Hmm, the video quality looks oddly bad. I'm guessing I messed up one of the settings in my video editing pipeline (it has been 11 months since I last did it...), I'll try to fix it for future days!
In case anyone's curious: the issue was that I was exporting a 4k video with a target bitrate of 2 Mbps... I turned it up to 60 Mbps and quality issues seem to be gone (for day 2).
in your python setup, can't you write a piece of code at the end that automatically copies `ans` to clipboard? I'm pretty sure this can be done and it will save you a few seconds.
I used set( ).intersection( ) with .count( ) didn't think of Counter. still pretty fast one I have to say that reading the problem statement is slower than writing the code. so it makes sense the guy that used llm to solve it would have got it in a few seconds. llms seem quite capable at reading and extracting meaningful data from the text. so with it being simple tast and has plenty of examples, llms would be very much capable of solving this instantly.
Yeah, I think the rank 1 person solved automatically with ChatGPT or similar - it sounds like they weren't aware that this was against the rules and they already apologized for it.
@@x87-64 they briefly had their github description set to something like "If you're here from Advent of Code, I apologize for not reading the FAQ and won't do it again"
Hmm, the video quality looks oddly bad. I'm guessing I messed up one of the settings in my video editing pipeline (it has been 11 months since I last did it...), I'll try to fix it for future days!
In case anyone's curious: the issue was that I was exporting a 4k video with a target bitrate of 2 Mbps... I turned it up to 60 Mbps and quality issues seem to be gone (for day 2).
Your typing speed is wild
The speed at which you understand the problem is insane.
Thanks Paul Dano
classic
really good, looking forward for the second day
welcome back!
one day I will code like you, just not today! very impressive man
in your python setup, can't you write a piece of code at the end that automatically copies `ans` to clipboard? I'm pretty sure this can be done and it will save you a few seconds.
Yeah, I saw that ecnerwala does this, I probably should too I just haven't gotten around to it.
yeah there is pyperclip, which makes it pretty easy
What the heck...
I went away for 2 minutes to fetch a beverage, and when I came back, you were already done?!
I used set( ).intersection( ) with .count( ) didn't think of Counter. still pretty fast one
I have to say that reading the problem statement is slower than writing the code. so it makes sense the guy that used llm to solve it would have got it in a few seconds. llms seem quite capable at reading and extracting meaningful data from the text. so with it being simple tast and has plenty of examples, llms would be very much capable of solving this instantly.
Randomly looking through AoC solutions for today and ran across your channel. Hope you are doing well Neil!
Hey Nick! Good to hear from you, and yep I'm doing well 😄
How does he download the input file so fast?
I raised video quality to 2160p 4K but your camera is still at a quality from dial-up Internet era
Ah yeah, see the pinned comment, I accidentally exported at 2 Mbps 🥲
Day 2's quality should be better!
Cool now do it in assembly
Bro, rank 1 is 9 seconds and rank 2 is 54 seconds.
Thats a bit much different, isnt it?
Yeah, I think the rank 1 person solved automatically with ChatGPT or similar - it sounds like they weren't aware that this was against the rules and they already apologized for it.
@@nthistlethwaite They will get away with the first days, but with harder problems LLMs sometimes struggle!
@@nthistlethwaite Where did they apologize?
@@x87-64 they briefly had their github description set to something like "If you're here from Advent of Code, I apologize for not reading the FAQ and won't do it again"