논문해외HF Daily Papers
D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding
Target-conditioned diffusion drafters predict candidate tokens in parallel, yet each new block must wait for updated target-model hidden states.
실시간
국내외 뉴스 · 기업 공식 발표 · 논문 · 영상 · 커뮤니티 소식을 한국시간 기준으로 제공합니다. 같은 소식은 하나로 묶고, 기본은 핵심만 보여 드립니다.
모든 수집 소식(구글 뉴스로 들어온 단건 보도 포함).
Target-conditioned diffusion drafters predict candidate tokens in parallel, yet each new block must wait for updated target-model hidden states.
The vibe-coded addon problem is getting so bad that Warcraft Logs is changing its API and developers are dealing with theft of their work.
명지전문대학(총장 권두승)이 지역 주민들이 직접 인공지능(AI) 로봇을 제작해 생활 속 문제를 해결하고, 중남미 학생들과 아이디어를 공유하는 글로벌 AI 해커톤을 개최한다.
Speculative decoding has been treated purely as a systems optimization that speeds up large language model decoding without changing the model behavior.
When a developer gives an AI agent a coding task, the agent does more than answer once.
Qodo brings company knowledge and coding standards into Codex, helping developers catch P0 issues across repositories before opening a pull request.
Recently, harness self-evolution has emerged as an important paradigm for improving agent performance.
Speculative decoding accelerates memory-bound LLM decoding with a lightweight drafter, and parallel drafters extend this speedup by drafting a whole block in…
Speculative decoding (SD) accelerates large language model inference by reducing the sequential cost of token-by-token generation.
LLM agents for GPU kernel optimization sit at two ends of a structure-freedom tradeoff.
As coding agents achieve increasingly high success rates on software engineering benchmarks, recent efforts have sought to increase task complexity through…
Coding agents solve repository issues through long sequences of exploration, editing, and validation.
Software development sessions often involve several unfinished tasks rather than one issue carried through to completion.
Amazon Bedrock AgentCore Session endpoints give each runtime session an authenticated URL that forwards any HTTP method, path, and WebSocket connection to the…
Host Sowbhagya Suresh shares some of the latest news and cutting-edge research from across Microsoft—from understanding the challenges facing coding agents to…
Mistral opens Large 4 weights for coding and security
Mistral Large 4의 공개 가중치는 코딩과 사이버 성능 향상을 제공하며 자체 호스팅이 가능하다.
The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.
Coding agents increasingly rely on automated approvers to decide whether a proposed shell command may run.
DyCraft: An Agentic System for Long-Horizon 4D Scene Generation
기존 코딩 에이전트는 고품질 3D 장면을 만들지만 물리적으로 불가능한 결과물을 내는 등 장기 애니메이션 생성에 어려움을 겪는다.
I've written before about how November 2025 was an inflection point for coding agents - the point where Opus 4.5 and GPT-5.1 crossed a tipping point that meant…
Since coding is solved, why code for the job?
Over 11 hours of coding videos on evaluation, inference-time scaling, reinforcement learning, and distillation, with Python and PyTorch.
Recursive releases Flint, 40B model for local coding
Recursive의 Flint는 오픈 웨이트, 512k 토큰 컨텍스트 창, Apple Silicon 최적화 추론 엔진을 바탕으로 빠른 로컬 코딩을 지원한다.
Google is now rolling out support for Gemini 4 Argon to Antigravity, its AI coding tool. more…
#Google Gemini#Google · 구글#AI 코딩
다른 보도 Matthew Berman