慣性聚合 高效追讀感興趣之博客、新聞、科技資訊
閱原文 以慣性聚合開啟

推薦訂閱源

博客园 - 司徒正美
V
V2EX
T
Tailwind CSS Blog
有赞技术团队
有赞技术团队
aimingoo的专栏
aimingoo的专栏
Apple Machine Learning Research
Apple Machine Learning Research
IT之家
IT之家
Blog — PlanetScale
Blog — PlanetScale
A
About on SuperTechFans
月光博客
月光博客
T
The Blog of Author Tim Ferriss
宝玉的分享
宝玉的分享
Martin Fowler
Martin Fowler
博客园 - 聂微东
The GitHub Blog
The GitHub Blog
V
Visual Studio Blog
WordPress大学
WordPress大学
酷 壳 – CoolShell
酷 壳 – CoolShell
Engineering at Meta
Engineering at Meta
GbyAI
GbyAI

Hacker News - Newest: "LLM"

The LLM Is Not a Junior Engineer GitHub - pleasedodisturb/llm-safe-haven: The missing security guide for solo developers running autonomous AI coding agents GitHub - vassiliylakhonin/agenda-intelligence-md: A markdown protocol for AI agents that analyze public agenda instead of summarizing it badly. Firefox for Web Developers (@firefoxwebdevs@mastodon.social) Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning Easy VM sandboxes for LLM agents on MacOS, Miami & Paris travel GitHub - TauricResearch/TradingAgents: TradingAgents: Multi-Agents LLM Financial Trading Framework LLM 0.32a0 is a major backwards-compatible refactor IKP — Incompressible Knowledge Probes The 6-Lever LLM Cost Stack: A Production Playbook (One Backfired, One I'd Reverse Today) GitHub - Javierlozo/llm-audit: Static analysis for TypeScript / JavaScript LLM-application code. OWASP LLM Top 10 at commit time. A complement to Semgrep's p/ai-best-practices for the TS/JS ecosystem. RFC 0010: Workflow Composition Extension | PromptPack GitHub - allocz/slm: zero-dependency TUI LLM chat GitHub - stevefan1999-personal/demcstify: Decompile Minecraft using Vineflower, reconstruct the code using LLM GitHub - victornominista/anp: The economic layer for agent-to-agent negotiation. Binary protocol, Ed25519 identity, price oracle. GitHub - lazyville/qsh: q - Slim LLM CLI I built OWASP-style security skill packs for LLM apps (NPM install) Incompressible Knowledge Probes: Estimating Black-Box LLM Parameter Counts via Factual Capacity We Upgraded to a Frontier Model and Our Costs Went Down [PUBLIC] 03/21/26 vLLM-compile @ KCD vLLM meetup Ask HN: What happens when you paste a screenshot, and ask questions in LLM? GitHub - gfernandf/agent-skills: Agents should execute whenever possible — runtime for composable AI agent skills Understanding the LLM Bubble - American Affairs Journal LLM Budget Guard — Hard Cutoffs Before Your Agents Burn You GitHub - vnmoorthy/pavo-bench: A 50K-turn voice pipeline benchmark and an 85K-param meta-controller that cuts P95 latency 10.3% and energy 71% vs fixed cloud. TMLR 2026. GitHub - redcaller/voice-goat: A purposely vulnerable voice agent application for security practitioners to practice exploiting voice-based (and text based) AI systems. HeLa-Mem: Hebbian Learning and Associative Memory for LLM Agents Machina Mirabilis - Michael Hla GitHub - xoai/sage-wiki: An LLM-compiled personal knowledge base. Drop in your papers, articles, and notes. sage-wiki compiles them into a structured, interlinked wiki — with concepts extracted, cross-references discovered, and everything searchable. GitHub - MTimma/knowerage: Local MCP server that tracks AI analysis coverage against your codebase The Environment Rewrites the Question Before I Ask It GitHub - epscylonb/1386.ai.rocm: A lightweight transformer language model built from scratch in PyTorch, trained on a single consumer GPU with a full pipeline for data processing, pretraining, and instruction tuning. AeSlides: Incentivizing Aesthetic Layout in LLM-Based Slide Generation via Verifiable Rewards I build my LLM a Brain GitHub - glama-ai/lightport: A lightweight AI gateway that makes LLM providers OpenAI-compatible. Vibe-exploring Stunt Island Customizing Karpathy's LLM wiki for fighting disease AI Usage Analytics – Real-time budget enforcement and PII redaction for LLM My Workflow for Understanding LLM Architectures Museum of Code: alphafold_2018 InterviewDeck — Master 50 LLM Interview Questions GitHub - fkyah3/opencode-fkyah3: The open source coding agent. GitHub - poinsettiaclg-gif/AETHER-core: The open-source core compiler for the AETHER Agent Reliability Framework. Replaces fuzzy prompts with strict Weighted Intent Token (WIT) vectors to prevent Context Rot. The New Linux Kernel AI Bot Uncovering Bugs Is A Local LLM On Framework Desktop + AMD Ryzen AI Max GitHub - rlops/rlix: Run more RL experiments. Wait less for GPUs. GitHub - lace-ai/gai: 🤖 GAI is a flexible Go library for building agent-style applications on top of LLMs GitHub - starface77/Neuro-Adaptive-Reasoning-Engine GitHub - artem-mangilev/ctxbrew: 📦 Ship & Use AI-friendly package context. Could creativy in LLM emerge by reframing language? Cowork on 3P: How to Run Any LLM in Claude Cowork and Claude Code
华夏于大语言模型之赛落伍,然犹可胜于人工智能,前腾讯AI主管言。
01-_- · 2026-05-24 · via Hacker News - Newest: "LLM"

曩岁,劉偉與華人科技巨擘騰訊控股之人工智能之業相關。彼為深澤之學者,亦為騰訊之 Hunyuan 隊主,此乃該公司為生成式人工智能之世所立之基礎模型開發單位也。

然于二零二四年末,刘氏去腾讯,逾八载,众遂纷议其故。 Hunyuan 乃前岁方引,今刘氏何故骤去华夏财雄之极者,方于 AI 之盛?

对南华晨报,刘氏所言屡称 fanshi,或谓“范式”。AI研习者常用此语,指技术之突破,足以界定AI创新之新纪元,卓然者如OpenAI之ChatGPT、Anthropic之Claude Code是也。

刘氏则谓,之缺乏, fanshi。之新创,实为中国人工智能之最大阿喀琉斯之踵。“中国企业或仿DeepSeek,或效美国公司,于核心技术层面,此言所指,特指大型语言模型(LLM)之发展,此乃全球人工智能之争之重心。”

自去岁DeepSeek崭露头角以来,世人屡议中土大语言模型是否已追及美利坚之辈。然刘氏言,公论标尺分数渐趋窄狭,实难映真实应用之别。

虽美利坚之工业巨擘,犹砥砺术业之疆,尤以Anthropic之Mythos模型四月问世为显例,然华夏本土之DeepSeek领袖,则未能复其V4模型昔时之盛。