crackthrough @crackthrough

🇰🇷 23yo, he/him (or at least I think so I can't count) A developer in 7th Beat Games, https://t.co/aEZSTNSnXj Representative. 이것저것 퍼먹는 얼불춤 개발자입니다. 잘 부탁드려요. cooki.dev Joined March 2019

Tweets

1K
Followers

258
Following

212
Likes

3K

crackthrough @crackthrough

a month ago

@bwenkun @jofuture_sdvx 이게머꼬;;

0 0 0 30 0

View Details

Merong @mer0n_x

a month ago

Laur - Gears of Fate (Fan M/V) #Blender #AfterEffects #Motiongraphics

3 25 150 4K 22

0 0 0 73 0

View Details

crackthrough @crackthrough

a month ago

@Basix1120 오...........

0 0 0 103 0

View Details

crackthrough @crackthrough

3 months ago

shoutout to @Ssimille577341 for his creations 여러분, 씨밀레 님의 멋진 작업물을 감상해주세요 youtube.com/watch?v=CE7Mwv…

0 0 4 251 0

View Details

crackthrough @crackthrough

5 months ago

오 오 AWC 2026 Semi-Finals

0 0 4 159 0

View Details

파차🕯 (@[email protected]) @finalchildmc

7 months ago

암호화 직접 짜지 마세여 여러분

0 18 34 5K 0

View Details

She dumped me last night. Not because I don't listen. Not because I'm always on my phone. Not even because I forgot our anniversary (twice). But because, in her exact words: "You only pay attention to the parts of what I say that you think are important." I stared at her for a moment and realized... She just perfectly described the attention mechanism in transformers. Turns out I wasn't being a bad boyfriend. I was being mathematically optimal. See, in conversations (and transformers), you don't give equal weight to every word. Some words matter more for understanding context. Attention figures out exactly HOW important each word should be. Here's the beautiful math: Attention(Q, K, V) = softmax(QK^T / √d_k)V Breaking it down: Q (Query): "What am I looking for?" K (Key): "What info is available?" V (Value): "What is that info?" d_k: Key dimension (for scaling) Think library analogy: You have a question (Query). Books have titles (Keys) and content (Values). Attention finds which books are most relevant. Step-by-step with "The cat sat on the mat": Step 1: Create Q, K, VEach word → three vectors via learned matrices W_Q, W_K, W_V For "cat": Query: "What should I attend to when processing 'cat'?" Key: "I am 'cat'" Value: "Here's cat info" Step 2: Calculate scoresQK^T = how much each word should attend to others Processing "sat"? High similarity with "cat" (cats sit) and "mat" (where sitting happens). Step 3: Scale by √d_kPrevents dot products from getting too large, keeps softmax balanced. Step 4: SoftmaxConverts scores to probabilities: "cat": 0.4 (subject) "sat": 0.3 (action) "mat": 0.2 (location) "on": 0.1 (preposition) "the": 0.1 (article) Step 5: Weight valuesMultiply each word's value by attention weight, sum up. Now "sat" knows it's most related to "cat" and "mat". Multi-Head Magic:Transformers do this multiple times in parallel: Head 1: Subject-verb relationships Head 2: Spatial ("on", "in", "under") Head 3: Temporal ("before", "after") Head 4: Semantic similarity Each head learns different relationship types. Why This Changed Everything: Before: RNNs = reading with flashlight (one word at a time, forget the beginning) After: Attention = floodlights on entire sentence with dimmer switches This is why ChatGPT can: Remember 50 messages ago Know "it" refers to something specific Understand "bank" = money vs river based on context The Kicker:Models learn these patterns from data alone. Nobody programmed grammar rules. It figured out language structure just by predicting next words. Attention is how AI learned to read between the lines. Just like my therapist helped me understand my focus patterns, maybe understanding transformers helps us see how we decide what matters. Now if only I could implement multi-head attention in dating... Still waiting for "scaled dot-product listening" to be invented.

187 493 7K 701K 2K

View Details

crackthrough @crackthrough

7 months ago

@g333choi1 오투잼

0 0 2 32 0

View Details

crackthrough @crackthrough

8 months ago

[enum.item] = new item { ... }로 바꾸고 ctrl alt L 쓰니까 해결됨 ㅎㅎ;;

0 0 0 179 0

View Details

crackthrough @crackthrough

8 months ago

c# indentation standard ㄹㅇ이해안가네 왜 이렇게만들었지 프로젝트마다 .editorconfig 바꾸는거 그냥 열받음

2 0 0 246 0

View Details

crackthrough @crackthrough

8 months ago

@rpip1020 제게 영감을 주셔서 감사합니다

1 0 0 27 0

View Details

crackthrough @crackthrough

8 months ago

@getnachonacho Sounds like a concrete rule! Do you happen to have any examples I can learn from?

0 0 0 14 0

View Details

crackthrough @crackthrough

8 months ago

shower thought: where is the appropriate place to put a paywall on SaaS? 갑자기 궁금) SaaS는 어디에 유료화 기능 추가해야 욕을 안 먹을까요

2 0 0 177 0

View Details

crackthrough @crackthrough

8 months ago

x.com/stupidtechtake… 이거 보고 갑자기 ssotax.org 생각나서 올렸어요

stupid tech takes @stupidtechtakes

8 months ago

is it just me or is it hilarious to paywall a colour

46 50 3K 91K 88

0 0 0 220 0

View Details

crackthrough @crackthrough

8 months ago

@cherry__k1m You have unlock the song! 😭

0 0 0 930 0

View Details

crackthrough @crackthrough

8 months ago

@Gur__vi looks cool! but i would like to cite emilkowal.ski/ui/you-dont-ne… > I don’t expect to be “delighted”, I don’t need to be. IT DEPENDS ON THE CASE, but users are already used to input fields with immediate feedback, so it's probably not a good idea to *always* have animations there

0 1 30 2K 12

View Details

crackthrough @crackthrough

8 months ago

@Basix1120 ㅜㅜ 조심히 다녀오세요

0 0 0 37 0

View Details

crackthrough @crackthrough

8 months ago

@theeluwin 해야 했던 경험이 당연해서, 또는 그런 라이브러리를 만들거나 사용하기 위한 노력을 거의 모든 개발사가 하는게 norm이 아니어서가 아니었을까 추측해봅니다.

0 0 1 13 0

View Details

crackthrough @crackthrough

8 months ago

@theeluwin 창의성을 중요하게 생각하지 않고 그리 자료가 많지 않던 시절, 추상적인 비즈니스 로직을 최대한 문제 없이 구현하는 훈련이지 않을까 싶습니다. 당시 지금처럼 세상이 라이브러리 떡칠이 되리라고는 예상하지 못했을 거고, 트리, 벡터 등의 구조는 많은 양의 데이터 처리를 라이브러리 없이 구현