DeepSeek V4.1 Flash mencapai 1 token/detik di MacBook dengan memori 24GB.
Ini merupakan peningkatan dari kecepatan aslinya karena streaming SSD dan pengoptimalan mesin inferensi. AI lokal juga terasa seperti eksperimen kreatif yang mengeksplorasi batasan antara perangkat keras dan perangkat lunak. melalui：@mfranz_on
“모든 AI가 같은 ‘나’를 기억하게 하자.”
Memmy의 슬로건은 아주 직관적입니다. Claude Code, Codex, Cursor, OpenCode, DeepSeek Harness를 오가다 보면 매번 프로젝트 배경부터 다시 설명해야 하죠. 이전에 무엇을 했는지, 어떤 결정을 내렸는지, 어디에서 막혔는지도요. Memmy는 로컬 AI
In an interview with Bloomberg, Karen Hao, the author of Empire of AI
Highlighted the significance of China’s open-source AI models, particularly DeepSeek, in shedding light on the inefficiencies in computing resources utilized by AI companies. Notably, China’s large AI models
GPT-6 Astra + Claude Fable 5.1 + DeepSeek 4.1 Flash for $0 😱
All in one app, no credit card needed free starter credits earn more daily (visits, tasks, reviews, referrals) switch models in one click how to use: 1. click and sign up 2. open the
Sudah sampai ujung arsip
Semua deepseek