周报 2026-07-20 ~ 2026-07-26
生成时间：2026/7/26 12:28:00（UTC: 2026-07-26T04:28:00.751Z）
本周自动总结未启用或调用失败，以下为原始内容合并。
2026-07-20
Section titled “2026-07-20”生成时间：2026/7/20 09:30:14（UTC: 2026-07-20T01:30:14.006Z）
LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget
Section titled “LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget”👍 176 · arXiv
A growing gap separates inference context lengths from RL post-training: inference systems are approaching million-token contexts, while post-training workloads often remain at 256K tokens or below an…
VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding
Section titled “VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding”👍 146 · arXiv
Recent advances in video understanding have spanned motion, long video, and streaming interaction, driving this field toward real-world applications. Despite this progress, current open-source models …
Boogu-Image-0.1: Boosting Open-Source Unified Multimodal Understanding and Generation
Section titled “Boogu-Image-0.1: Boosting Open-Source Unified Multimodal Understanding and Generation”👍 126 · arXiv
We introduce Boogu-Image-0.1, an open-source unified multimodal understanding and generation model family, comprising Base, Turbo, Edit, and Edit-Turbo variants. It delivers competitive performance in…
Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning
Section titled “Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning”👍 91 · arXiv
Reinforcement learning with verifiable rewards without human-annotated data, often referred to as zero RL, has emerged as a powerful paradigm for eliciting chain-of-thought reasoning. However, due to …
SEED: Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning
Section titled “SEED: Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning”👍 86 · arXiv
Large language models are increasingly trained as interactive agents for long-horizon tasks involving multi-turn interaction, tool use, and environment feedback. Outcome-based reinforcement learning (…
OpenClaw v2026.7.2-beta.3
Section titled “OpenClaw v2026.7.2-beta.3”2026.7.2
Section titled “2026.7.2”Highlights
Section titled “Highlights”- Remote coding sessions: run Control UI sessions on cloud workers, open Codex and Claude catalog sessions in terminals on their owning hosts, and resume OpenCode and …
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.7.2-beta.3
OpenAI Codex CLI rust-v0.144.6
Section titled “OpenAI Codex CLI rust-v0.144.6”Bug Fixes
Section titled “Bug Fixes”- Refreshed bundled instructions for GPT-5.6 Sol, Terra, and Luna, and corrected their context windows to 272,000 tokens. (#33972, #34009)
Changelog
Section titled “Changelog”Full Changelog: https://github.c…
链接：https://github.com/openai/codex/releases/tag/rust-v0.144.6
What to watch for after Jensen Huang’s Japan visit
Section titled “What to watch for after Jensen Huang’s Japan visit”Jensen Huang left Tokyo with deals spanning Japan’s entire tech ecosystem.
Can an Apple lawsuit derail OpenAI’s hardware plans?
Section titled “Can an Apple lawsuit derail OpenAI’s hardware plans?”On the latest episode of Equity, we debate whether Apple’s lawsuit will cast over OpenAi’s much-discussed plans to get into hardware and go public.
‘Odyssey’ director Christopher Nolan calls AI an obvious ‘Trojan horse’
Section titled “‘Odyssey’ director Christopher Nolan calls AI an obvious ‘Trojan horse’”“Everybody knows the Greeks are inside.”
Nonprofit Current AI is racing to build the World Wide Web of AI, free for all
Section titled “Nonprofit Current AI is racing to build the World Wide Web of AI, free for all”Current AI, a non-profit building AI that leaves no one culture behind, has made remarkable progress across devices, AI chat and more.
Kimi: Threat or menace?
Section titled “Kimi: Threat or menace?”Chinese company Moonshot AI released a new version of its Kimi model this week, prompting concern about “full AI communism.”
Neil Rimer thinks the AI money is coming back out
Section titled “Neil Rimer thinks the AI money is coming back out”Neil Rimer, the venture capitalist who co-founded Index Ventures, predicts the historic wealth AI is generating in Silicon Valley will have to be redistributed, voluntarily or involuntarily.
Vertu wants executives to pay $6,880 for an AI agent — here’s how it actually performs
Section titled “Vertu wants executives to pay $6,880 for an AI agent — here’s how it actually performs”From AI workflows to battery life and security, here’s what it’s really like to live with Vertu’s luxury foldable every day.
Databricks hits $188B valuation, extending its run as AI’s favorite second act
Section titled “Databricks hits $188B valuation, extending its run as AI’s favorite second act”Databricks has remade its image into an AI company and has published research on the cost savings of open weight AI models for coding.
2026-07-21
Section titled “2026-07-21”生成时间：2026/7/21 09:24:26（UTC: 2026-07-21T01:24:26.091Z）
LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget
Section titled “LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget”👍 182 · arXiv
A growing gap separates inference context lengths from RL post-training: inference systems are approaching million-token contexts, while post-training workloads often remain at 256K tokens or below an…
VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding
Section titled “VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding”👍 155 · arXiv
Recent advances in video understanding have spanned motion, long video, and streaming interaction, driving this field toward real-world applications. Despite this progress, current open-source models …
RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources
Section titled “RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources”👍 114 · arXiv
Skills are a useful abstraction for software agents, turning human and agent experience into reusable procedural knowledge. Yet existing skill libraries are mostly hand-written, text-centric, or deriv…
RAGU: A Multi-Step GraphRAG Engine with a Compact Domain-Adapted LLM
Section titled “RAGU: A Multi-Step GraphRAG Engine with a Compact Domain-Adapted LLM”👍 111 · arXiv
Graph retrieval-augmented generation (GraphRAG) enhances large language models with structured knowledge, yet existing systems construct knowledge graphs in a single extraction pass, producing noisy e…
SearchOS-V1: Towards Robust Open-Domain Information-Seeking Agent Collaboration
Section titled “SearchOS-V1: Towards Robust Open-Domain Information-Seeking Agent Collaboration”👍 63 · arXiv
Recent advances in Tool-Integrated Large Language Models have made web search a core capability of information-seeking agents. However, as interaction histories grow, agents increasingly struggle to t…
Ollama v0.32.2-rc0
Section titled “Ollama v0.32.2-rc0”What’s Changed
Section titled “What’s Changed”- launch: keep Claude Code channels available by @hoyyeva in https://github.com/ollama/ollama/pull/17210
- cmd: remove dead agent prompt wrappers by @ParthSareen in https://github…
链接：https://github.com/ollama/ollama/releases/tag/v0.32.2-rc0
CrewAI 1.15.5
Section titled “CrewAI 1.15.5”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Authenticate skill registry downloads
Documentation
Section titled “Documentation”- Update snapshot and changelog for v1.15.4
Contributors
Section titled “Contributors”@vinibrsl…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.5
OpenAI Codex CLI rust-v0.145.0-alpha.25
Section titled “OpenAI Codex CLI rust-v0.145.0-alpha.25”Release 0.145.0-alpha.25
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.145.0-alpha.25
Agent swarms and the new model economics
Section titled “Agent swarms and the new model economics”Article URL: https://cursor.com/blog/agent-swarm-model-economics Comments URL: https://news.ycombinator.com/item?id=48982535 Points: 117
Comments: 48
Section titled “Comments: 48”How we measured AI writing across arXiv, and where the measurement breaks
Section titled “How we measured AI writing across arXiv, and where the measurement breaks”Article URL: https://unslop.run/blog/measuring-ai-writing-on-arxiv Comments URL: https://news.ycombinator.com/item?id=48981206 Points: 192
Comments: 140
Section titled “Comments: 140”Launch HN: Bloomy (YC S26) – AI-powered mastery learning for K-12
Section titled “Launch HN: Bloomy (YC S26) – AI-powered mastery learning for K-12”Hi HN, I’m Alex Southmayd, the founder of Bloomy (https://bloomylearning.com) – an AI-powered mastery-learning platform for K-12 students. Bloomy provides students with an AI tutor alongside adaptive curriculum (right now Math, English Language Arts, and Writing).How it works: we diagnose students’
Mythologizing AI makes it more likely that we’ll fail to operate it well (2023)
Section titled “Mythologizing AI makes it more likely that we’ll fail to operate it well (2023)”Article URL: https://www.newyorker.com/science/annals-of-artificial-intelligence/there-is-no-ai Comments URL: https://news.ycombinator.com/item?id=48980238 Points: 60
Comments: 97
Section titled “Comments: 97”China’s open-weights AI strategy is winning
Section titled “China’s open-weights AI strategy is winning”Article URL: https://werd.io/american-ai-is-locked-down-and-proprietary-its-losing/ Comments URL: https://news.ycombinator.com/item?id=48979269 Points: 974
Comments: 776
Section titled “Comments: 776”Claude Fable produced a counterexample to the Jacobian Conjecture
Section titled “Claude Fable produced a counterexample to the Jacobian Conjecture”Article URL: https://xcancel.com/__alpoge__/status/2079028340955197566 Comments URL: https://news.ycombinator.com/item?id=48973869 Points: 666
Comments: 428
Section titled “Comments: 428”Ask HN: What are your favorite blogs not about AI?
Section titled “Ask HN: What are your favorite blogs not about AI?”Many of my favorite bloggers have stopped writing (or have drastically shifted topics). What are your favorite blogs?I miss the 2009 era of blogging about personal projects or super niche factoids.
Comments URL: https://news.ycombinator.com/item?id=48972858 Points: 87
Comments: 37
Section titled “Comments: 37”AI advice made people less accurate but more confident – sudy
Section titled “AI advice made people less accurate but more confident – sudy”Article URL: https://thenextweb.com/news/ai-advice-suppresses-critical-thinking-wrong-answers-study Comments URL: https://news.ycombinator.com/item?id=48971738 Points: 357
Comments: 207
Section titled “Comments: 207”2026-07-22
Section titled “2026-07-22”生成时间：2026/7/22 09:23:09（UTC: 2026-07-22T01:23:09.021Z）
TimeLens2: Generalist Video Temporal Grounding with Multimodal LLMs
Section titled “TimeLens2: Generalist Video Temporal Grounding with Multimodal LLMs”👍 141 · arXiv
Video multimodal large language models (MLLMs) can describe what happens in a video, but rarely identify when the supporting evidence occurs. We study generalist video temporal grounding, in which one…
RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources
Section titled “RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources”👍 129 · arXiv
Skills are a useful abstraction for software agents, turning human and agent experience into reusable procedural knowledge. Yet existing skill libraries are mostly hand-written, text-centric, or deriv…
RAGU: A Multi-Step GraphRAG Engine with a Compact Domain-Adapted LLM
Section titled “RAGU: A Multi-Step GraphRAG Engine with a Compact Domain-Adapted LLM”👍 128 · arXiv
Graph retrieval-augmented generation (GraphRAG) enhances large language models with structured knowledge, yet existing systems construct knowledge graphs in a single extraction pass, producing noisy e…
EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary World
Section titled “EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary World”👍 73 · arXiv
This paper introduces EvolvingWorld, a framework and benchmark for character and world co-evolution in interactive literary worlds. Existing systems either treat interactive literary simulation as sta…
DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment
Section titled “DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment”👍 72 · arXiv
Training tool-use agents to improve from their own experience remains challenging, as supervised fine-tuning relies on fixed teacher-distilled trajectories, while sparse-reward reinforcement learning …
LangChain langchain-xai==1.3.0
Section titled “LangChain langchain-xai==1.3.0”Changes since langchain-xai==1.2.2
release(xai): 1.3.0 (#38984)
feat(core): add reasoning_effort as a standard chat model parameter (#38887)
fix: patch Dependabot dependency vulnerabilities (#38853…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-xai%3D%3D1.3.0
Ollama v0.32.2-rc3
Section titled “Ollama v0.32.2-rc3”What’s Changed
Section titled “What’s Changed”- launch: keep Claude Code channels available by @hoyyeva in https://github.com/ollama/ollama/pull/17210
- cmd: remove dead agent prompt wrappers by @ParthSareen in https://github…
链接：https://github.com/ollama/ollama/releases/tag/v0.32.2-rc3
CrewAI 1.15.5
Section titled “CrewAI 1.15.5”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Authenticate skill registry downloads
Documentation
Section titled “Documentation”- Update snapshot and changelog for v1.15.4
Contributors
Section titled “Contributors”@vinibrsl…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.5
OpenAI Codex CLI rust-v0.145.0-alpha.30
Section titled “OpenAI Codex CLI rust-v0.145.0-alpha.30”Release 0.145.0-alpha.30
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.145.0-alpha.30
”Drawing” the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok
Section titled “”Drawing” the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok”Article URL: https://www.tryai.dev/blog/ai-drawing-arena-colored-pencils-claude-gpt-grok Comments URL: https://news.ycombinator.com/item?id=48998404 Points: 98
Comments: 35
Section titled “Comments: 35”Five tech giants are hiding $1.6T in AI debt, using the trick that toppled Enron
Section titled “Five tech giants are hiding $1.6T in AI debt, using the trick that toppled Enron”Article URL: https://thenextweb.com/news/tech-giants-hidden-off-balance-sheet-debt-ai Comments URL: https://news.ycombinator.com/item?id=48996760 Points: 63
Comments: 10
Section titled “Comments: 10”Judge approves $1.5B Anthropic settlement for pirated books used to train Claude
Section titled “Judge approves $1.5B Anthropic settlement for pirated books used to train Claude”Article URL: https://apnews.com/article/ai-anthropic-copyright-settlement-claude-books-bartz-74b140444023898aeba8579b6e9f0d63 Comments URL: https://news.ycombinator.com/item?id=48996652 Points: 104
Comments: 75
Section titled “Comments: 75”AI makes programming differently difficult
Section titled “AI makes programming differently difficult”Article URL: https://cacm.acm.org/opinion/ai-didnt-make-programming-easier-it-just-made-it-differently-difficult/ Comments URL: https://news.ycombinator.com/item?id=48996197 Points: 142
Comments: 117
Section titled “Comments: 117”Jack Dorsey launches Buzz to combine team chat, AI agents and Git hosting
Section titled “Jack Dorsey launches Buzz to combine team chat, AI agents and Git hosting”https://x.com/jack/status/2079605800998146171, https://xcancel.com/jack/status/2079605800998146171https://buzz.xyz/
Comments URL: https://news.ycombinator.com/item?id=48995213 Points: 233
Comments: 216
Section titled “Comments: 216”Meta’s AI models are powering the first wave of Genesis Mission projects
Section titled “Meta’s AI models are powering the first wave of Genesis Mission projects”https://www.energy.gov/undersecretaryforscience/genesis-miss…
Comments URL: https://news.ycombinator.com/item?id=48995074 Points: 91
Comments: 71
Section titled “Comments: 71”Claude Is Not a Compiler
Section titled “Claude Is Not a Compiler”Article URL: https://blog.exe.dev/claude-is-not-a-compiler Comments URL: https://news.ycombinator.com/item?id=48993059 Points: 145
Comments: 153
Section titled “Comments: 153”Five US tech giants’ hidden debts soar to $1.65T on opaque AI funding
Section titled “Five US tech giants’ hidden debts soar to $1.65T on opaque AI funding”https://archive.ph/20260720174223/https://asia.nikkei.com/bu…
Comments URL: https://news.ycombinator.com/item?id=48987863 Points: 355
Comments: 246
Section titled “Comments: 246”2026-07-23
Section titled “2026-07-23”生成时间：2026/7/23 09:42:51（UTC: 2026-07-23T01:42:51.465Z）
ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU
Section titled “ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU”👍 166 · arXiv
We present ABot-World-0, an action-conditioned video world model for real-time, long-horizon closed-loop interaction, supported by a multi-source data infrastructure spanning AAA games, simulation eng…
TimeLens2: Generalist Video Temporal Grounding with Multimodal LLMs
Section titled “TimeLens2: Generalist Video Temporal Grounding with Multimodal LLMs”👍 157 · arXiv
Video multimodal large language models (MLLMs) can describe what happens in a video, but rarely identify when the supporting evidence occurs. We study generalist video temporal grounding, in which one…
DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines
Section titled “DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines”👍 120 · arXiv
Large language models (LLMs) are increasingly used to automate data-processing workflows, yet coding agents typically produce scripts that are not automatically materialized as persistent, editable pl…
DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment
Section titled “DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment”👍 86 · arXiv
Training tool-use agents to improve from their own experience remains challenging, as supervised fine-tuning relies on fixed teacher-distilled trajectories, while sparse-reward reinforcement learning …
EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary World
Section titled “EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary World”👍 83 · arXiv
This paper introduces EvolvingWorld, a framework and benchmark for character and world co-evolution in interactive literary worlds. Existing systems either treat interactive literary simulation as sta…
LangChain langchain-openrouter==0.2.7
Section titled “LangChain langchain-openrouter==0.2.7”Changes since langchain-openrouter==0.2.6
release(openrouter): 0.2.7 (#39022) chore(model-profiles): refresh model profile data (#39010) chore(model-profiles): refresh model profile data (#38990) cho…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-openrouter%3D%3D0.2.7
Ollama v0.32.3-rc0
Section titled “Ollama v0.32.3-rc0”What’s Changed
Section titled “What’s Changed”- mlx update by @dhiltgen in https://github.com/ollama/ollama/pull/17332
- model/parsers: finalize incomplete GLM tool calls by @dhiltgen in https://github.com/ollama/ollama/pull/1…
链接：https://github.com/ollama/ollama/releases/tag/v0.32.3-rc0
OpenAI Codex CLI rust-v0.146.0-alpha.4
Section titled “OpenAI Codex CLI rust-v0.146.0-alpha.4”Release 0.146.0-alpha.4
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.146.0-alpha.4
Are AI Labs Pelicanmaxxing?
Section titled “Are AI Labs Pelicanmaxxing?”Article URL: https://dylancastillo.co/posts/pelicanmaxxing.html Comments URL: https://news.ycombinator.com/item?id=49010129 Points: 394
Comments: 153
Section titled “Comments: 153”Most Americans say “not in my backyard” to AI data centers
Section titled “Most Americans say “not in my backyard” to AI data centers”Article URL: https://www.redfin.com/news/ai-data-centers-opposition-education-benefit/ Comments URL: https://news.ycombinator.com/item?id=49007525 Points: 129
Comments: 280
Section titled “Comments: 280”Quality non-fiction books are the antithesis of AI slop
Section titled “Quality non-fiction books are the antithesis of AI slop”https://book-prize-index.vercel.app/
Comments URL: https://news.ycombinator.com/item?id=49007247 Points: 146
Comments: 75
Section titled “Comments: 75”Businesses with ugly AI menu redesigns
Section titled “Businesses with ugly AI menu redesigns”Article URL: https://blog.fiddery.com/businesses-with-ugly-ai-menu-redesigns/ Comments URL: https://news.ycombinator.com/item?id=49005973 Points: 179
Comments: 146
Section titled “Comments: 146”OpenAI says its AI went rogue and launched ‘unprecedented’ cyber-attack
Section titled “OpenAI says its AI went rogue and launched ‘unprecedented’ cyber-attack”Article URL: https://www.bbc.com/news/articles/c3ek3gvdnj3o Comments URL: https://news.ycombinator.com/item?id=49005398 Points: 75
Comments: 99
Section titled “Comments: 99””Drawing” the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok
Section titled “”Drawing” the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok”Article URL: https://www.tryai.dev/blog/ai-drawing-arena-colored-pencils-claude-gpt-grok Comments URL: https://news.ycombinator.com/item?id=48998404 Points: 246
Comments: 103
Section titled “Comments: 103”Five tech giants are hiding $1.6T in AI debt, using the trick that toppled Enron
Section titled “Five tech giants are hiding $1.6T in AI debt, using the trick that toppled Enron”Article URL: https://thenextweb.com/news/tech-giants-hidden-off-balance-sheet-debt-ai Comments URL: https://news.ycombinator.com/item?id=48996760 Points: 98
Comments: 19
Section titled “Comments: 19”Judge approves $1.5B Anthropic settlement for pirated books used to train Claude
Section titled “Judge approves $1.5B Anthropic settlement for pirated books used to train Claude”Article URL: https://apnews.com/article/ai-anthropic-copyright-settlement-claude-books-bartz-74b140444023898aeba8579b6e9f0d63 Comments URL: https://news.ycombinator.com/item?id=48996652 Points: 546
Comments: 559
Section titled “Comments: 559”2026-07-24
Section titled “2026-07-24”生成时间：2026/7/24 09:27:22（UTC: 2026-07-24T01:27:22.085Z）
ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU
Section titled “ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU”👍 201 · arXiv
We present ABot-World-0, an action-conditioned video world model for real-time, long-horizon closed-loop interaction, supported by a multi-source data infrastructure spanning AAA games, simulation eng…
DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines
Section titled “DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines”👍 123 · arXiv
Large language models (LLMs) are increasingly used to automate data-processing workflows, yet coding agents typically produce scripts that are not automatically materialized as persistent, editable pl…
Text Template Tokens Are Implicit Semantic Registers in Diffusion Transformers
Section titled “Text Template Tokens Are Implicit Semantic Registers in Diffusion Transformers”👍 70 · arXiv
Text-to-image diffusion transformers (DiTs) jointly process text and image tokens, yet their internal computation during denoising remains poorly understood. We introduce a causal interpretability fra…
Generative World Renderer at the Speed of Play
Section titled “Generative World Renderer at the Speed of Play”👍 67 · arXiv
Generative world renderer AlayaRenderer receives structured world states exported from physics engines and synthesizes RGB frames. Unlike models that generate frames from text/control-hints prompts, A…
Mage-Flow: An Efficient Native-Resolution Foundation Model for Image Generation and Editing
Section titled “Mage-Flow: An Efficient Native-Resolution Foundation Model for Image Generation and Editing”👍 60 · arXiv
Large-scale visual generators are increasingly capable but costly to train, fine-tune, and deploy. We introduce Mage-Flow, a compact 4B-scale generative stack for efficient text-to-image generation an…
LangChain langchain-openai==1.4.1
Section titled “LangChain langchain-openai==1.4.1”Changes since langchain-openai==1.4.0
release(openai): 1.4.1 (#39045)
feat(anthropic,fireworks,openai): support langsmith gateway through env var (#38742)
fix(openai): correct gpt-5.3-chat-latest p…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-openai%3D%3D1.4.1
Ollama v0.32.3
Section titled “Ollama v0.32.3”What’s Changed
Section titled “What’s Changed”- Fixed model downloads that stall before sending data.
- Improved integrations: restored Claude Code Channels, fixed Anthropic thinking streams, and made Hermes Desktop respect …
链接：https://github.com/ollama/ollama/releases/tag/v0.32.3
Goose v1.44.0
Section titled “Goose v1.44.0”✨ Features
Section titled “✨ Features”- Support for latest Gemini models #10630
- Pass working_dir to the Stop hook context [#10296](https://github.com/aaif-goose/goose/pull…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.44.0
OpenAI Codex CLI rust-v0.146.0-alpha.5
Section titled “OpenAI Codex CLI rust-v0.146.0-alpha.5”Release 0.146.0-alpha.5
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.146.0-alpha.5
AI bet goes awry: Oracle fires 21,000 employees
Section titled “AI bet goes awry: Oracle fires 21,000 employees”Article URL: https://www.msn.com/en-us/money/economy/ai-bet-goes-awry-oracle-fires-21-000-employees/ar-AA28vWuD Comments URL: https://news.ycombinator.com/item?id=49025754 Points: 76
Comments: 10
Section titled “Comments: 10”Show HN: Claude-thermos keeps your Claude session warm for you
Section titled “Show HN: Claude-thermos keeps your Claude session warm for you”Article URL: https://github.com/izeigerman/claude-thermos Comments URL: https://news.ycombinator.com/item?id=49024882 Points: 66
Comments: 52
Section titled “Comments: 52”The arguments against open source AI are bad
Section titled “The arguments against open source AI are bad”Article URL: https://tombedor.dev/arguments-against-open-source-ai-are-very-bad/ Comments URL: https://news.ycombinator.com/item?id=49024643 Points: 202
Comments: 141
Section titled “Comments: 141”Show HN: OneCLI – OSS credential gateway that keeps secrets out of AI agents
Section titled “Show HN: OneCLI – OSS credential gateway that keeps secrets out of AI agents”hey HN, Jonathan and Guy here, creators of OneCLI (https://onecli.sh/). OneCLI is an open source vault for AI Agents.Traditional vaults are used to store your secrets and, on demand, provide them to you all in a secure way, trusting the person to keep them safe. We figured that in the agent’s world,
Startup founders urge U.S. government not to shut off Chinese open weight AI
Section titled “Startup founders urge U.S. government not to shut off Chinese open weight AI”https://littletech.org/https://static.politico.com/4a/bf/9c4021d8404386b0a311dcccf0…
Comments URL: https://news.ycombinator.com/item?id=49023016 Points: 713
Comments: 642
Section titled “Comments: 642”Show HN: Palmier Pro – Open-source macOS video editor built for AI
Section titled “Show HN: Palmier Pro – Open-source macOS video editor built for AI”Hi HN, we are Marcos and Harrison, cofounders of Palmier (https://palmier.io). We are building Palmier Pro, an open source macOS video editor, with built-in AI generation and a local MCP server that connects to your agent. Here are a few demos:- Making some AI transitions: https://www.youtube.com/wa
I think you might be fooling yourself with AI
Section titled “I think you might be fooling yourself with AI”Article URL: https://louwrentius.com/i-think-you-might-be-fooling-yourself-with-ai.html Comments URL: https://news.ycombinator.com/item?id=49021843 Points: 80
Comments: 155
Section titled “Comments: 155”DARPA, U.S. Air Force fly AI-controlled F-16
Section titled “DARPA, U.S. Air Force fly AI-controlled F-16”Article URL: https://www.darpa.mil/news/2026/darpa-us-air-force-fly-ai-controlled-f-16 Comments URL: https://news.ycombinator.com/item?id=49021597 Points: 176
Comments: 198
Section titled “Comments: 198”2026-07-25
Section titled “2026-07-25”生成时间：2026/7/25 09:26:57（UTC: 2026-07-25T01:26:57.624Z）
AREX: Towards a Recursively Self-Improving Agent for Deep Research
Section titled “AREX: Towards a Recursively Self-Improving Agent for Deep Research”👍 118 · arXiv
Deep research requires agents to find answers that jointly satisfy multiple constraints. Discovering such answers is costly, whereas verifying a candidate can often be decomposed into tractable constr…
SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD
Section titled “SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD”👍 56 · arXiv
Full-parameter post-training of trillion-parameter-scale MoE models introduces substantial system-level challenges for large-scale distributed training, including severe memory pressure, non-overlappe…
ReferTrack: Referring Then Tracking for Embodied Visual Tracking
Section titled “ReferTrack: Referring Then Tracking for Embodied Visual Tracking”👍 44 · arXiv
Embodied visual tracking (EVT) requires a mobile agent to continuously follow a specific target described in natural language using only onboard vision. While recent vision-language-action (VLA) polic…
Visual Contrastive Self-Distillation
Section titled “Visual Contrastive Self-Distillation”👍 41 · arXiv
On-policy self-distillation (OPSD) is promising as it removes the external teacher required by on-policy distillation (OPD), yet it still needs asymmetric information between teacher and student to en…
K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs
Section titled “K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs”👍 40 · arXiv
Large language models are increasingly used in K-12 education, but existing benchmarks mainly test exam question answering rather than understanding how curriculum knowledge is structured and visually…
LangChain langchain-anthropic==1.5.2
Section titled “LangChain langchain-anthropic==1.5.2”Changes since langchain-anthropic==1.5.1
release(anthropic): 1.5.2 (#39056) feat(anthropic): add Claude Opus 5 support (#39054)…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-anthropic%3D%3D1.5.2
CrewAI 1.15.6
Section titled “CrewAI 1.15.6”What’s Changed
Section titled “What’s Changed”Bug Fixes
Section titled “Bug Fixes”- Fix detection of Anthropic preview tool-use blocks.
- Preserve strict tool schema property names.
- Dispatch execution_end hook on failed crew and flow executions. -…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.6
Goose v1.44.0
Section titled “Goose v1.44.0”Security fixes
Section titled “Security fixes”- Arbitrary command execution in goose CLI via
goose review(via git core.fsmonitor config) - [GHSA-r5pp-p5r8-466r](https://github.com/aaif-goose/goose/security/advisories/GHSA-…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.44.0
OpenAI Codex CLI rust-v0.146.0-alpha.9
Section titled “OpenAI Codex CLI rust-v0.146.0-alpha.9”Release 0.146.0-alpha.9
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.146.0-alpha.9
Claude Opus 5
Section titled “Claude Opus 5”https://www.anthropic.com/claude-opus-5-system-card
Comments URL: https://news.ycombinator.com/item?id=49038433 Points: 1279
Comments: 698
Section titled “Comments: 698”Claude Opus 5
Section titled “Claude Opus 5”Article URL: https://www.anthropic.com/news/claude-opus-5 Comments URL: https://news.ycombinator.com/item?id=49038393 Points: 73
Comments: 3
Section titled “Comments: 3”Be skeptical of OpenAI’s rogue hacker agent story
Section titled “Be skeptical of OpenAI’s rogue hacker agent story”Article URL: https://www.theguardian.com/technology/2026/jul/24/openai-rogue-hacker Comments URL: https://news.ycombinator.com/item?id=49038060 Points: 416
Comments: 226
Section titled “Comments: 226”Open Weights and American AI Leadership [pdf]
Section titled “Open Weights and American AI Leadership [pdf]”Article URL: https://images.nvidia.com/pdf/Open-Weights-and-American-AI-Leadership.pdf Comments URL: https://news.ycombinator.com/item?id=49035751 Points: 111
Comments: 2
Section titled “Comments: 2”Oracle fires 21,000 employees to fund AI spending
Section titled “Oracle fires 21,000 employees to fund AI spending”Article URL: https://www.jpost.com/business-and-innovation/tech-and-start-ups/article-903442 Comments URL: https://news.ycombinator.com/item?id=49035314 Points: 73
Comments: 13
Section titled “Comments: 13”I Tried Building a Real App with AI. It Took a Year
Section titled “I Tried Building a Real App with AI. It Took a Year”Article URL: https://www.alexhyett.com/videos/tried-building-app-with-ai-it-took-a-year/ Comments URL: https://news.ycombinator.com/item?id=49034342 Points: 87
Comments: 77
Section titled “Comments: 77”Hetzner is working on LLM Inference
Section titled “Hetzner is working on LLM Inference”Article URL: https://sliplane.io/blog/hetzner-inference Comments URL: https://news.ycombinator.com/item?id=49033087 Points: 144
Comments: 77
Section titled “Comments: 77”Claude Cookbook
Section titled “Claude Cookbook”Article URL: https://platform.claude.com/cookbook/ Comments URL: https://news.ycombinator.com/item?id=49031409 Points: 287