JKSHARK

AI 관련 일 하구 있어요

[JKSHARK @JKSHARK3]

@Angaisb_ I won't risk using Chinese model for that price difference with gpt5.6 or Fable. Chinese model was competitive b/o their extremely cheaper price!

[JKSHARK @JKSHARK3]

Similar K3 생각보다 엄청 비싸네? 흠....2.8T 모델 정도인거 감안하면 당연히 동일 수준의 opus4.8 보다 살짝 싼 수준이긴 한데. 중국산 모델 인거 생각하면 겨우 이 정도 차이나는걸론 Moonshot꺼 안 쓸듯. 중국애들이 뭔 짓 할줄 알고?

[JKSHARK @JKSHARK3]

@jun_song If this is true then who would use fable5 or gpt 5.6 sol? They cost multiple than Chinese models. I am so worried if Chinese AI would crash US ai economy.

[Jun Song @jun_song]

Upcoming models performance predictions :

Opus-5 = Slightly worse than Fable, similar to GPT-5.6
Kimi-K3 = Very fast, slightly worse than Fable (tested)
Deepseek-V4 = Opus-4.8
GLM-5.5 = Fable
MiniMax-M3 Pro = Slightly worse than Fable
Gemini-3.5-Pro = Only good at benchmarks

[JKSHARK @JKSHARK3]

@Multi_Serio_Ai @devpost 오 감사합니다 좋은 정보네요 크레딧 받아먹기로도 좋을듯 ㅋㅋ

[JKSHARK @JKSHARK3]

@SimyungChang 아하 감사합니다

[JKSHARK @JKSHARK3]

@SimyungChang 오오옹!!! 재미로 "Sk하이닉스 살까말까" 라고 물어봤는데 agentnews.md/finance-ko 읽으라고 했을때 답이 훨 정확한데요?? 아니였을땐 그냥 아무 최근 뉴스 긁어서 ADR 상장이랑 미국 CPI 얘기만 하는데 agentnews.md 읽고 하니까 "옵션 레버리지 가 올렸다"는 평가도 하네요 ㅋㅋ

[JKSHARK @JKSHARK3]

@SimyungChang 오 ㄷㄷㄷㄷ함 해보겠습니다

[JKSHARK @JKSHARK3]

@editxshub @Bhavani_00007 Hahaha nice one

[JKSHARK @JKSHARK3]

@thsottiaux Cooool. Anthropic needs to stop marketing Claude like it’s Skynet. We’re sick and tired of the safety-theater hype cycle. I like OpenAI's policy much better

[JKSHARK @JKSHARK3]

실시간 IFC 시위 현장...홈플러스가 회생에 필요한 긴급운영자금 2,000억 원을 못 구해 파산 위기에 놓이자, 노조가 최대 채권자인 메리츠를 대상으로 시위하는거. 반면 메리츠는 경영 실패가 최대주주 MBK에 있다면서 서로 책임 떠넘기기 하는중 ㅋㅋㅋ

[JKSHARK @JKSHARK3]

gpt 하네스 에 대한 원성이 자자하군.. 토큰 빨아들이는 기계라나. 별 태스크도 아닌데 백그라운드로 계속 subagent 돌고 있고 등등... 하네스가 이렇게나 중요하다! 차라리 claude 하네스 에 gpt 5.6 sol 붙인게 더 낫다는 평이 훨 많음

[JKSHARK @JKSHARK3]

@ddururiiiiiii 직원들 토큰이라도 쓰게 해주지..쿼타 주고..비싸서 부담되는건 이해되는데 그럼 클로드라고 못 박지나 말든갸

[Simyung Chang @SimyungChang]

How to use AgentNews:

Open your AI assistant and paste:

“Read https://t.co/ctuov5VsXh first.
Then explain the main scenarios for US stocks this week.”

For agents that read markdown directly:

https://t.co/CXS0P0KdaN

Then ask:

What changed?
What is contested?
What would falsify this view?

AgentNews is not another finance chatbot.
It is live context your AI can read before it searches, reasons, or answers.

[JKSHARK @JKSHARK3]

@SimyungChang So is it like giving weights to let the AI which question is more important? Hmm interesting... i wonder why can't they direct themselves to the right question in the first place. Maybe lack of experiences? Haha. The anyway

[JKSHARK @JKSHARK3]

@SimyungChang @garrytan True true. What's the point of pouring more GPUs into training bigger models if they still can't reliably use everyday software? They spend half their time recovering from errors instead of getting work done. Maybe we should focus on harnesses for once.

[JKSHARK @JKSHARK3]

@SimyungChang So you're arguing that "context should be treated as infrastructure rather than something every agent reconstructs independently". But with modern RAG pipelines and built-in web search, aren't today's models already pretty good at building that context from external news?

[Simyung Chang @SimyungChang]

Ask an AI what matters for US stocks this week.

It can search.
But it often still misses the market's "now."

What is leading?
What is contested?
What changed?
What would prove the frame wrong?

That's Nowless Search.

So we built AgentNews:
a falsifiable macro context board for AI agents.