Breaks help a whole lot. Learning to embrace that helped me a ton with projects where I got mentally stuck

[magic oxide]

yeah i spent like a huge chunk of time yesterday on day 12 purescript

and brain's kinda refusing to work

[devout tusk]

nvm, I just suck at grid based problems 😭
876/400 :/

[pine tiger]

we'll get em next time

[warm gazelle]

i think i may have written one of the most inefficient search algorithms of all time for day 16 part 1

[woven oriole]

erm

||my p2 does not maintain multiple ancestors how||

[warm gazelle]

it's probably one of your ||pruning parameters in part1||

[karmic coral]

i dont understand ppl participating in aoc. if you don't want people utilizing part of your solution then dont share them at all literally anywhere on the internet

[tall quest]

Part of showing off, and just sharing different solutions, as to gain insights from others.

[karmic coral]

im sharing all of my solutions copy them idc if you solved the theory of general relativity problem with it then ty i helped with that 😂

[karmic coral]

dont be a Scrooge its the holidays bro 🙌

[woven oriole]

since when did nim became a language and not a game

all nim reminds me of is stupid grundy problems 😔

[scarlet locust]

im really hoping tomorrow isn't also grid based, im kinda sick of them LOL

(incoming monkey paw curling and we get a 3d puzzle)

[hybrid gull]

Or an infinitely recursive fractal grid, like 2019 day 20.

[scarlet locust]

https://tenor.com/view/look-up-scared-willem-dafoe-gif-17761333

[lime fulcrum]

you fool now we'll get beacons 3.0

[twin ingot]

fractal-looking beacons on an infinite 3d grid with asteroids

[rancid creek]

and teleporters and keys!

[ripe meteor]

Real, im so tired of these grid problems

[twin ingot]

and robots that arrange into multiple different shapes!

[scarlet locust]

throw in some elephants and some rocks too

[lime fulcrum]

i kind of like grid problems

[twin ingot]

grid problems can be fun but like the last few days have only been grids

[mossy basin]

haven't even gotten our cellular automata problem yet

[lime fulcrum]

lets see 4,6,8,10,12,13(debatable),14,15,16 are grid problems

yeah they are abundant

but to be fair grid problems are a still a very wide field

[modern terrace]

mazes are scary

[lime fulcrum]

agreed. pathfinding problems in general are tricky

[modern terrace]

my program paused output for some reason

i don't know if i should be worried for a cycle

[lime fulcrum]

run it in debug. check where it hangs

[modern terrace]

oh

it's just proceeding ever so slowly

..i can just recursively define this, right

[magic oxide]

i think i'm kinda burnt out from doing purescript

functional programming is interesting, but day 12 was hard enough as it was

and then another grid problem today is

[modern terrace]

2 approaches running at the same time because why not

[blazing vector]

oh no we didn't need to find shortest distance 🤦‍♂️ day16

[rancid creek]

do you have a seen set/dict ? if not cycing is certainly possible

[modern terrace]

i do have one

i'm gonna blame it on all the copying though

okay i found a faster way that takes like 10 minutes instead of an hour or so and so

there has to be a smarter way right-

[terse comet]

iirc all problems from all years should have a solution that takes under 1 minute
some might take a few seconds even with an optimal solution though

[modern terrace]

oh

part 1 is done :D

[cold plover]

hello avents of code

hru doin

[rancid creek]

Yes. A typical solution should work under a few seconds here.

[rancid creek]

Hullo!

[hollow wharf]

Gotta get back to it

[cold plover]

🎀

[cyan sleet]

some might take a few seconds even with an optimal solution though
huh?

the years I was learning rust I was well <1s in total across all days

[slender ravine]

Looking at my 30mins python and 7mins30s csharp
...

[rapid wolf]

the rule of thumb is at most 30 seconds in a slow interpreter.

[cyan sleet]

even that seems immensely slow

[magic oxide]

isn't the rule of thumb this line

[modern terrace]

at most 15 seconds on ten-year-old hardware
(for people with more than a little problem solving skills)

[tall quest]

For Day 16: (no coding as of yet) but I'm trying to come up with different concept by ||creating a node into graph network...At every time there is a fork, there shall be a coordinates of maze and when it is branched out, there shall be a coordinates of the next step, the same is true for another branches...However, there shall not be any coordinates for the long shaft of wall until there is another intersection/fork (and that's where we create another node and making a link between)...And so on and so forth until I have a complete graph network of the maze. Then what's left of it...is just by learning which in between the S & E has the fewest node bridging between it.||

[magic oxide]

you may want to spoiler that

||but yeah i did that too||

||text|| -> ||text||

[humble copper]

well today hit like a truck

😭

[mild hill]

Guys, in the today's first example am I tripping or is the given shortest path not actually the shortest ?
Is there some rule I am missing ?
This is what I find as the shortest path

||```

# # # # # # # # # # # # #

. . . . . . . # x x x x x

. # . # # # . # x # # # .

. . . . . # . # x x x # .

. # # # . # # # # # x # .

. # . # . . . . x x x # .

. # . # # # # # x # # # .

. . . . x x x x x . . # .

# # . # x # # # # # . # .

. . . # x . . . . # . # .

. # . # x # # # . # . # .

x x x x x # . . . # . # .

x # # # . # . # . # . # .

S . . # . . . . . # . . .

# # # # # # # # # # # # #

28 Steps

[tender swallow]

||read the bit about scores again||

[mild hill]

I am not calculating the scores here though, just the steps

I should be getting 36 no ?

[tender swallow]

||you're supposed to find the path with the least score, not least steps.||

[tall quest]

@mild hill (Number of steps) + 1000*(Number of Turn + 1 for initial forward) = Total Score

[hollow wharf]

Come onnnnn I hate those grid challenges

[austere basin]

I had the same issue, spent an hour on it. For me it was because of the random newline characters in the instructions, which somehow didn't affect the answer on the example, but obviously did for the big one. Idk why they even exist, but just make sure your code accounts for the "\n"s.

[compact wasp]

same boat lol, been having a fever

will get back to it when i'm feeling better

[earnest forge]

I didn't get enough sleep so today probably won't go great

[karmic coral]

drivers start your engines lol

[shut trellis]

<@&518565788744024082> Good morning! Day 17 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/17. Good luck!

[tiny tendon]

Dope

Got first Part correct!!!

The 2nd part is taking time dudee

[pure sphinx]

lie

[pure sphinx]

frrrrrrrrrrr

[devout tusk]

are y'all using some ||data structure||??

[karmic coral]

yea

[scarlet locust]

input file is only 86 characters

https://tenor.com/view/kataman-scared-gif-5117551213292804995

[devout tusk]

they're just not fast enough

[pure sphinx]

i will do this tomorrow

my precious sleep time is getting wasted

[humble copper]

||brute force.......||

[woven oriole]

wait

we have to output WITH commas right?

[humble copper]

potentially not ideal but......

[woven oriole]

why tf is it failling

[humble copper]

yes

[woven oriole]

another day of sample works tc doesnt smh

[woven scaffold]

||are your opcodes 1 and 4 correct?||

[woven oriole]

lmao thanks

||i forgot to do something in 4th opcode||

||i was updating the integer variable C, not the list i created||

[covert arrow]

503/41!!!

[woven scaffold]

Nice

[covert arrow]

Time to catch up on the last 2 days now

This is the first time I feel like my top 100 is deserved

[slender ravine]

||Is bf second part possible||

[devout tusk]

what data structure are you even using??

[woven oriole]

||sirs my program is not stopping||

how fucking big

[karmic coral]

if you have while loops in your solution you're basically brute forcing a solution that is not logically sound

[covert arrow]

am I allowed to share this yet?

[woven oriole]

||it has more than 6 digits?||

[woven scaffold]

leaderboard is full

[slender ravine]

Umm yep lb just full

[covert arrow]

ah yes ok

[karmic coral]

||eg infinite loops||

[slender ravine]

https://discord.com/channels/267624335836053506/1318453896095469578

[covert arrow]

||DFS extaspacesoitsnotsuspicious||

[scarlet locust]

||for bxl/opcode 1, do i write it in binary representation or just decimal to b? ex. 3 or 011?||

[woven scaffold]

||You just have to store the number||

[scarlet locust]

thank you!

[pine tiger]

diabolical problem today

[humble copper]

i was waiting for the opcode problem honestly

[devout tusk]

took exactly an hour to solve p2 😭

[scarlet locust]

||hey so what do i do if my input results in an infinite loop? asking for a friend||

[scarlet locust]

update: nvm i accidentally used an i instead of a j and didn't notice because i don't have my glasses on

[candid berry]

Much more

[scarlet locust]

how many?

[humble copper]

im completely stuck with part 2

[candid berry]

Mine had ||15 digits||

[scarlet locust]

||so brute forcing today is impossible, understood||

[zinc kindle]

In most AOC years (including this year) if they give you a programming language that you have to implement and run for part1, part2 requires ||studying your input to figure out what the given code is doing||. Drove me craze in my first AoC year before realizing that.

[warm gazelle]

added a note-to-self for future me lol:

def main():
    print("warning: unless you want to be computing this for the next ~90 years, don't run this on the input data")

[warm gazelle]

rip mine had ||16||
edit: oops i thought you were talking about input digits, not output for p2

[karmic coral]

im a newb 😂 the ||input data is literally less than 10 lines of information||

[warm gazelle]

i did the math (based on someone's estimation that the part 2 output is roughly correlated to some formula) for ||16 digits|| and found that if ||your solution is 100x faster than my (unoptimized) python solution, running on 32 cores|| it will take roughly ||10 days||

so not impossible but definitely not something you want to brute force

[scarlet locust]

well, time to get to reverse engineering

[warm gazelle]

however my math is probably wrong, as usual

[scarlet locust]

we're programmers, if we get something right on the first try something is wrong

[karmic coral]

im upgrading my software packages leaderboards can wait

[scarlet locust]

this was the most difficult one this year so far for me, good luck to everyone who hasn't attempted it yet o7

[woven oriole]

holy shit

the funny thing is

i dont even know whats happening

👍

[scarlet locust]

how'd you do it? i'm curious

i feel like the way i did it was insanely hacky

[woven oriole]

like most people did

||bruteforce in packets of 3 bits recursively||

[scarlet locust]

||same||

[cyan sleet]

just go to the spoiler channel 😛

[scarlet locust]

||i wonder if there's an input out there where that approach wouldn't work?||

[thin topaz]

well my program basically did ||A >> 3 every time and the program is 16 numbers so A must be at least 8^15||

so ||2^45||

I don't really know if you can do this for some arbitrary program

prob no

I don't have an idea how to write code for it either

I imagine it would be very complicated if you can jmp to odd position but my program didn't have that

[wispy saddle]

||ediio||

[hollow wharf]

:^)

[thin topaz]

finally got it

feels good when I first tried the answer

without having to spent hours debugging

[cyan sleet]

@minor cave I have the bestest idea, I'm guessing you might already have ||python as the last day|| but what about ||python 0.9.1||

[woven oriole]

what about ||pypy|| 😔

[magic oxide]

here are my thoughts on language roulette so far

[minor cave]

Lmfao this is incredible

[floral hare]

oh wow, this sums up every single language that has come up

[magic oxide]

lean is the only one in there that hasn't appeared

but i didn't really want gleam to be the only ok functional language

[sharp mauve]

Question. What's your total time for day 5 part 1 and 2?

I got 0.009589757000867394 and was wondering if others got sth much better.

[distant igloo]

Spoilers

[sharp mauve]

Not giving the solution though, just asking how long it took.

[tender swallow]

in Python? mine is 0.3s for both to run

[sharp mauve]

Yeah in python.

If I run both together, it'll be 0.015182451999862678.

[magic oxide]

0.002518958004657179 for both

0.0017596250108908862 with a small optimization

[sharp mauve]

In python?

[magic oxide]

yeah

[sharp mauve]

I was able to make it 0.007004650000453694 for both.

Nice hehe.

[cyan sleet]

what was day 5 again?

[sharp mauve]

Could I see your solution?

[sharp mauve]

The rules and updates one.

23|53.

Had data like that.

[cyan sleet]

that was...err...pascal?

[sharp mauve]

I did it in python.

I don't know what you mean by pascal.

[cyan sleet]

oh no

it was forth

[sharp mauve]

Could I see your solution?

[magic oxide]

sent it in the thread

[cyan sleet]

using time I see like 10ish ms with gforth, but I suspect I can't trust time to give me too useful resolution

[cyan sleet]

I've started remembering the problems based on the roulette language used ok

I'm not damaged

I swear

[magic oxide]

how do you run forth?

specifically how do you time it

because iirc i was running it via repl

[covert arrow]

My initial solution (both parts) was ~0.2s, down to ~0.003s with a few optimizations

[sharp mauve]

Nice!

Could I see the solution please? Trying to learn.

definitely not plotting to take all of you down after stealing your techniques

[gray mantle]

@admin

@graceful condor

@sharp mauve

Need some help

[cyan sleet]

time gforth main.fs -e bye < input

so I'm measuring with a bunch of overhead of random stuff

[magic oxide]

damn

mine is like .5s

[cyan sleet]

I did a 100 * 100 lookup table to do comparisons in not O(terrible) time

[magic oxide]

ah

i just did lin search i think

[hollow wharf]

4-5 hours later, finally got the part 2 solution ||I was doing one way and then tried doing the opposite and it worked fine lmao||

[tiny tendon]

7 hours for new day

[jolly echo]

Is it completely free?

Cuse 2nd day the firefox browser doesn't seem to work just kept buffering for 2 days

[tender swallow]

AoC is free, yes. Are you saying the day 2 page isn't loading for you? That seems unrelated. I can load it fine.

[jolly echo]

No the page is loading but the firefox browser in the VM isn't working

To open elastic SIEM

[tender swallow]

Sorry, how is this related to Advent of Code?

[jolly echo]

Cuse it's an issue that came up in the 2nd day of AOC room ig

Wait isn't this Advent of Christmas of THM?

[compact wasp]

if it's so free then get on the leaderboard 🙄

[peak dock]

it is not

[jolly echo]

Ohh my bad

[peak dock]

https://adventofcode.com

[tender swallow]

No, TryHackMe's event is different. This channel is about Advent of Code. https://adventofcode.com/2024

[jolly echo]

Ahh cool lme check

[next viper]

Today's part 2: WILDLY fun

It has reminded me how fascinating I find compilers... Is the dragon book worth reading still? I kind of want to get my hands on a copy but they seem pretty expensive

[tender swallow]

I think it's a nice book. If you want a more to-the-point/practical introduction, Crafting Interpreters is great.

[scarlet locust]

advent of code gods, please bless me with an easy one today, i have so much work 😭

[peak dock]

same

[karmic coral]

2nd day im early

i think i will get stuck

[next viper]

thanks, Maybe I'll start with that one 🙂

[shut trellis]

<@&518565788744024082> Good morning! Day 18 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/18. Good luck!

[tender swallow]

truly the Advent of Grid

[scarlet locust]

enough with the GRIDS alreaduy

no more!!!

[coarse nymph]

https://www.youtube.com/watch?v=4-J4duzP8Ng

[scarlet locust]

died december 16 born december 18

welcome back reindeer race

[next viper]

its giving https://youtu.be/fuEJWmxWkKw?t=7

[scarlet locust]

part 2 please don't be evil

[next viper]

Hahaha When I saw 'nanosecond' was a link I was like "I bet thats RADM Grace Hopper".

[devout tusk]

428/206 :/

[scarlet locust]

how bad is part 2 on a scale from 1 to 10

[peak dock]

||fairly straightforward||

[earnest forge]

||simple enough that BF can work||

[devout tusk]

this.

[karmic coral]

i beat oo31 today haha less go

oh wait

are you cefrqn on the lb 0o31

[peak dock]

starwort is 0o31

[earnest forge]

I mean I got 173/171

[scarlet locust]

oh lol nice

[earnest forge]

Draw your own conclusions

[scarlet locust]

thank goodness, i really needed time today

[karmic coral]

i made it to global

[earnest forge]

Oh nice

[worthy arch]

Advent of Grids

[modern terrace]

totally lost on part 1

[devout tusk]

i'm tired of grid based problems :/
I'm slow at them and they completely messed up my ranking in the past 4 days

[earnest forge]

you should use aoc-helper, it comes with a handy dandy Grid class

[worthy arch]

Haha same. Doesn't help that I code everything from scratch each day 💀

[karmic coral]

i want to share my ranking with the python community we all ranked 🫶

[modern terrace]

how to pathfind

[earnest forge]

Grid.pathfind

:P

[modern terrace]

oh my god it was easier to implement the algorithm from wikipedia

wow

i'm so dumb

[marsh current]

Available tags

» args-kwargs
» async-await
» blocking
» botvar
» class
» classmethod
» codeblock
» comparison
» contribute
» customchecks
» customcooldown
» customhelp
» dashmpip
» decorators

To show a tag, use /tag <tagname>. (Page 1/8)

[slender ravine]

Oops, what I even typed

[woven oriole]

god damn it i always wake up late on easy days

[warm gazelle]

can part 2 be bruteforced?

[woven oriole]

||yes||

[peak dock]

||yes||

[warm gazelle]

thank you

||this is a nice break.. last few days were hard||

[woven oriole]

much needed

i would have lost my mind after yest disaster

[scarlet locust]

||im very grateful||

||solved both, time to relax for the next 23 hours||

[warm gazelle]

i still have part 2 from yesterday to do. i'll tackle that in the morning

[candid berry]

This was weirdly easy for week 3

[mossy basin]

maybe it was an apology for yesterday

[worthy arch]

hardest problem imo

[woven oriole]

so grateful

[warm gazelle]

i was so worried when reading it lol

[woven oriole]

same

[hollow wharf]

Try not to make a grid challenge yet again challenge impossible

[magic oxide]

honestly thought p2 would be ||find shortest path where every step you take, a new block drops||

[peak dock]

yeah same

[magic oxide]

tho that would've been too easy ig

[hybrid gull]

Day 18 ||networkx|| to the rescuuuuuuue

[woven oriole]

i thought it would be blocks going down every second due to gravity in y

[hybrid gull]

I thought it was gonna be both 😫

[slender ravine]

||Took me some time to confirm it's not gravity simulation (cuz without it's easy but with gravity it's too hard)||

[humble copper]

i was honestly expecting something along the lines of ||every step you take, the next byte falls into place||

but yeah this was significantly easier than expected

especially this late

[karmic coral]

||every step you take every move you make 🎶 ||

[unique osprey]

Yeah i was thinking this + ||a byte dissappears after X steps||

[lime fulcrum]

oh and i would have the question ||when is the latest time you start so that you can still make it.||

[hollow wharf]

honestly quite lots of people thought about that

[karmic coral]

200k+ to 20k+ in day 18 sheesh might need to have tiered levels for the events based on difficulty no?

[hollow wharf]

?

[karmic coral]

Day 1 200k+ participants completed the puzzles. Day 18 only 20k+ so far

[hollow wharf]

161k you mean?

ah edited to day 18

well, compare with other years and it's exactly the same trend:

[karmic coral]

Ow my eyes that’s too many numbers Krypton 😂

[hollow wharf]

Yeah wait, trying to make it smaller x)

[minor cave]

Yeah this is expected. The puzzles do get more difficulty as it goes on and as things get closer to the holidays people's schedules get fickle

It's not a bad thing. The days towards the end are supposed to be more challenging

[hollow wharf]

2023 differences

2022

2021

it's currently also at a fairly normal state

people stop solving them either because of lack of time, motivation or knowledge

sometimes people even forget about it, especially people that do it for the first time

they read somewhere it exists, solved the first (maybe second) day and then forget about it or doesn't like it

[karmic coral]

Reminds me of a marathon really cool event

Does AoC do prizes? Or merch or something

I’m surprised big tech hasn’t jumped on this as a marketing opportunity

[minor cave]

They have merch you can buy, but no prizes. The event is just for fun
https://cottonbureau.com/people/advent-of-code

[hollow wharf]

they have sponsors

some communities make private leaderboards and then distribute prizes, but there's no prize for global leaderboard

[cyan sleet]

"I did the AoC roulette and all I got was this ||pink role||"

[karmic coral]

What is AoC roulette

[distant igloo]

Doing AoC in a different language every day -> https://discord.com/channels/267624335836053506/1312400095123341465

[karmic coral]

How do I participate the channel is read-only on my end I’m a newb x)

[distant igloo]

The channel is just an announcement telling you which language to use when solving the puzzles

[karmic coral]

Ah I see hmm

[karmic coral]

Someone said advent of bruteforce the other day idremeber which day it was

[royal gull]

https://youtu.be/uZ8DcbhojOw?feature=shared

(Don't think I saw this posted , it's two weeks new)

[hybrid gull]

Is this recent? I feel like I've watched this talk?

[royal gull]

It says 2 weeks?

[hybrid gull]

Yeah, I've just seen this graphs and all this before. Looks like maybe a talk he gave before, just presented it newly again.

[royal gull]

(Looks like the conference was in July, but just posted videos)

[tender swallow]

he also did a behind-the-scenes talk in 2021 IIRC

[royal gull]

Ooh, he shows something from @cyan sleet at 30:59

I was hoping to see something from salt die

[scarlet locust]

good luck everyone, final week starting in 3 minutes o7

[shut trellis]

<@&518565788744024082> Good morning! Day 19 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/19. Good luck!

[scarlet locust]

please don't let part 2 be too bad

[humble copper]

what....

its

over

lmfao

[late frigate]

bruh

i thought i got record time but naw

[scarlet locust]

how bad was part 2 from 1 to 10

[humble copper]

2

minor roadblock when i just changed what i thought i needed to from part 1

but the first (and most obvious) solution, without giving too much away, worked flawlessly

[woven oriole]

wtf thats it

god damn it i || fucking messed up parsing due to splitting by ',' and not ', ' and my dp[-1] was giving 0||

[humble copper]

yeah

thats it

😭

[karmic coral]

its all Eric's fault

[humble copper]

i cant help but feel like this is to lure us into a false sense of security

[karmic coral]

so are we looking for a Historian or Santa Clause or something. this is my first year doing aoc

[scarlet locust]

isn't day 20 usually the evil one

i feel like thats on average the worst one

[humble copper]

i only have 2021 to reference off of

day 19 on that (beacon scanners) was hell

[woven oriole]

same

[woven oriole]

okay lmao i too remembe

the coordinates hell

[devout tusk]

Literally in the middle of nowhere. Somehow got internet access and leaderboarded. 48/50 😭

[modern terrace]

thankfully a somewhat easier day

[rancid creek]

😦 beacons/scanners was one of my favorites.

[lime fulcrum]

oh the concept is awesome. its just very hard

[rancid creek]

Got to make a Beacon class.
||

class Beacon(tuple):
    def __add__(p1,p2): return Beacon((a+b for a,b in zip(p1,p2)))
    def __sub__(p1,p2): return Beacon((a-b for a,b in zip(p1,p2)))
    def rotate(s,a,b,c,i,j,k): return Beacon((a*s[i],b*s[j],c*s[k]))

||

[lime fulcrum]

i'm glad it has unlimited towels of the same same pattern. would have been annoying otherwise

(for part 1)

[jovial aspen]

Overall difficulty (part1 + part2) for me (higher rank = more difficult):
||```
Day 16 (maze problem)
Day 14 (christmas tree; u need to spend a long time on it)
Day 12 (garden group)
Day 13 (claw machine; it is a math problem)
Day 17 (3 bit pc)
Day 15 (push box)

rest of the day are easier than these

[naive stone]

damn, my answer is correct for a different input, but not mine. So i'm close, but not close enough. My code works for the example, hmmmm
-# edit - i figured out the issue

[hollow wharf]

||memoization banger||

[hollow wharf]

i found that day 14 much easier

silly way of doing part 2 though

[pine tiger]

cool script runner! how’d you get it to look cool like that?

[hollow wharf]

Oh that's just hyperfine for the benchmarks

https://github.com/sharkdp/hyperfine

At first I calculated myself, but then discovered that tool

Run it multiple times to get that mean, min and max

[magic oxide]

it grows

[cyan sleet]

oh god that was a long time ago

neat

[scarlet locust]

that was such a monstrous 6 day stretch

[scarlet locust]

you know the phrase "calm before the storm"

also huh, TIL every day's description has a secret hidden word that when hovered over provides some extra flavor text

||hover over the word "stripe" right before when the colors are listed in today's description... turns out there's one of these for every single day so far||

[naive wharf]

||userscript for highlighting when||?

[scarlet locust]

||the website automatically does this when you get all 50 stars for that year!||

[cyan sleet]

advent of roulette repeats 😔

at least ruby is a very sane language

[magic oxide]

agreed

[cyan sleet]

interesting, apparently I don't have ruby installed on my desktop

[hollow wharf]

||styled em and span elements differently||

[jovial aspen]

oh you ||highlight them by adding this to css||
||```css
em {
background: red;
}

span {
background: green;
}

[next viper]

start todays
this seems simple
implement very simple brute force solution for test input
works great
look at actual input
ohno.jpg
try to run brute force solution?
computer melts
dynamic programming it is

[magic oxide]

i looked at p1 and was thought p2 was gonna be like ||oh no now each towel is duplicated 15 billion times||

[normal loom]

||Day 3 of @cache||

[unique osprey]

These channels have been quiet the last couple days. Are people just dropping off?

[hollow wharf]

im working on solving advent of code in goboscript

been fairly easy days recently, solved quite quickly by lots of people and then moved on with the daily life things

and others may have stopped indeed

(which is a language i made that compiles into scratch)

[tulip anvil]

its christmas soon wth

[tender swallow]

I guess so. Like with every year

[unique osprey]

To be fair i'm usually dropped off by this point

[humble copper]

probably

though in my case i’ve already finished today’s problems so i’m out and about doing other things until the next day’s set gets released

[hollow wharf]

these are actually hilarious

the do_not_mul with the mul being highlighted

and the undo() here

[cyan sleet]

yes

[pine tiger]

wait the numbers still seem a lot bigger ngl

[woven oriole]

true

20k

[scarlet locust]

good luck everyone, 20's usually the worst from what i've seen and heard 🙏

[modern terrace]

haha

funny

[shut trellis]

<@&518565788744024082> Good morning! Day 20 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/20. Good luck!

[scarlet locust]

this truly is the advent of grid

[gentle cape]

advent of BFS

[karmic coral]

nooo i have to wait 4 minutes lol

welp there goes lb for me

found this cool new algorithm tho i wanna see what it does

[devout tusk]

219/51

I DESPISE GRIDS

[hexed axle]

https://discord.com/channels/267624335836053506/1319533788300054629

[karmic coral]

you beat me to lbs lol

i think i will start just not caring about leaderboards anymore

[devout tusk]

Also, my dumbass had 1411 as the answer and typed out 1412 (not my actual answers) and got a 1 minute penalty on part 1 😭

[devout tusk]

neither did I until this AoC when I actually started leaderboarding and am in top 100 global lol

[magic oxide]

dang 151 for p2 isn't bad

[devout tusk]

congrats!

[karmic coral]

yeh i got global past 2 days its getting eehh

[magic oxide]

my dumbass ||thought through it, wrote >= 100, then changed it to > 0 for the test input, and then changed it back to > 100 and lost like 5 mins||

truly an advent of "make a grid util"

[modern terrace]

i wasted time with a non-bug TT

[earnest forge]

How dare you summon this puzzle /j

[scarlet locust]

at least it isn't 2019

[earnest forge]

I liked 2019

I can't get my pathfinding algorithm to work 🙃

[magic oxide]

this one wasn't too bad, tho ||my runtime for p2 is terrible||

[modern terrace]

||@cache|| doesn't help me here..

[karmic coral]

this challenge would be perfect for people interested in robotics

[earnest forge]

Well it didn't help that ||I accidentally turned the goal into a wall|| 🤦

[karmic coral]

3300/1400

i just realized something. graph neural networks

gl those who haven't completed the puzzles you got this

[humble copper]

today was nice !

[modern terrace]

858/1782

[scarlet locust]

today wasn't as bad as i thought

[peak dock]

i finished aoc 2023 👍

only around 360 days late

[blazing vector]

are there any crossroads in grid day20?

[magic oxide]

||no. ||

[blazing vector]

omg off by 1 is like daily routine 🤦‍♂️ forgot to remove E from grid...

[wary river]

congratulations 👏👏

no mean feat

[thin topaz]

this is my first year attempting

[fervent anchor]

today I'm gonna rest

I have mega project to do

[oak flax]

my solution for part 1 runs in 5 minutes, but if it works it works ⭐

[urban urchin]

@fervent anchor
Can u elaborate about your project?

[fervent anchor]

it's a machine learning project I have to send tomorrow

[royal gull]

Mega project due tomorrow? I hope you started 🙂

[fervent anchor]

it's the final date

[hollow wharf]

@modern terrace help me

[cyan sleet]

if char != '#' 😌

[distant igloo]

I think you meant to ask this in #discord-bots

[blazing vector]

thats before even solving just to remember start and end

[cyan sleet]

just saying, I didn't remove anything

[paper folio]

TGIF time to catch up on AoC 🔥 anyone else also behind now? I'm on day 17 p2 reverse engineering my registers

[modern terrace]

sorry

i'm not awake that early

[hollow wharf]

my friend helped me

i was doing condition1 or condition2 while i had to do condition1 + condition2

[modern terrace]

13 minutes until i decide if today is worth my one hour before i leave

[earnest forge]

Good luck everyone

[slender ravine]

4 minutes until I decide if I just go back to sleep and wake up 2 hours later to solve it instead

[shut trellis]

<@&518565788744024082> Good morning! Day 21 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/21. Good luck!

[magic oxide]

wtf

ok yeah i'm taking a break before i do this

[jovial aspen]

oh what iS tHaT???

the question seems soo long

[slender ravine]

yeah nope I am going back to sleep for 2h

[covert arrow]

The moment I read it I was like "hey I've seen this before" ||https://www.youtube.com/watch?v=X4YHpZ0EfOE||

[slender ravine]

so we have:
||
The numereric keypad that control by robot 1 that is depressurized
The directional keypad to control robot 1 by robot 2 that is in high levels of radiation
The directional keypad to control robot 2 by robot 3 that is in -40 degrees
The directional keypad to control robot 3 by you in an area full of historian
||

[jovial aspen]

I got a smaller result (from the example given)💀

[slender ravine]

I have an idea but gn

[covert arrow]

i might try this manually at this point

[devout tusk]

test gives correct ans

input doesn't work 😦

[somber trench]

How on earth do people do these so quickly

[earnest forge]

This is ridiculous

[karmic coral]

looks like ||they referenced numerous aoc challenges from years past||

[devout tusk]

already have a 5 min penalty now :/

[karmic coral]

6 people finished day 21 already wth 😂 are they bots or something

these puzzles are so cap -40 degrees what farenheit or celsius wth Eric 😂

[devout tusk]

i think everyone ||has the same input||

[jovial aspen]

bro

whats wrong with my pc
how come the answer is different when re run it??

[magic oxide]

ok this problem is so cursed pycharm died

lmfao

[devout tusk]

123/14
finally!

[magic oxide]

pycharm actually died wtf

[devout tusk]

weirdly, chatgpt helped a bit today

[pine tiger]

ru on this discords leaderboard?

[devout tusk]

how do I get on it? 🤔

[covert arrow]

why not kelvin

[devout tusk]

i'm doubting your science skills more than coding skills

[pine tiger]

&aoc join

[devout tusk]

&aoc join

[shut trellis]

Command Help

**```
&adventofcode

**Can also use:** `aoc`

*All of the Advent of Code commands.*

**Subcommands:**
**`adventofcode about`**
*Learn about Advent of Code*
**`adventofcode countdown`**
*Return time left until next day*
**`adventofcode link [aoc_name]`**
*Tie your Discord account with your Advent of Code name.*
**`adventofcode unlink`**
*Untie your Discord account from your Advent of Code name.*

[pine tiger]

at least i think so

oh yeah i think you have to link first

also do it in #aoc-bot-commands

[devout tusk]

hold up, how do I link my name tho?

[karmic coral]

yeh i am beginning to doubt the real-world logic of these aoc puzzles 😂

[minor cave]

#1319908014118928476

[jovial aspen]

my code works for example input
but not the puzzle input🗿

[hybrid gull]

OOF. ||two same length, but different order, of sequences of button presses for the robot you're controlling can make your sequence of buttons presses shorter or longer|| 😭

[jovial aspen]

yes

[jovial aspen]

I made it
rank 579, (the first time I got on lb)

[woven oriole]

||how is 379A 64 and not 68||

[covert arrow]

Perhaps I should have kept the /j

[woven oriole]

smh

[covert arrow]

||<<^^ instead of ^^<< somehow||

[woven oriole]

wow

okay

wait

[jovial aspen]

p2:
||length of string is already > 300,000 for 11 iterations||

I don think this method will work

[devout tusk]

happened to me as well ahah

[devout tusk]

lol tru

code golfing time? 💀

[covert arrow]

Nope, I just finished p1

Is p2 ||just dp||?

[karmic coral]

i have to wait 6 mins still on part1 😂

[jovial aspen]

||Dont you need to keep track of both (str, len(str)) instead of just len(str)? maybe I am incorrect||

[pine tiger]

yeah it’s exponential like that

[covert arrow]

Not sure, just thinking of ideas

[jovial aspen]

||if it is exponential then we have to use other method||

||even if you only retain the minimum length of the string after each iter, the length will still exceed 1,000,000||

[devout tusk]

the only thing exponential is the submission graph in the past few days 💀

[pine tiger]

||something like the stones wouldn’t work because position actually matters||

[jovial aspen]

yep

i have no idea on it now 😭

[woven oriole]

||are we talking about numeric pad||

||not sure but what i did was, we have to prioritize moving left right as much as we can before up down, after which if it was the left most column should go one more left, but this does not work smh||

[jovial aspen]

will ||something like ^>^> make the final string shorter, in some cases?||

||um i am sure the answer is no||

[woven oriole]

i need proof of what the fuck is going on

[jovial aspen]

what proof

[woven oriole]

||379A giving 68 for me||

[covert arrow]

My idea was ||right or up first because it's closer to A but apparently not||

[covert arrow]

Yea, ||the final arm will trace out ^A<<^^A>>AvvvA||

[woven oriole]

okay i get|| ^A<^^<A>>AvvvA||

interesting

[covert arrow]

||I'm not doing any of that, just assume the optimal movements on the intermediate keypads are fixed||

[jovial aspen]

heres some of the examples of 379A, len=64
||```
v<<A>>^AvA^A<vA<AA>>^AAvA^<A>AAvA^A<vA>^AA<A>A<vA<A>>^AAAvA^<A>A
v<<A>>^AvA^A<vA<AA>>^AAvA^<A>AAvA^A<vA^>AA<A>A<vA<A>>^AAA<A>vA^A
v<<A>>^AvA^A<vA<AA>>^AAvA<^A>AAvA^A<vA^>AA<A>Av<<A>A^>AAAvA<^A>A
v<<A>>^AvA^Av<A<AA>>^AAvA^<A>AAvA^A<vA^>AA<A>A<vA<A>>^AAA<A>vA^A
v<<A>>^AvA^A<vA<AA>>^AAvA<^A>AAvA^Av<A>^AA<A>Av<<A>A^>AAAvA^<A>A
v<<A>>^AvA^Av<A<AA>>^AAvA<^A>AAvA^A<vA>^AA<A>A<vA<A>>^AAAvA<^A>A
v<<A>>^AvA^Av<A<AA>>^AAvA<^A>AAvA^Av<A^>AA<A>Av<<A>A^>AAA<A>vA^A
v<<A>>^AvA^Av<A<AA>>^AAvA^<A>AAvA^Av<A^>AA<A>Av<A<A>>^AAA<Av>A^A
v<<A>>^AvA^Av<A<AA>>^AAvA^<A>AAvA^A<vA>^AA<A>Av<<A>A>^AAAvA^<A>A
v<<A>>^AvA^Av<A<AA>>^AAvA<^A>AAvA^A<vA>^AA<A>A<vA<A>>^AAA<Av>A^A
v<<A>>^AvA^A<vA<AA>>^AAvA^<A>AAvA^A<vA>^AA<A>Av<A<A>>^AAAvA<^A>A
v<<A>>^AvA^Av<A<AA>>^AAvA<^A>AAvA^Av<A^>AA<A>A<vA<A>>^AAA<A>vA^A
v<<A>>^AvA^A<vA<AA>>^AAvA^<A>AAvA^Av<A>^AA<A>A<vA<A>>^AAA<A>vA^A
v<<A>>^AvA^Av<A<AA>>^AAvA^<A>AAvA^A<vA>^AA<A>Av<A<A>>^AAA<A>vA^A
v<<A>>^AvA^Av<A<AA>>^AAvA<^A>AAvA^A<vA^>AA<A>A<vA<A>>^AAAvA<^A>A
v<<A>>^AvA^Av<A<AA>>^AAvA^<A>AAvA^A<vA^>AA<A>Av<<A>A>^AAAvA^<A>A
v<<A>>^AvA^A<vA<AA>>^AAvA<^A>AAvA^A<vA>^AA<A>Av<A<A>>^AAA<Av>A^A
v<<A>>^AvA^A<vA<AA>>^AAvA^<A>AAvA^A<vA>^AA<A>Av<<A>A^>AAAvA^<A>A
v<<A>>^AvA^Av<A<AA>>^AAvA<^A>AAvA^Av<A>^AA<A>Av<A<A>>^AAA<Av>A^A

I have a lot more answer than that btw

[woven oriole]

okay now ||456|| became costlier lol

okay i understand why

||^<<^A>A>AvvA||

[woven scaffold]

same for me, been trying to debug for the last half an hour and still not really sure what I'm doing wrong

Managed to narrow it down to one particular part
||```
379A:

depth 1
my code: ^A^^<<A>>AvvvA
correct: ^A<<^^A>>AvvvA

depth 2 v v (this part messes me up)
my code: <A>A<AAv<AA^>>AvAA^Av<AAA^>A
correct: <A>Av<<AA>^AA>AvAA^A<vAAA>^A

depth 3
my code: v<<A^>>AvA^Av<<A^>>AAv<A<A^>>AA<A>vAA^Av<A^>AA<A>Av<A<A^>>AAA<A>vA^A
correct: <v<A>>^AvA^A<vA<AA>>^AAvA<^A>AAvA^A<vA>^AA<A>A<v<A>A>^AAAvA<^A>A

[woven oriole]

okay wow

sample works

tc too high

[jovial aspen]

some tips
||
so take a look of these if your result is wrong.
1: if one of the fromLocation and toLocation is in 0A^, what will be the possible paths?
2: if one of the fromLoc and toLoc is in 0A^ and the others is not at column 0 (i.e. in the column that doesnt contain a gap), what will be the possible paths?
3: if *none of them * is in 0A^, what will be the possible paths?
||

[jovial aspen]

visual explaination for 1 and 2 (big spoiler for part 1)

[woven oriole]

||does this also imply for the directional pad||

[jovial aspen]

||yes||

[woven oriole]

damn

||and those two paths in the second grid both lead to optimal score i suppose||

[magic oxide]

damn big spoiler wall today

[worthy arch]

Today 💀

[jovial aspen]

||btw can anyone give some tips on today's part 2||
||I dont have any idea on how||

[twin widget]

guys

did you do part 2?

jonathna paulson did it in 3 hours

what about you stickie?

[turbid urchin]

my p2 soln is in https://discord.com/channels/267624335836053506/1319908014118928476 if anyone wants to look

[ivory vigil]

Just finished. 5ms runtime for p2.

[twin widget]

man you´re the only one that has posted a solution

HAHAHAHAHAHA

[woven oriole]

god damn it

finally

and its not over

[magic oxide]

haven't been working on it much

i'll prob finish it tonight

[fervent anchor]

ok good news is I'm done I'll come back to do them soon

[covert arrow]

||yes (kinda)||

[woven oriole]

okay lmao now

||my memo code works for depth = 2 but too high answer for depth = 25||

weird

how cancer can this day be

[covert arrow]

||mess with your input path, the optimal one isn't the same for d=2 and d=25||

[woven oriole]

||i generated all input paths actually and am taking the min, but only for numeric keypad||

do i have to do this shit for || directional too|| smh smh smh

[covert arrow]

||yeah I had that from before||

[woven oriole]

welp

aight

hopefully the last cancerous thing to do today

woohooo

the ||modification was pretty quick all i had to do was run the same script again, cuz the forbidden coordinate is technically the same for both numeric and directional (-2, 0) for me || lmao

[thin topaz]

oh god finally

and I completed part2 instantly lmao

just forgot to cache the function as I was debugging

[humble copper]

is today easy

im just getting started on it now

[lime fulcrum]

today might just be the first day i can't do. we'll see but i can't really make heads or tails of it yet. might also have to do with the fact that i'm not that comfortable with recursion

you know what i want my grid problems back XD less of a headache

[magic oxide]

this problem is just the right amount of fiddly to make me not want to focus on it

keep getting distracted

[thin topaz]

only 4k solves for part2 now

[modern terrace]

this is so real

i thought of a solution but thinking of making it just makes me say "eh, i'll do it later" for some reason

[magic oxide]

i tried working on it a bit in the afternoon and then spent 2 hours watching youtube instead

[magic oxide]

so true

[cyan sleet]

wisdom of today, editing in vim over ssh connected to my desktop which in another country on a phone in a train that's going through tunnels is not optimal

[magic oxide]

bro is doing today's problem irl while doing today's problem

[cyan sleet]

weirdly fitting

yeah ok, I'm not debugging like this

especially in C#

[humble copper]

my attempt keeps undershooting for the example input

:D

FOUND THE ISSUE!

that one paragraph helped so much

now it's overshooting, which is totally fixable

[magic oxide]

just take the average of your two algorithms 👍

[cyan sleet]

wisdom of the crowd

[humble copper]

nvm undershooting again

[magic oxide]

today actually is advent of sit down and concentrate for 10 minutes

i think i've written like 3 lines of code in the last hour

[cyan sleet]

I have a solution for p1, it's just written in pseudo C#

[thin topaz]

a rare time I need to write on paper

[cyan sleet]

because I couldn't read docs reliably in the tunnels

so I just made up some methods and syntax

[humble copper]

perfect.

[cyan sleet]

I have yet to see what the p2 twist is, from your comments it sounds like it might be terrible 😛

[devout tusk]

10 minutes hours

[humble copper]

passes the example input

[thin topaz]

it might be the best thing actually

[cyan sleet]

at least p1 sounds like a wonderful fit for ||generator functions||

[humble copper]

oh

thats what p2 is

[magic oxide]

honestly i would've maybe enjoyed a functional language today

extra type safety seems kinda nice about now

[lime fulcrum]

in the mean time i'm trying to do 16 part 2... not that happy about this

[cyan sleet]

C# isn't type safe enough? 👀

[magic oxide]

it's already midnight

there's no way i'm doing this in C# without a python sol first

[strange zinc]

for anyone, day 21, did ||test case 379A produce something different than what was given? I keep getting a length of 68 and not 64, but for all the other test cases, they are correct||

[humble copper]

i had a lot of issues with that but ||it kept guessing below the actual length instead of above. depending on how your solution is set up, it may be a problem with finding the "ideal" paths between any two keys on the keypad||

[strange zinc]

using ||manhattan distance, then with that determining the path taken||

[humble copper]

it'd be pretty difficult to diagnose the problem without seeing the code being run (especially in here since no real spoilers) but the best advice would probably be to reread the problem carefully, it places a few restrictions that might help trim down problematic cases

[candid berry]

yes, this happens. ||two sequences of keypresses that are equivalent at one layer, like <<^^A vs ^^<<A, can take different number of keypresses to key in at the next layer up||

you found ||a solution, but not the shortest one.||

[strange zinc]

||found my error. I pass them now. Just not getting the correct answer for my input||

[scarlet locust]

ack,, this is such a headache

this might be the day that i finally don't pass 😔

[pine tiger]

||is there a case where one sequence is always peferable over the other?||

[humble copper]

(to my knowledge) no

[pine tiger]

||so there's cases where you might want >v over v> and vice versa||

[humble copper]

i believe so?

i’m still trying to wrap my head around how the problem works, but my solution didnt attempt to solve this i think

[covert arrow]

||the optimal path can change between p1 and p2 even||

[pine tiger]

p1/p2?

[scarlet locust]

i give up, i made it a good 20 days though

good luck everyone!!

[humble copper]

parts 1 and 2 are slightly different

[pine tiger]

well part 2 is really just part 1 ++

[humble copper]

yeah

i feel like...

part 2 could have honestly been harder

so im glad it was only what it was

[zinc kindle]

Part2 took way too much effort and too many false starts and failed approaches, but I finally managed to solve it.
This was a weekend one, wasn

wasn't it. Here's hoping that after day22 it's relatively easier.

[humble copper]

should be, if its anything like 2021

actually....

there is tomorrow.......

but its a sunday, and right before christmas

[pine tiger]

oh wait was it a thing

that weekends are usually harder

[naive stone]

in his talk he mentions that he tends to put harder puzzles on weekends because people have more time

[pine tiger]

well today is going to be awful then

[warm gazelle]

for day 21 part 1 do we need to account for ||the cursor not hovering over A when going to the 2nd, 3rd, 4th, etc line of the input||

[thin topaz]

No

it is impossible to happen

[earnest forge]

Good luck everyone

Let's hope it's not quite as nasty as yesterday

[magic oxide]

just rolled out of bed

[shut trellis]

<@&518565788744024082> Good morning! Day 22 is ready to be attempted. View it online now at https://adventofcode.com/2024/day/22. Good luck!

[obsidian shell]

shat up

imma remove the role

[distant igloo]

Be nice to Sir Robin

[dusty dagger]

HOW 😭

[obsidian shell]

#roles

[dusty dagger]

I got the role like 2 years ago and have just been enduring the pings...

Damn that was easier than I thought lmao

[hexed axle]

You're also always welcome to ask in the #python-discussion channel, modmail, etc

[compact wasp]

we're so back

weeklong hiatus due to illness but got 45/7

don't think i'm getting back into the top 5 this time though

[hexed axle]

https://discord.com/channels/267624335836053506/1320257701829152790

[devout tusk]

😭

36/108

missed top 100 on part 2 due to a stupid error

[compact wasp]

rip

[pine tiger]

thank GOD

an easy day

[compact wasp]

[shrewd thicket]

im tweaking out what is this difficulty curve

fuckin mountain range

[compact wasp]

[magic oxide]

bruh no way they changed the test input for p2

i could've finished like 5 minutes ago

[magic oxide]

the king has returned 🙇

[compact wasp]

no king just an arrogant peasant

[karmic coral]

what is up with aoc and old challenges. go live in the past then aoc

[humble copper]

huh

well that was something

here i was thinking part 2 was gonna be something like ||"find the minimum starting value such that the 2000th price is 2024"||

[magic oxide]

i thought part 2 was gonna be ||find the 2 billionth price||

[shrewd thicket]

same

[woven oriole]

okay didnt expect easy day for sunday

except i blundered a lot of time

[thin topaz]

brute force in progress xd

[woven scaffold]

anyone else getting that you have the correct answer for someone else?

I'm definitely using my own input, and my p1 was accepted

[humble copper]

sure

lets do that

nvm it takes forever to even finish 1 😭

[compact wasp]

what if it was ||finding when the secret loops||

[thin topaz]

I don't think the bitwise ops have anything to do with part2 on the surface

or does it

[woven oriole]

is it guaranteed to happen

wow

this would just be ||visisted set and pray||

[compact wasp]

yes

[humble copper]

for any one number, the 100 millionth iteration takes 25 seconds

[woven oriole]

right there's mod

[humble copper]

it would only take....

9 days

to fully do all ||2 billion||

[feral hazel]

yeag
||you can prove it with the pigeonhole principle since there are 2^24 possibilities and the evolution process is a function (same input = same output)||

[compact wasp]

how many cores are you using

[humble copper]

im gonna be so real with you

i have no idea how to check

ive just been gunning these for however long they took

oh wait

[compact wasp]

actually tbh idk why i asked i'm not sure if this can be parallelized lol

[humble copper]

multiprocessig would def work

[humble copper]

different processes for individual numbers?

[compact wasp]

how are you getting the numbers

[humble copper]

i mean from the input

[compact wasp]

oh yeah

[humble copper]

lets try that

how many different processes can be spawned

[compact wasp]

my brain still not in peak form it seems

[compact wasp]

enough

[humble copper]

are we feeling daring...... ||1714?||

[compact wasp]

yeah you should be able to do that

parallelism will only help you up to how many cores you have

but processes are an abstraction so it's ok

[covert arrow]

Could you reduce that by 4x with simd?

[compact wasp]

probably wanna tell them to use threads instead of processes first lol

at least that can be done without 3rd party libraries

[humble copper]

we're already breaking importless with multiprocessing :P

there's def some optimization that could be done though

[compact wasp]

i feel like python perhaps isn't the right language if we're interested in optimization

[karmic coral]

I gave up. i'm implementing ML/DL approaches now on my own input data no lb for me today

[compact wasp]

c++ or rust might be more like it

[magic oxide]

i'm curious ||how often cycles happen||

[humble copper]

for 2021, whichever day was the one where you had a maze that repeated itself some number of times

[magic oxide]

might ||run floyd's a bit later and see||

[humble copper]

my part 2 solution for that day ran for nearly 3 hours straight

[compact wasp]

don't even need ||floyd's|| since it's ||deterministic||

maybe that's not the right word

[magic oxide]

wdym

||it has to be deterministic for floyd to work no?||

[compact wasp]

that's why i said it's not the right word

[karmic coral]

what if it has deterministic ||for robert to work on x)||

[jovial aspen]

um why it doesnt work

[tender swallow]

you're better off making it iterative

[magic oxide]

food for thought
||if you treat the ints as elements of the vector space (Z/2Z)^24, then bitshifting is a linear operator and xor is addition
so getting the next secret number looks like a product of 3 (I + B) terms, which is also a matrix
so you could do fast pow||

[humble copper]

crazy

in just 90 minutes, there have been 35% of the full day completions today than there have been total for yesterday

[karmic coral]

omgerd that puzzle was annoying 😂

[cyan sleet]

ah yes, the perfect setup

[warm gazelle]

some back-of-the-napkin math shows that my part 2 solution for today will take 4 hours to complete

this would be tolerable once i get multiprocessing running on all 16 cores

that's going to be a task for me in the morning

(i could have it running all night, but then i have to deal with my computer lights and my room being hot)

[jovial aspen]

is this 9 minutes or 9 seconds?

if it is 9 minutes then why the next one is 54 minutes

[royal vector]

Whats this?

[jovial aspen]

global leaderboard

[royal vector]

For what

[jovial aspen]

day1

[karmic coral]

9 seconds

[jovial aspen]

even faster than chatGPT 🗿

if u use chatGPT it takes at least 30 seconds and may not be correct

[cyan sleet]

🥴

what are you even doing?

[hybrid gull]

Probably similar, but better, things than I was doing on my first swipe at it. 😆 My first swipe at pt2 was going to take 8hrs (according to my tqdm progress bar), so while it was running, I rewrote the whole thing and now it takes 2.4s. 😅

[elder quest]

AOC bug moment:

6 [[[...], [...], [...], [...], [...], [...], [...], [...], [...]], [[...], [...], [...], [...], [...], [...], [...], [...], [...]], [[...], [...], [...], [...], [...], [...], [...], [...], [...]], [[...], [...], [...], [...], [...], [...], [...], [...], [...]]]
[[...], [...], [...], [...], [...], [...], [...], [...], [...]]

[pine tiger]

lmao how

does that even happen

[blazing vector]

day21 part1 runs in 0.0003s part2 is running for 10+ mins already xD

[normal loom]

I fear it may not finish for a while

[proud cosmos]

i made a solution for today that works on everything i've tested it on (even someone elses input) EXCEPT my own input

going insane

what am I even meant to do about this

[blazing vector]

parsing or off by1

[proud cosmos]

i copy pasted their input the same way i do mine so parsing should be fine
and it's not off by 1 i tried it as a joke the site says it's more than 5 off

there must be some weird edge case in my input that doesn't appear in theirs

and i have no idea what i am supposed to do about it

i even looked at how others did it and it's litearlly the exact same solution idea

it just works for them and not for me

[hybrid gull]

Make sure ||you are getting the price at the end of the sequence you find|| or make sure ||you are getting only the first price find by that sequence||

[proud cosmos]

i do both of those

i have no clue

i just ran someone elses solution, and it also gives my answer (which is apparently wrong)

funky edge case somewhere i assume

[magic oxide]

maybe ||you're missing the first price?||

[proud cosmos]

unsure how that would happen

I found the issue

||if the sequence was the very first 4 differences my code skipped it on accident, and this happened to be the case on exactly one number. this just isn't the case for some inputs so yea||

[grizzled shard]

What is the best plan to learn Python?

[proud cosmos]

Find a project youd like to do and research of how to make it in python

[thin topaz]

solve all of advent of code in python

[magic oxide]

yeah ||this is what i meant by missing the first price||

||i had the same bug||

[proud cosmos]

oh

if only i read that correctly

that would've saved a lot of time

actualyl 15m isn't that crazy but still

[wary river]

just finding a passion project and going through with it will alone teach you incredible amounts of techniques, tips and tricks. it's amazing to start there

[compact wasp]

passion project is fine when possible but you cannot force passion

what you can force is discipline

if sufficiently disciplined you could try reading through the documentation

or watch a yt series

[warm gazelle]

||going through every possible combination from -9 -9 -9 -9 to 9 9 9 9 then finding the ‘score’ at the combination. then print the max score||

[cyan sleet]

🥴

[warm gazelle]

how else am i supposed to do it?

[cyan sleet]

you do things ||the other way around||

rather than ||choosing a value and seeing what the scores it produces for a sequence|| you ||look at a sequence|| and ||see what scores/values it can produce||

[warm gazelle]

||but wouldn’t i have to check basically all sequences anyway? i’d imagine there’s so many numbers that basically every sequence comes up||

[cyan sleet]

with your approach you ||process every sequence a quite large number of times||

||19**4 times||

with the other approach you just ||process once but with some slightly more expensive processing||

[warm gazelle]

||i precompute the sequences and their deltas so that’s not a concern, right?||

[cyan sleet]

that doesn't change much, you're still ||going through all that ~19**4 times||

you can ||extract all possible non-zero scores/values for one sequence by only going through it once||

[humble copper]

is it expected to be smooth sailing for the rest of the calendar

[paper folio]

yo day18 was so ez. I copy pasta my day 16 code for the most part

I was so shook after day 17

i guess on to 19

[paper folio]

which day was the hardest for you?

[humble copper]

yesterday’s

or 17

[humble copper]

19 isnt so bad

its more reliant on concepts from earlier days

[paper folio]

yeah holymoly 17 had me to tears

I was writing down my modulos and bit shifts on a piece of paper for hours

[humble copper]

yeahhhh

17 was hard on me execution wise

21 was hard on me conceptually

but once i got it the execution is somewhat straightforward

[paper folio]

have you done all the previous years too?

[humble copper]

i’ve been participating for the last 5ish years

but the only year i managed to complete fully was 2021