周报 2026-07-13 ~ 2026-07-19
生成时间：2026/7/19 12:21:59（UTC: 2026-07-19T04:21:59.684Z）
本周自动总结未启用或调用失败，以下为原始内容合并。
2026-07-13
Section titled “2026-07-13”生成时间：2026/7/13 09:28:02（UTC: 2026-07-13T01:28:02.016Z）
Vidu S1: A Real-Time Interactive Video Generation Model
Section titled “Vidu S1: A Real-Time Interactive Video Generation Model”👍 122 · arXiv
We introduce Vidu S1, a real-time interactive video generation model supporting voice control of digital characters. Users can control video generation content at any moment through voice instructions…
Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning
Section titled “Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning”👍 84 · arXiv
Structure-property relationships are foundational to biology, chemistry and materials science, where function, reactivity and physical response emerge from spatial, chemical and periodic organization…
Video-Oasis: Rethinking Evaluation of Video Understanding
Section titled “Video-Oasis: Rethinking Evaluation of Video Understanding”👍 55 · arXiv
The inherent complexity of video understanding makes it difficult to determine whether Video-LLM benchmark performance stems from visual perception, linguistic reasoning, or knowledge priors. While ma…
Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation
Section titled “Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation”👍 51 · arXiv
Mainstream Vision-Language-Action (VLA) models predict actions primarily from the current observation under a Markovian assumption, thus struggling with long-horizon, temporally dependent tasks. Exist…
Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence
Section titled “Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence”👍 49 · arXiv
Despite the recent promise in robot control, video generative models suffer from a domain mismatch due to their primary focus on content creation. For example, their design inherently prioritizes visu…
OpenClaw v2026.7.1-beta.5
Section titled “OpenClaw v2026.7.1-beta.5”2026.7.1
Section titled “2026.7.1”Highlights
Section titled “Highlights”- Conversational onboarding: Crestodian now runs a real agent-loop setup across the CLI, web install, and macOS app, with AI-guided provider setup, model-judged approv…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.7.1-beta.5
vLLM v0.25.0
Section titled “vLLM v0.25.0”vLLM v0.25.0 Release Notes
Section titled “vLLM v0.25.0 Release Notes”Highlights
Section titled “Highlights”This release features 558 commits from 232 contributors (64 new)!
- Model Runner V2 is now the default for all dense models (#44443). Building o…
链接：https://github.com/vllm-project/vllm/releases/tag/v0.25.0
Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k
Section titled “Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k”This started based off of a hunch. We usually use OpenCode, but were ‘forced’ to use Claude Code for a while due to issues with Meridian. In that time, we saw the usage meter rise much, much more quickly than when using OpenCode.This was the initial anecdotal evidence, but we undertook this small st
Migrating a production AI agent to GPT-5.6: 2.2x faster, 27% cheaper
Section titled “Migrating a production AI agent to GPT-5.6: 2.2x faster, 27% cheaper”Article URL: https://ploy.ai/blog/migrating-a-production-ai-agent-to-gpt-5-6 Comments URL: https://news.ycombinator.com/item?id=48882716 Points: 132
Comments: 46
Section titled “Comments: 46”AI boosts research careers but narrow the span of ideas explored: study
Section titled “AI boosts research careers but narrow the span of ideas explored: study”Article URL: https://spectrum.ieee.org/ai-science-research-flattens-discovery Comments URL: https://news.ycombinator.com/item?id=48881043 Points: 137
Comments: 100
Section titled “Comments: 100”Show HN: Mindwalk – Replay coding-agent sessions on a 3D map of your codebase
Section titled “Show HN: Mindwalk – Replay coding-agent sessions on a 3D map of your codebase”Article URL: https://github.com/cosmtrek/mindwalk Comments URL: https://news.ycombinator.com/item?id=48878682 Points: 149
Comments: 62
Section titled “Comments: 62”Mesh LLM: distributed AI computing on iroh
Section titled “Mesh LLM: distributed AI computing on iroh”Article URL: https://www.iroh.computer/blog/mesh-llm Comments URL: https://news.ycombinator.com/item?id=48876505 Points: 334
Comments: 79
Section titled “Comments: 79”Stop Telling Me to Ask an LLM
Section titled “Stop Telling Me to Ask an LLM”Article URL: https://blog.yaelwrites.com/stop-telling-me-to-ask-an-llm/ Comments URL: https://news.ycombinator.com/item?id=48876441 Points: 192
Comments: 113
Section titled “Comments: 113”AI 2040 and the cult of intelligence
Section titled “AI 2040 and the cult of intelligence”Article URL: https://geohot.github.io//blog/jekyll/update/2026/07/11/ai-2040.html Comments URL: https://news.ycombinator.com/item?id=48874200 Points: 220
Comments: 261
Section titled “Comments: 261”Reverse centaurs are the answer to the AI paradox (2025)
Section titled “Reverse centaurs are the answer to the AI paradox (2025)”Article URL: https://pluralistic.net/2025/09/11/vulgar-thatcherism/#there-is-an-alternative Comments URL: https://news.ycombinator.com/item?id=48873855 Points: 107
Comments: 69
Section titled “Comments: 69”2026-07-14
Section titled “2026-07-14”生成时间：2026/7/14 09:17:05（UTC: 2026-07-14T01:17:05.218Z）
Vidu S1: A Real-Time Interactive Video Generation Model
Section titled “Vidu S1: A Real-Time Interactive Video Generation Model”👍 128 · arXiv
We introduce Vidu S1, a real-time interactive video generation model supporting voice control of digital characters. Users can control video generation content at any moment through voice instructions…
Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning
Section titled “Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning”👍 85 · arXiv
Structure-property relationships are foundational to biology, chemistry and materials science, where function, reactivity and physical response emerge from spatial, chemical and periodic organization…
Video-Oasis: Rethinking Evaluation of Video Understanding
Section titled “Video-Oasis: Rethinking Evaluation of Video Understanding”👍 58 · arXiv
The inherent complexity of video understanding makes it difficult to determine whether Video-LLM benchmark performance stems from visual perception, linguistic reasoning, or knowledge priors. While ma…
Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading
Section titled “Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading”👍 48 · arXiv
AI agents have become capable of autonomously completing short, well-specified tasks. However, existing terminal benchmarks largely focus on simple problems that finish within minutes and are evaluate…
Why Can’t I Open My Drawer? Mitigating Object-Driven Shortcuts in Zero-Shot Compositional Action Recognition
Section titled “Why Can’t I Open My Drawer? Mitigating Object-Driven Shortcuts in Zero-Shot Compositional Action Recognition”👍 48 · arXiv
Zero-Shot Compositional Action Recognition (ZS-CAR) requires recognizing novel verb-object combinations composed of previously observed primitives. In this work, we tackle a key failure mode: models p…
OpenClaw v2026.7.1
Section titled “OpenClaw v2026.7.1”2026.7.1
Section titled “2026.7.1”Highlights
Section titled “Highlights”- New models and providers: add Featherless, Claude Sonnet 5 and Mythos 5, Meta Muse Spark 1.1, and ClawRouter; make GPT-5.6 the new-setup default with
/think ultra…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.7.1
Goose v1.42.0
Section titled “Goose v1.42.0”✨ Features
Section titled “✨ Features”- OllamaCloudProvider with dynamic model discovery and context limits #10264
- Per-message usage stats UI (tokens, cost, TTFT, tok/s) […
链接：https://github.com/aaif-goose/goose/releases/tag/v1.42.0
OpenAI Codex CLI rust-v0.144.3
Section titled “OpenAI Codex CLI rust-v0.144.3”Chores
Section titled “Chores”- Published a version-only release with no merged pull request changes since
rust-v0.144.2.
Changelog
Section titled “Changelog”Full Changelog: https://github.com/openai/codex/compare/rust-v0.144.2…rust-v0…
链接：https://github.com/openai/codex/releases/tag/rust-v0.144.3
Uber’s product chief on hotels, robotaxis, and why the company doesn’t want to be “everything for everyone”
Section titled “Uber’s product chief on hotels, robotaxis, and why the company doesn’t want to be “everything for everyone””Uber Chief Product Officer Sachin Kansal walks TechCrunch through the company’s financial-services ambitions, its increasingly complicated relationship with Waymo, its new AV Labs data operation, and how AI is starting to show up in ways riders and drivers will actually notice.
Video-generation startup PixVerse raises $439M, valuation soars past $2B
Section titled “Video-generation startup PixVerse raises $439M, valuation soars past $2B”With the cash, the company aims to expand its world model offering and reach customers across geographies.
Hermes agent maker Nous Research in talks for new funding at $1.5B valuation
Section titled “Hermes agent maker Nous Research in talks for new funding at $1.5B valuation”The company is raising at least $75 million, led by Robot Ventures, with significant participation from USV and other prominent investors.
Satya Nadella has issued a shocking warning to companies using AI
Section titled “Satya Nadella has issued a shocking warning to companies using AI”Of all the debates raging about the potential downsides of AI, there is one worry causing the most hand-wringing among AI enthusiasts in Silicon Valley — that the giant AI labs that sell proprietary models are somehow acting like Trojan horses.
The wildest allegations in Apple’s trade secrets lawsuit against OpenAI
Section titled “The wildest allegations in Apple’s trade secrets lawsuit against OpenAI”Apple’s trade secrets lawsuit against OpenAI contains allegations that range from employees joking about unauthorized access to Apple’s systems to claims that job candidates were asked to bring Apple hardware to interviews. Here are the complaint’s most eye-catching claims.
Sam Altman’s space data center trash talk is what most experts already believe
Section titled “Sam Altman’s space data center trash talk is what most experts already believe”Responding to Musk accusing him of being a scammer, Altman said, “homeboy you’re the one sellling [sic] public market investors on short-term space datacenters.”
Should AI help you get away with killing your spouse?
Section titled “Should AI help you get away with killing your spouse?”What does a world of total user-aligned AI actually look like?
Anthropic starts localizing Claude pricing for India, its biggest market after the US
Section titled “Anthropic starts localizing Claude pricing for India, its biggest market after the US”Claude users in India are starting to see Indian rupee-denominated subscription plans.
2026-07-15
Section titled “2026-07-15”生成时间：2026/7/15 09:14:21（UTC: 2026-07-15T01:14:21.271Z）
Weak-to-Strong Generalization via Direct On-Policy Distillation
Section titled “Weak-to-Strong Generalization via Direct On-Policy Distillation”👍 93 · arXiv
Reinforcement learning with verifiable rewards (RLVR) is a powerful recipe for improving language-model reasoning, but it is expensive to repeat on every new strong model because the target model must…
ABot-N1: Toward a General Visual Language Navigation Foundation Model
Section titled “ABot-N1: Toward a General Visual Language Navigation Foundation Model”👍 81 · arXiv
Visual Language Navigation foundation models aim to unify deep reasoning for grounded spatial decisions with broad versatility for diverse embodied tasks. Current approaches typically achieve this int…
ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memory
Section titled “ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memory”👍 68 · arXiv
Recent VLM and VLA systems have improved robotic perception and action prediction, yet long-horizon embodied agents still require a general runtime layer for reasoning, memory, tool use, verification,…
Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading
Section titled “Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading”👍 61 · arXiv
AI agents have become capable of autonomously completing short, well-specified tasks. However, existing terminal benchmarks largely focus on simple problems that finish within minutes and are evaluate…
Video-Oasis: Rethinking Evaluation of Video Understanding
Section titled “Video-Oasis: Rethinking Evaluation of Video Understanding”👍 61 · arXiv
The inherent complexity of video understanding makes it difficult to determine whether Video-LLM benchmark performance stems from visual perception, linguistic reasoning, or knowledge priors. While ma…
OpenClaw v2026.7.1
Section titled “OpenClaw v2026.7.1”2026.7.1
Section titled “2026.7.1”OpenClaw v2026.7.1 brings major Control UI and onboarding overhauls, major updates to the official iOS, Android, and macOS apps, expanded model and provider support including GPT-5.6 comp…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.7.1
vLLM v0.25.1
Section titled “vLLM v0.25.1”vLLM v0.25.1
Section titled “vLLM v0.25.1”Highlights
Section titled “Highlights”This release features 2 commits from 2 contributors (1 new)!
v0.25.1 is a patch release containing two targeted bug fixes on top of v0.25.0.
Bug Fixes
Section titled “Bug Fixes”- *…
链接：https://github.com/vllm-project/vllm/releases/tag/v0.25.1
Goose v1.43.0
Section titled “Goose v1.43.0”✨ Features
Section titled “✨ Features”- Reconnect desktop ACP sessions after sleep and connection loss #10411
- Per-message usage stats UI (tokens, cost, TTFT, tok/s) [#1021…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.43.0
OpenAI Codex CLI rust-v0.145.0-alpha.12
Section titled “OpenAI Codex CLI rust-v0.145.0-alpha.12”Release 0.145.0-alpha.12
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.145.0-alpha.12
Financing the AI boom: from cash flows to debt [pdf]
Section titled “Financing the AI boom: from cash flows to debt [pdf]”Article URL: https://www.bis.org/publ/bisbull120.pdf Comments URL: https://news.ycombinator.com/item?id=48913443 Points: 77
Comments: 28
Section titled “Comments: 28”Are we offloading too much of our thinking to AI?
Section titled “Are we offloading too much of our thinking to AI?”Article URL: https://www.artfish.ai/p/offloading-thinking-to-ai Comments URL: https://news.ycombinator.com/item?id=48908178 Points: 376
Comments: 376
Section titled “Comments: 376”Proof of care in the age of AI
Section titled “Proof of care in the age of AI”Article URL: https://jacobfilipp.com/care/ Comments URL: https://news.ycombinator.com/item?id=48906125 Points: 169
Comments: 102
Section titled “Comments: 102”Guardian Angels: LLM Personalization for Productivity and Security
Section titled “Guardian Angels: LLM Personalization for Productivity and Security”Article URL: https://gwern.net/guardian-angel Comments URL: https://news.ycombinator.com/item?id=48906041 Points: 55
Comments: 7
Section titled “Comments: 7”Show HN: I RL-trained an agent that trains models with RL (for ~$1.3k)
Section titled “Show HN: I RL-trained an agent that trains models with RL (for ~$1.3k)”Article URL: https://github.com/Danau5tin/ai-trains-ai Comments URL: https://news.ycombinator.com/item?id=48905919 Points: 96
Comments: 42
Section titled “Comments: 42”How to stop Claude from saying load-bearing
Section titled “How to stop Claude from saying load-bearing”Article URL: https://jola.dev/posts/how-to-stop-claude-from-saying-load-bearing Comments URL: https://news.ycombinator.com/item?id=48905248 Points: 429
Comments: 492
Section titled “Comments: 492”Codex starts encrypting sub-agent prompts
Section titled “Codex starts encrypting sub-agent prompts”Article URL: https://github.com/openai/codex/issues/28058 Comments URL: https://news.ycombinator.com/item?id=48905028 Points: 408
Comments: 240
Section titled “Comments: 240”Demis Hassabis has a plan to harness AI safely
Section titled “Demis Hassabis has a plan to harness AI safely”https://xcancel.com/i/article/2076957440109625718https://www.economist.com/business/2026/07/14/demis-hassabis…, https://archive.ph/GOUcN
Comments URL: https://news.ycombinator.com/item?id=48904095 Points: 135
Comments: 180
Section titled “Comments: 180”2026-07-16
Section titled “2026-07-16”生成时间：2026/7/16 09:23:45（UTC: 2026-07-16T01:23:45.085Z）
Weak-to-Strong Generalization via Direct On-Policy Distillation
Section titled “Weak-to-Strong Generalization via Direct On-Policy Distillation”👍 112 · arXiv
Reinforcement learning with verifiable rewards (RLVR) is a powerful recipe for improving language-model reasoning, but it is expensive to repeat on every new strong model because the target model must…
ABot-N1: Toward a General Visual Language Navigation Foundation Model
Section titled “ABot-N1: Toward a General Visual Language Navigation Foundation Model”👍 84 · arXiv
Visual Language Navigation foundation models aim to unify deep reasoning for grounded spatial decisions with broad versatility for diverse embodied tasks. Current approaches typically achieve this int…
ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memory
Section titled “ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memory”👍 72 · arXiv
Recent VLM and VLA systems have improved robotic perception and action prediction, yet long-horizon embodied agents still require a general runtime layer for reasoning, memory, tool use, verification,…
Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading
Section titled “Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading”👍 64 · arXiv
AI agents have become capable of autonomously completing short, well-specified tasks. However, existing terminal benchmarks largely focus on simple problems that finish within minutes and are evaluate…
Video Generation Models are General-Purpose Vision Learners
Section titled “Video Generation Models are General-Purpose Vision Learners”👍 59 · arXiv
Driven by next-token prediction, NLP shifted from task-specific models into powerful generalist foundation models. What, then, is the equivalent catalyst needed to achieve a general-purpose model in c…
OpenClaw v2026.7.2-beta.1
Section titled “OpenClaw v2026.7.2-beta.1”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.1
vLLM v0.25.1
Section titled “vLLM v0.25.1”vLLM v0.25.1
Section titled “vLLM v0.25.1”Highlights
Section titled “Highlights”This release features 2 commits from 2 contributors (1 new)!
v0.25.1 is a patch release containing two targeted bug fixes on top of v0.25.0.
Bug Fixes
Section titled “Bug Fixes”- *…
链接：https://github.com/vllm-project/vllm/releases/tag/v0.25.1
Goose v1.43.0
Section titled “Goose v1.43.0”✨ Features
Section titled “✨ Features”- Reconnect desktop ACP sessions after sleep and connection loss #10411
- Per-message usage stats UI (tokens, cost, TTFT, tok/s) [#1021…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.43.0
OpenAI Codex CLI rust-v0.145.0-alpha.15
Section titled “OpenAI Codex CLI rust-v0.145.0-alpha.15”Release 0.145.0-alpha.15
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.145.0-alpha.15
Microsoft is reportedly training salespeople to talk down OpenAI and Anthropic
Section titled “Microsoft is reportedly training salespeople to talk down OpenAI and Anthropic”Microsoft is looking to sell its in-house AI models as more efficient and cost-effective than its competitors’ models.
Amid hardware legal battle, OpenAI releases a $230 keyboard for Codex
Section titled “Amid hardware legal battle, OpenAI releases a $230 keyboard for Codex”OpenAI, which is in the middle of a legal battle with Apple over hardware trade theft allegations, just released a light-up keyboard designed to be paired with its agentic coding app.
SpaceX falls to $135 IPO price ahead of Starship launch
Section titled “SpaceX falls to $135 IPO price ahead of Starship launch”The stock has steadily fallen from the euphoric post-IPO high, showing that markets may be sobering up to the promises CEO Elon Musk made before and after SpaceX went public.
Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling
Section titled “Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling”It’s the company’s first public proof point after a year and a half spent building AI infrastructure largely out of public view.
Hack suggests AI music generator Suno scraped YouTube for training data
Section titled “Hack suggests AI music generator Suno scraped YouTube for training data”The hacker used an employee’s credentials to access source code, which revealed how Suno scraped decades of audio.
Whatnot acquires Shaped to power real-time live shopping recommendations
Section titled “Whatnot acquires Shaped to power real-time live shopping recommendations”Livestream shopping platform Whatnot has acquired AI startup Shaped, a machine learning company focused on real-time recommendations and search. The deal will bolster Whatnot’s personalization and discovery features as it expands into new product categories.
Microsoft patches record number of security vulnerabilities, citing its use of AI
Section titled “Microsoft patches record number of security vulnerabilities, citing its use of AI”Microsoft’s monthly release of security fixes, dubbed Patch Tuesday, resolved a record 570 security vulnerabilities across the company’s product line, thanks to discoveries with AI.
Apple Intelligence approved for launch in China with Alibaba’s Qwen AI
Section titled “Apple Intelligence approved for launch in China with Alibaba’s Qwen AI”The deal, which was rumored to be in the works last year, marks an important step for Apple’s AI ambitions in a key market.
2026-07-17
Section titled “2026-07-17”生成时间：2026/7/17 09:26:34（UTC: 2026-07-17T01:26:34.573Z）
Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable
Section titled “Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable”👍 165 · arXiv
The capability of a modern AI agent depends not only on its foundation model but also on its harness, which constructs prompts, manages state, invokes tools, and coordinates execution. As models, APIs…
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”👍 109 · 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…
Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models
Section titled “Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models”👍 90 · arXiv
Coding agents must integrate external tool returns into ongoing reasoning - a capability that standard left-to-right pretraining on code exposes only in its forward direction. We observe that the acti…
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”👍 79 · 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 …
Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation
Section titled “Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation”👍 72 · arXiv
Visual generators excel at rendering, but they confidently fabricate what they do not know. User requests are unbounded, evolving, and deeply long-tailed: new characters, trending entities, post-cutof…
OpenClaw v2026.7.2-beta.1
Section titled “OpenClaw v2026.7.2-beta.1”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.1
LangChain langchain==1.3.14
Section titled “LangChain langchain==1.3.14”Changes since langchain==1.3.13
release(langchain): 1.3.14 (#38883)
fix(langchain): only retry retryable exceptions in ToolRetryMiddleware (#38845)
feat(langchain): ToolErrorMiddleware (#38781)…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain%3D%3D1.3.14
Ollama v0.32.1
Section titled “Ollama v0.32.1”What’s Changed
Section titled “What’s Changed”- Improved Gemma 4 tool calling and multi-turn reasoning, including more reliable tool-response continuations
- Fixed a recurrent MLX model cache leak that could increase memory us…
链接：https://github.com/ollama/ollama/releases/tag/v0.32.1
CrewAI 1.15.3
Section titled “CrewAI 1.15.3”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Add organization ID parameter to PlusAPI client
- Add step interception points and rework execution hooks documentation around @on
- Wire execution-boundary intercept…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.3
OpenAI Codex CLI rust-v0.144.5
Section titled “OpenAI Codex CLI rust-v0.144.5”Bug Fixes
Section titled “Bug Fixes”- Improved dangerous-command detection, including more forced
rmforms, and provides clearer rejection reasons when commands are denied. (#33455)
Changelog
Section titled “Changelog”Full Changelog: https:/…
链接：https://github.com/openai/codex/releases/tag/rust-v0.144.5
LM Studio Bionic: the AI agent for open models
Section titled “LM Studio Bionic: the AI agent for open models”Article URL: https://lmstudio.ai/blog/introducing-lm-studio-bionic Comments URL: https://news.ycombinator.com/item?id=48939662 Points: 150
Comments: 62
Section titled “Comments: 62”$100 AI Music Video: Claude Fable 5 vs. GPT-5.6 Sol
Section titled “$100 AI Music Video: Claude Fable 5 vs. GPT-5.6 Sol”Article URL: https://www.tryai.dev/blog/ai-music-video-arena-claude-vs-gpt-5.6 Comments URL: https://news.ycombinator.com/item?id=48939524 Points: 114
Comments: 121
Section titled “Comments: 121”German AI consortium releases Soofi S, an open 30B model that tops benchmarks
Section titled “German AI consortium releases Soofi S, an open 30B model that tops benchmarks”Article URL: https://the-decoder.com/german-ai-consortium-releases-soofi-s-an-open-30b-model-that-tops-benchmarks-in-both-english-and-german/ Comments URL: https://news.ycombinator.com/item?id=48937756 Points: 121
Comments: 27
Section titled “Comments: 27”Detecting LLM-Generated Texts with “Classical” Machine Learning
Section titled “Detecting LLM-Generated Texts with “Classical” Machine Learning”Article URL: https://blog.lyc8503.net/en/post/llm-classifier/ Comments URL: https://news.ycombinator.com/item?id=48936880 Points: 153
Comments: 103
Section titled “Comments: 103”How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM
Section titled “How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM”Article URL: https://www.zhinit.dev/blog/training-a-kick-drum-diffusion-model Comments URL: https://news.ycombinator.com/item?id=48935687 Points: 97
Comments: 54
Section titled “Comments: 54”Generative AI Is an Engineering Disaster
Section titled “Generative AI Is an Engineering Disaster”Article URL: https://www.theatlantic.com/technology/2026/07/generative-ai-engineering-disaster/687901/ Comments URL: https://news.ycombinator.com/item?id=48934046 Points: 100
Comments: 74
Section titled “Comments: 74”The LLM Critics Are Right. I Use LLMs Anyway
Section titled “The LLM Critics Are Right. I Use LLMs Anyway”Article URL: https://www.theocharis.dev/blog/llm-critics-are-right-i-use-llms-anyway/ Comments URL: https://news.ycombinator.com/item?id=48933310 Points: 188
Comments: 197
Section titled “Comments: 197”Stop saying that AI is just a tool and it only matters how it is used
Section titled “Stop saying that AI is just a tool and it only matters how it is used”Article URL: https://www.frank.computer/blog/2025/05/just-a-tool.html Comments URL: https://news.ycombinator.com/item?id=48930363 Points: 103
Comments: 112
Section titled “Comments: 112”2026-07-18
Section titled “2026-07-18”生成时间：2026/7/18 09:18:48（UTC: 2026-07-18T01:18:48.015Z）
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”👍 122 · 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…
VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding
Section titled “VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding”👍 108 · 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 …
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”👍 84 · 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”👍 67 · 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 (…
KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill
Section titled “KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill”👍 52 · arXiv
OpenClaw has emerged as a leading agent framework for complex task automation, yet it faces insufficient cross-platform GUI interaction support and a well-built self-evolution mechanism. These flaws l…
OpenClaw v2026.7.2-beta.2
Section titled “OpenClaw v2026.7.2-beta.2”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.2
LangChain langchain==1.3.14
Section titled “LangChain langchain==1.3.14”Changes since langchain==1.3.13
release(langchain): 1.3.14 (#38883)
fix(langchain): only retry retryable exceptions in ToolRetryMiddleware (#38845)
feat(langchain): ToolErrorMiddleware (#38781)…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain%3D%3D1.3.14
Ollama v0.32.1
Section titled “Ollama v0.32.1”What’s Changed
Section titled “What’s Changed”- Improved Gemma 4 tool calling and multi-turn reasoning, including more reliable tool-response continuations
- Fixed a recurrent MLX model cache leak that could increase memory us…
链接：https://github.com/ollama/ollama/releases/tag/v0.32.1
CrewAI 1.15.4
Section titled “CrewAI 1.15.4”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Promote Skills Repository out of experimental status
Documentation
Section titled “Documentation”- Add Flows in Studio documentation
Contributors
Section titled “Contributors”@jessemiller, @joaomdmoura, @vinibrsl…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.4
OpenAI Codex CLI rust-v0.145.0-alpha.23
Section titled “OpenAI Codex CLI rust-v0.145.0-alpha.23”Release 0.145.0-alpha.23
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.145.0-alpha.23
Kaiser nurses say AI, workplace surveillance are making their jobs, care worse
Section titled “Kaiser nurses say AI, workplace surveillance are making their jobs, care worse”Article URL: https://localnewsmatters.org/2026/07/15/kaiser-nurses-say-ai-workplace-surveillance-are-making-their-jobs-and-patient-care-worse/ Comments URL: https://news.ycombinator.com/item?id=48952880 Points: 247
Comments: 146
Section titled “Comments: 146”Everybody’s Weirded Out by AI–Except the People Who Foist It on Us
Section titled “Everybody’s Weirded Out by AI–Except the People Who Foist It on Us”Article URL: https://newrepublic.com/article/213004/everybody-weirded-ai-except-people-foist-us Comments URL: https://news.ycombinator.com/item?id=48952445 Points: 59
Comments: 60
Section titled “Comments: 60”Ask HN: Did Fable disappear from your Claude usage and requires credits now?
Section titled “Ask HN: Did Fable disappear from your Claude usage and requires credits now?”Looks like it happened. I actually assumed they would just enable it for all today, considering the Sol and Kimmi releases this past week. They doubled down instead.EDIT: Was an outage, they just fixed it (https://status.claude.com/)
Comments URL: https://news.ycombinator.com/item?id=48950477 Point
The state of open source AI
Section titled “The state of open source AI”Article URL: https://stateofopensource.ai/ Comments URL: https://news.ycombinator.com/item?id=48947825 Points: 367
Comments: 266
Section titled “Comments: 266”Claude Code: Anatomy of a Misfeature
Section titled “Claude Code: Anatomy of a Misfeature”Article URL: https://www.olafalders.com/2026/07/17/claude-code-anatomy-of-a-misfeature/ Comments URL: https://news.ycombinator.com/item?id=48947776 Points: 135
Comments: 116
Section titled “Comments: 116”AI Meets Cryptography 2: What AI Found in OpenVM’s ZkVM
Section titled “AI Meets Cryptography 2: What AI Found in OpenVM’s ZkVM”Article URL: https://blog.zksecurity.xyz/posts/openvm-bugs/ Comments URL: https://news.ycombinator.com/item?id=48947714 Points: 82
Comments: 6
Section titled “Comments: 6”VulnHunter: Capital One’s agentic AI code security tool
Section titled “VulnHunter: Capital One’s agentic AI code security tool”Article URL: https://www.capitalone.com/tech/open-source/announcing-vulnhunter/ Comments URL: https://news.ycombinator.com/item?id=48946692 Points: 60
Comments: 29
Section titled “Comments: 29”LM Studio Bionic: the AI agent for open models
Section titled “LM Studio Bionic: the AI agent for open models”Article URL: https://lmstudio.ai/blog/introducing-lm-studio-bionic Comments URL: https://news.ycombinator.com/item?id=48939662 Points: 319