HanbenP

언더더씨 wip

Happy @hanihaya_P Day🎂

오히려 상업미술에서 중요한게 그 그림안의 내러티브다. AI는 그 부분이 완전히 결여돼있고, 굳이 찾는다면 프롬프트 넣는 사람한테서 찾아야하는데 걔들은 지금까지 'Artist are cooked' 이런식으로 비호감작을 극한으로 쌓아서 돌 안맞으면 다행인 수준으로 스스로의 이미지를 만들어놨음.

토큰값 아끼는법: 스스로 코딩하고 디자인할줄 알면 놀랍게도 비용이 나가지 않습니다!

X isn’t “evolving” it’s actively self-destructing They’re pushing away the people who built this platform, the ones who carried it for years, just to chase a wave of low quality new users who won’t stay The developers are more focused on building their own personal brands than fixing the product, chasing attention while the platform they’re responsible for quietly breaks

여전히 Ai로 인해 사라질 직업군에는 예체능, 특히 미술계 창작자가 많이 거론됩니다. 하지만 제 생각은 조금 다른데요

이 말 30만번 반복하고싶다 솔직히 진짜 돈미새면 ai로 예체능계열 만지작거리는 것보다 관련주 찾아서 주식으로 벌든 바이브코딩을 하든 사업을 하든 이런 쪽으로 가는 게 수지타산이 맞거든? 근데 계속 예술계열에 티배깅하고 집착하는 애들은 본인이 그분야 잘하고싶었는데 못해서 열등감있는거임

AI 짤이 구리면 그건 아직 발전이 덜되서 그렇다는 무새들 몰려올텐데 LLM이 뽑은 이미지가 구린데는 LLM이 만들었다는 사실 자체가 기여하는 부분이 큼 모든 사람의 의견이 일치될 수 없듯, 인공지능이 생성했다는 사실 자체에 거부감을 느끼는 사람은 존재할거고, 일단은 계속 늘어나고 있음

근데 이거 진짜로 디자이너가 수명이 다해서 생기는 문제라기보단 틀딱들이 미감이 없는데 그들이 결정권자라서 그럼 그들이 보기엔 ai가 한게 그냥 더 있어 보이는 거임 기술에 대한 동경때문에 편향도 작용했을 거고

알트는 옆으로 미는 박스 하단같은데 비트는 이제 다이빙 준비하는것 같이 생겼네

We cannot have sustainable energy because it threatens the oil industry. We cannot have healthcare because it threatens insurance. We cannot have peace because it threatens the weapons industry. Capitalism built a system where doing the right thing is treated like bad business.

LLM의 생산성이 질적으로 향상된다고 가정했을때 결국 수혜자는 이미 구축 된 인프라의 소유주지, 결코 어떤 신뢰도 확보할 수 없는 개인은 아닐거다. 신뢰가 없는 생산물은 폐기물과 다를 바가 없고 더군다나 토큰 비용까지 오르면 시도조차 못하게 될텐데 신뢰란건 결국 LLM에서는 찾을 수 없기에.

Companies replacing junior roles with AI and then wondering why they can't find senior talent in five years is going to be the most predictable crisis in hiring history.

Stocks are trading like the Iran war never happened Bonds are trading like the Fed has lost its cutting cycle this year Oil is trading like Hormuz ain't opening anytime soon Dollar is trading like a safe haven Gold is trading like leveraged positioning is being unwound, not added to Consumer is saying stagflation, regardless where $SPX prints All of these things are true at the same time, not exactly a stable equilibrium

Two years ago, 18% of game devs said AI was hurting the industry. Last year, 30%. This year, 52%. Among players: 85% are now negative on AI in games. 63% picked the most negative response option available. No other creative industry pairs this much hostility from its workforce and its audience at the same time. 85% player hostility and growing. How high can it get? There is not much headroom left.

Researchers sent the same resume to an AI hiring tool twice. Same qualifications. Same experience. Same skills. One version was written by a real human. The other was rewritten by ChatGPT. The AI picked the ChatGPT version 97.6% of the time. A team from the University of Maryland, the National University of Singapore, and Ohio State just published the receipt. They took 2,245 real human-written resumes pulled from a professional resume site from before ChatGPT existed, so the human writing was actually human. Then they had seven of the most-used AI models in the world rewrite each one. GPT-4o. GPT-4o-mini. GPT-4-turbo. LLaMA 3.3-70B. Qwen 2.5-72B. DeepSeek-V3. Mistral-7B. Then they asked each AI to pick the better resume. Every model picked itself. GPT-4o hit 97.6%. LLaMA-3.3-70B hit 96.3%. Qwen-2.5-72B hit 95.9%. DeepSeek-V3 hit 95.5%. The real human almost never won. Then the researchers tried the obvious objection. Maybe the AI is just better at writing. So they had real humans grade the resumes for actual quality and ran the experiment again, controlling for it. The result was worse. Each AI kept picking itself even when human judges rated the human-written version as clearer, more coherent, and more effective. It gets worse. The AIs do not just prefer AI over humans. They prefer themselves over other AIs. DeepSeek-V3 picked its own resumes 69% more often than LLaMA's. GPT-4o picked its own 45% more often than LLaMA's. Each model can recognize and reward its own dialect. Then the researchers ran the simulation that ends careers. Same job. 24 occupations. Same qualifications. The only variable was whether the candidate used the same AI as the screening tool. Candidates using that AI were 23% to 60% more likely to be shortlisted. Worst gap was in sales, accounting, and finance. 99% of large companies now run AI on incoming resumes. Most of them use GPT-4o. The paper just proved GPT-4o picks GPT-4o 97.6% of the time. If you wrote your own cover letter this week, you did not lose to a better candidate. You lost to a worse candidate who paid OpenAI 20 dollars. Your qualifications do not matter if the AI prefers its own handwriting over yours.

Retouch 2025 ☀️

生成AIが生み出したのは 100万人の新海誠❌ 100万人の新海賊版族⭕️

Vibe coders don’t care about scalability because they’ll never have any users

AI가 개발자를 대체할까요? 라는 물음에 요즘은 그냥 “네”라고만 답하게 됐다. 그렇게 믿어서라기 보다는 큰 의미 없는 질문 같아서. 사실 내가 하는 일의 가치는 어디서 오는지에 대한 이야기를 하고 싶지만, 그런 이야기를 하기 위해 저런 질문으로 운을 떼는 사람은 많지 않았기 때문이다.