周报 2026-06-22 ~ 2026-06-28
生成时间：2026/6/28 13:18:56（UTC: 2026-06-28T05:18:56.686Z）
本周自动总结未启用或调用失败，以下为原始内容合并。
2026-06-22
Section titled “2026-06-22”生成时间：2026/6/22 10:36:27（UTC: 2026-06-22T02:36:27.811Z）
Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance
Section titled “Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance”👍 113 · arXiv
While 10B-level industrial foundation models have pushed the boundaries of image inpainting, their prohibitive computational costs severely hinder practical deployment. Constructing a highly optimized…
DragMesh-2: Physically Plausible Dexterous Hand-Object Interaction with Articulated Objects
Section titled “DragMesh-2: Physically Plausible Dexterous Hand-Object Interaction with Articulated Objects”👍 69 · arXiv
Dexterous interaction with articulated objects is important for household, assistive, and humanoid manipulation, where multi-finger hands can provide compliant contact patterns beyond parallel-jaw gra…
Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages
Section titled “Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages”👍 49 · arXiv
LiveCodeBench (LCB) has recently become a widely adopted benchmark for evaluating large language models (LLMs) on code-generation tasks. By curating competitive programming problems, constantly adding…
MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction
Section titled “MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction”👍 47 · arXiv
Motion forecasting is central to visual intelligence: agents must anticipate how objects will move in order to plan actions, reason about physical interactions, and synthesize realistic futures. We ar…
Playful Agentic Robot Learning
Section titled “Playful Agentic Robot Learning”👍 44 · arXiv
Current agentic robot systems can write executable Code-as-Policy programs, observe feedback, and revise behavior across multiple attempts, but they remain largely task-driven: reusable skills are acq…
OpenClaw v2026.6.10-beta.1
Section titled “OpenClaw v2026.6.10-beta.1”Highlights
Section titled “Highlights”- More reliable agent turns and session state: OpenClaw now preserves pending subagent completion announcements, keeps chat history transcripts non-empty, maintains media index al…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.10-beta.1
OpenAI Codex CLI rust-v0.142.0-alpha.9
Section titled “OpenAI Codex CLI rust-v0.142.0-alpha.9”Release 0.142.0-alpha.9
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.142.0-alpha.9
When the Trump administration cracks down on Anthropic, who benefits?
Section titled “When the Trump administration cracks down on Anthropic, who benefits?”On the new episode of Equity, we discussed what actually prompted the administration’s latest moves against Anthropic, and what this might mean for the AI ecosystem.
Beyond Siri: Here are the practical AI features coming to your iPhone in iOS 27
Section titled “Beyond Siri: Here are the practical AI features coming to your iPhone in iOS 27”Siri’s AI overhaul may have grabbed the headlines at WWDC, but some of Apple’s most useful AI features are arriving elsewhere in iOS 27.
Signal’s Meredith Whittaker wants you to remember that AI chatbots ‘are not your friends’
Section titled “Signal’s Meredith Whittaker wants you to remember that AI chatbots ‘are not your friends’”“These are not your friends. These are not conscious beings. These are not sentient interlocutors.”
In the Weights is your new AI-centric vanity search
Section titled “In the Weights is your new AI-centric vanity search”So … what’s your In the Weights score?
Nobel laureate John Jumper is leaving DeepMind for rival Anthropic
Section titled “Nobel laureate John Jumper is leaving DeepMind for rival Anthropic”Jumper isn’t the only big name leaving Google DeepMind.
From PGP to Mythos: a brief history of export controls that didn’t stop anyone
Section titled “From PGP to Mythos: a brief history of export controls that didn’t stop anyone”For the last 30 years, stopping the flow of cybersecurity-related software has proven to be ineffective. It’s unclear why it would work now with Anthropic’s cybersecurity model Mythos.
Is the US government’s Anthropic ban accidentally helping the brand?
Section titled “Is the US government’s Anthropic ban accidentally helping the brand?”Just as last week was ending, the US government forced Anthropic to pull its two newest models, Fable 5 and Mythos 5, citing national security concerns after Amazon researchers allegedly found a way to bypass Fable 5’s guardrails. Cybersecurity researchers have since
The US banned Anthropic’s Fable 5 release, but the numbers don’t seem to care
Section titled “The US banned Anthropic’s Fable 5 release, but the numbers don’t seem to care”Just as last week was ending, the US government forced Anthropic to pull its two newest models, Fable 5 and Mythos 5, citing national security concerns after Amazon researchers allegedly found a way to bypass Fable 5’s guardrails. Cybersecurity researchers have since
2026-06-23
Section titled “2026-06-23”生成时间：2026/6/23 10:02:38（UTC: 2026-06-23T02:02:38.461Z）
Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance
Section titled “Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance”👍 118 · arXiv
While 10B-level industrial foundation models have pushed the boundaries of image inpainting, their prohibitive computational costs severely hinder practical deployment. Constructing a highly optimized…
DragMesh-2: Physically Plausible Dexterous Hand-Object Interaction with Articulated Objects
Section titled “DragMesh-2: Physically Plausible Dexterous Hand-Object Interaction with Articulated Objects”👍 69 · arXiv
Dexterous interaction with articulated objects is important for household, assistive, and humanoid manipulation, where multi-finger hands can provide compliant contact patterns beyond parallel-jaw gra…
Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages
Section titled “Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages”👍 56 · arXiv
LiveCodeBench (LCB) has recently become a widely adopted benchmark for evaluating large language models (LLMs) on code-generation tasks. By curating competitive programming problems, constantly adding…
PerceptionDLM: Parallel Region Perception with Multimodal Diffusion Language Models
Section titled “PerceptionDLM: Parallel Region Perception with Multimodal Diffusion Language Models”👍 51 · arXiv
Multimodal large language models (MLLMs) have achieved remarkable progress in visual understanding tasks. However, most existing MLLMs rely on autoregressive generation, which limits their efficiency …
Playful Agentic Robot Learning
Section titled “Playful Agentic Robot Learning”👍 45 · arXiv
Current agentic robot systems can write executable Code-as-Policy programs, observe feedback, and revise behavior across multiple attempts, but they remain largely task-driven: reusable skills are acq…
OpenClaw v2026.6.10-beta.2
Section titled “OpenClaw v2026.6.10-beta.2”2026.6.10
Section titled “2026.6.10”Highlights
Section titled “Highlights”- Automatic fast mode for talks: OpenClaw can enable fast mode for short conversational turns, then return to normal mode for longer runs with bounded fallback and de…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.10-beta.2
LangChain langchain==1.3.11
Section titled “LangChain langchain==1.3.11”Changes since langchain==1.3.10
release(langchain): 1.3.11 (#38377)
fix(langchain,openai): only set strict=True on tools for OpenAI-compatible models in ProviderStrategy (#38370)
chore: bump pyda…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain%3D%3D1.3.11
OpenAI Codex CLI rust-v0.142.0
Section titled “OpenAI Codex CLI rust-v0.142.0”New Features
Section titled “New Features”/usagecan now show and redeem earned usage-limit reset credits, with confirmation, retry, and refreshed availability states. (#28154, #28793)/pluginsnow organizes remote pl…
链接：https://github.com/openai/codex/releases/tag/rust-v0.142.0
The running list: major tech layoffs in 2026 where employers cited AI
Section titled “The running list: major tech layoffs in 2026 where employers cited AI”A running look — in reverse chronological order — at the bigger tech companies that have announced significant layoffs this year with AI as a stated factor.
OpenAI launches new initiative to help find and patch open-source bugs
Section titled “OpenAI launches new initiative to help find and patch open-source bugs”OpenAI is attempting to tackle the security issues of the open source software community.
The AI world is getting ‘loopy’
Section titled “The AI world is getting ‘loopy’”The loop takes agentic AI a step further by authorizing a swarm of agents to work continuously in the background, endlessly.
AI chipmaker Groq confirms $650M raise, re-staffs after Nvidia’s $20B not-acqui-hire deal
Section titled “AI chipmaker Groq confirms $650M raise, re-staffs after Nvidia’s $20B not-acqui-hire deal”What does an AI company do after one of those not-acqui-hire deals? Groq raised money, is leaning into its neocloud business, and is hiring new execs.
Nvidia wants to cut data center water use, but that’s not the same as fixing AI’s water problem
Section titled “Nvidia wants to cut data center water use, but that’s not the same as fixing AI’s water problem”Nvidia announced a new cooling system that cuts water use inside the data center. But it does nothing to address AI’s biggest water use — fossil fuel power plants.
Google DeepMind bets $75M on AI’s future in Hollywood with A24 deal
Section titled “Google DeepMind bets $75M on AI’s future in Hollywood with A24 deal”Google DeepMind and A24 are teaming up to build AI filmmaking tools.
Amazon is testing Alexa+ in India with Hindi support
Section titled “Amazon is testing Alexa+ in India with Hindi support”Amazon is planning to increase the footprint of its new conversational AI assistant Alexa+ to India and is inviting users in the country to test out a Hindi-language version.
SpaceX inks compute deal with Reflection AI, an open source AI lab
Section titled “SpaceX inks compute deal with Reflection AI, an open source AI lab”Reflection AI will pay $150 million a month beginning July 1, 2026 through 2029 for immediate access to Nvidia’s latest GB300 AI chips and supporting hardware across SpaceX’s Colossus 2 data center near Memphis, Tennessee.
2026-06-24
Section titled “2026-06-24”生成时间：2026/6/24 10:04:26（UTC: 2026-06-24T02:04:26.201Z）
PlanBench-XL: Evaluating Long-Horizon Planning of LLM Tool-Use Agents in Large-Scale Tool Ecosystems
Section titled “PlanBench-XL: Evaluating Long-Horizon Planning of LLM Tool-Use Agents in Large-Scale Tool Ecosystems”👍 80 · arXiv
LLM agents increasingly operate in large tool ecosystems, where real-world tasks require discovering relevant tools, inferring implicit sub-goals, and adapting to dynamic environments over long horizo…
OpenRath: Session-Centered Runtime State for Agent Systems
Section titled “OpenRath: Session-Centered Runtime State for Agent Systems”👍 68 · arXiv
Modern agent systems often suffer from fragmented runtime state: transcripts, tool effects, memory events, workspace placement, branch provenance, and replay evidence are recorded separately and becom…
DataClaw0: Agentic Tailoring Multimodal Data from Raw Streams
Section titled “DataClaw0: Agentic Tailoring Multimodal Data from Raw Streams”👍 65 · arXiv
Massive unstructured multimodal streams suffer from high “data entropy,” impeding both efficient human knowledge acquisition and high-quality AI post-training. Existing passive annotation paradigms, h…
EnterpriseClawBench: Benchmarking Agents from Real Workplace Sessions
Section titled “EnterpriseClawBench: Benchmarking Agents from Real Workplace Sessions”👍 57 · arXiv
Enterprise agents increasingly operate inside workspaces: they read heterogeneous files, invoke tools, and deliver business artifacts. We introduce EnterpriseClawBench, an enterprise agent benchmark c…
Grouped Query Experts: Mixture-of-Experts on GQA Self-Attention
Section titled “Grouped Query Experts: Mixture-of-Experts on GQA Self-Attention”👍 46 · arXiv
Self-attention is central to Transformer performance and is often the most expensive part of the Transformer at long context lengths because its pairwise token interactions scale quadratically with se…
OpenClaw v2026.6.10-beta.2
Section titled “OpenClaw v2026.6.10-beta.2”2026.6.10
Section titled “2026.6.10”Highlights
Section titled “Highlights”- Automatic fast mode for talks: OpenClaw can enable fast mode for short conversational turns, then return to normal mode for longer runs with bounded fallback and de…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.10-beta.2
LangChain langchain-openrouter==0.2.4
Section titled “LangChain langchain-openrouter==0.2.4”Changes since langchain-openrouter==0.2.3
release(openrouter): 0.2.4 (#38381)
chore(openrouter): bump openrouter floor to 0.9.2, drop file workaround (#38216)
test(openrouter): cover `cache_control…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-openrouter%3D%3D0.2.4
CrewAI 1.14.8a3
Section titled “CrewAI 1.14.8a3”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Add unified declarative flow loading
- Improve crewai run startup UX
- Consolidate
crewai runandcrewai flow kickoff - Keep flow method progress visible for nest…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.14.8a3
OpenAI Codex CLI rust-v0.143.0-alpha.11
Section titled “OpenAI Codex CLI rust-v0.143.0-alpha.11”Release 0.143.0-alpha.11
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.143.0-alpha.11
Claude Tag
Section titled “Claude Tag”Article URL: https://www.anthropic.com/news/introducing-claude-tag Comments URL: https://news.ycombinator.com/item?id=48648039 Points: 233
Comments: 160
Section titled “Comments: 160”AI’s Affordability Crisis
Section titled “AI’s Affordability Crisis”Article URL: https://blog.dshr.org/2026/06/ais-affordability-crisis.html Comments URL: https://news.ycombinator.com/item?id=48646276 Points: 257
Comments: 344
Section titled “Comments: 344”The Low-Tech AI of Elden Ring
Section titled “The Low-Tech AI of Elden Ring”Article URL: https://nega.tv/posts/low-tech-ai-of-elden-ring.html Comments URL: https://news.ycombinator.com/item?id=48643489 Points: 108
Comments: 55
Section titled “Comments: 55”AI Built a Nuke and Still Lost
Section titled “AI Built a Nuke and Still Lost”Article URL: https://www.lwilko.com/blog/i-gave-an-ai-a-civilization Comments URL: https://news.ycombinator.com/item?id=48641927 Points: 86
Comments: 93
Section titled “Comments: 93”Ask HN: Anthropic banned me from using Claude Code and I don’t know what to do
Section titled “Ask HN: Anthropic banned me from using Claude Code and I don’t know what to do”After using Claude Code at work for months I wanted to use it on my own projects too. Most probably because my vpn was on I got banned after 1 hour of usage. I got my 120$ back. 1 month later I signed up with vpn off. But this time probably because I used the same credit cart (and that’s the only ca
OpenAI DayBreak – GPT-5.5-Cyber
Section titled “OpenAI DayBreak – GPT-5.5-Cyber”Article URL: https://openai.com/index/daybreak-securing-the-world/ Comments URL: https://news.ycombinator.com/item?id=48639063 Points: 206
Comments: 166
Section titled “Comments: 166”Meta pauses AI training program tracking employee keystrokes after internal leak
Section titled “Meta pauses AI training program tracking employee keystrokes after internal leak”Article URL: https://www.businessinsider.com/meta-ai-training-data-leak-exposed-employee-activity-across-company-2026-6 Comments URL: https://news.ycombinator.com/item?id=48636632 Points: 113
Comments: 27
Section titled “Comments: 27”AI has already killed academia as we know it?
Section titled “AI has already killed academia as we know it?”Article URL: https://truths-and-loves.ghost.io/ai-has-already-killed-academia-as-we-know-it/ Comments URL: https://news.ycombinator.com/item?id=48634966 Points: 59
Comments: 49
Section titled “Comments: 49”2026-06-25
Section titled “2026-06-25”生成时间：2026/6/25 10:05:03（UTC: 2026-06-25T02:05:03.945Z）
Qwen-AgentWorld: Language World Models for General Agents
Section titled “Qwen-AgentWorld: Language World Models for General Agents”👍 90 · arXiv
A world model predicts environment dynamics based on current observations and actions, serving as a core cognitive mechanism for reasoning and planning. In this work, we investigate how world modeling…
NatureBench: Can Coding Agents Match the Published SOTA of Nature-Family Papers?
Section titled “NatureBench: Can Coding Agents Match the Published SOTA of Nature-Family Papers?”👍 49 · arXiv
We introduce NatureBench, a cross-discipline benchmark of 90 tasks distilled from peer-reviewed Nature-family publications, designed to evaluate whether AI coding agents can move beyond reproduction t…
MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization
Section titled “MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization”👍 35 · arXiv
MLLM-based mobile GUI agents have made substantial progress in UI understanding and action execution, but adapting them to real target apps remains costly because mobile apps are numerous, frequently …
MemGUI-Agent: An End-to-End Long-Horizon Mobile GUI Agent with Proactive Context Management
Section titled “MemGUI-Agent: An End-to-End Long-Horizon Mobile GUI Agent with Proactive Context Management”👍 33 · arXiv
MLLM-based mobile GUI agents have made substantial progress on short-horizon tasks, yet remain unreliable on long-horizon tasks that require retaining intermediate facts across many steps and app tran…
OpenThoughts-Agent: Data Recipes for Agentic Models
Section titled “OpenThoughts-Agent: Data Recipes for Agentic Models”👍 28 · arXiv
Agentic language models dramatically expand the applications of AI yet little is publicly known about how to curate training data for broadly capable agents. Existing open efforts such as SWE-Smith, S…
OpenClaw v2026.6.11-beta.1
Section titled “OpenClaw v2026.6.11-beta.1”2026.6.11
Section titled “2026.6.11”Highlights
Section titled “Highlights”- More capable channel control: Slack relay mode, native Mattermost
/oc_queue, and per-DM model overrides make channel operations easier to automate and tune. (#947…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.11-beta.1
LangChain langchain-openrouter==0.2.4
Section titled “LangChain langchain-openrouter==0.2.4”Changes since langchain-openrouter==0.2.3
release(openrouter): 0.2.4 (#38381)
chore(openrouter): bump openrouter floor to 0.9.2, drop file workaround (#38216)
test(openrouter): cover `cache_control…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-openrouter%3D%3D0.2.4
Ollama v0.30.11-rc0
Section titled “Ollama v0.30.11-rc0”What’s Changed
Section titled “What’s Changed”- launch: add thinking capability detection to opencode by @hoyyeva in https://github.com/ollama/ollama/pull/15434
- launch: auto-install Claude Code by @hoyyeva in https://github…
链接：https://github.com/ollama/ollama/releases/tag/v0.30.11-rc0
LlamaIndex v0.14.23
Section titled “LlamaIndex v0.14.23”Release Notes
Section titled “Release Notes”[2026-06-24]
Section titled “[2026-06-24]”llama-index-callbacks-argilla [0.5.0]
Section titled “llama-index-callbacks-argilla [0.5.0]”- chore(deps): bump the uv group across 32 directories with 3 updates ([#21664](https://github.com/run-llama/llama_index/pu…
链接：https://github.com/run-llama/llama_index/releases/tag/v0.14.23
CrewAI 1.14.8a5
Section titled “CrewAI 1.14.8a5”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Make declarative refs work across flows and crews (#6326)
Bug Fixes
Section titled “Bug Fixes”- Fix JSON schema flow state kickoff inputs (#6325)
Documentation
Section titled “Documentation”- Nest One Card per St…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.14.8a5
OpenAI Codex CLI rust-v0.142.1
Section titled “OpenAI Codex CLI rust-v0.142.1”New Features
Section titled “New Features”- Added opt-in Windows system proxy support for authentication, including PAC, WPAD, static proxies, and bypass rules. (#26708)
Changelog
Section titled “Changelog”Full Changelog: https://github.com/opena…
链接：https://github.com/openai/codex/releases/tag/rust-v0.142.1
Anthropic says Alibaba illicitly extracted Claude AI model capabilities
Section titled “Anthropic says Alibaba illicitly extracted Claude AI model capabilities”Article URL: https://www.reuters.com/world/china/anthropic-says-alibaba-illicitly-extracted-claude-ai-model-capabilities-2026-06-24/ Comments URL: https://news.ycombinator.com/item?id=48664814 Points: 53
Comments: 87
Section titled “Comments: 87”Big AI labs are hiring philosophers
Section titled “Big AI labs are hiring philosophers”Comments URL: https://news.ycombinator.com/item?id=48662452 Points: 127
Comments: 113
Section titled “Comments: 113”For most of the world, open-source AI is the only way forward
Section titled “For most of the world, open-source AI is the only way forward”Article URL: https://techstrong.ai/articles/for-most-of-the-world-open-source-ai-is-the-only-way-forward/ Comments URL: https://news.ycombinator.com/item?id=48660839 Points: 203
Comments: 131
Section titled “Comments: 131”RubyLLM: A Ruby framework for all major AI providers
Section titled “RubyLLM: A Ruby framework for all major AI providers”Article URL: https://rubyllm.com/ Comments URL: https://news.ycombinator.com/item?id=48660711 Points: 347
Comments: 55
Section titled “Comments: 55”OpenAI and Broadcom unveil LLM-optimized inference chip
Section titled “OpenAI and Broadcom unveil LLM-optimized inference chip”Article URL: https://openai.com/index/openai-broadcom-jalapeno-inference-chip/ Comments URL: https://news.ycombinator.com/item?id=48659257 Points: 141
Comments: 1
Section titled “Comments: 1”Reid Hoffman says SpaceX ‘not an AI company’, xAI ‘complete train wreck’
Section titled “Reid Hoffman says SpaceX ‘not an AI company’, xAI ‘complete train wreck’”Article URL: https://fortune.com/2026/06/24/reid-hoffman-spacex-musk-openai-anthropic-gen-z-mistake/ Comments URL: https://news.ycombinator.com/item?id=48658647 Points: 222
Comments: 255
Section titled “Comments: 255”Haystack: Open-Source AI Framework for Production Ready Agents, RAG
Section titled “Haystack: Open-Source AI Framework for Production Ready Agents, RAG”Article URL: https://haystack.deepset.ai/ Comments URL: https://news.ycombinator.com/item?id=48658095 Points: 86
Comments: 21
Section titled “Comments: 21”How to burst the AI bubble: Strike at its roots
Section titled “How to burst the AI bubble: Strike at its roots”Article URL: https://arstechnica.com/gadgets/2026/06/how-to-burst-the-ai-bubble-strike-at-its-roots/ Comments URL: https://news.ycombinator.com/item?id=48657518 Points: 62
Comments: 49
Section titled “Comments: 49”2026-06-26
Section titled “2026-06-26”生成时间：2026/6/26 10:06:43（UTC: 2026-06-26T02:06:43.472Z）
Are We Ready For An Agent-Native Memory System?
Section titled “Are We Ready For An Agent-Native Memory System?”👍 91 · arXiv
Memory for large language model (LLM) agents has rapidly evolved from simple retrieval-augmented mechanisms into a data management system that supports persistent information storage, retrieval, updat…
DomainShuttle: Freeform Open Domain Subject-driven Text-to-video Generation
Section titled “DomainShuttle: Freeform Open Domain Subject-driven Text-to-video Generation”👍 55 · arXiv
Open domain subject-driven text-to-video (S2V) generation has drawn significant interest in academia and industry. Open domain S2V mainly involves two scenarios: in-domain, which requires retaining th…
Wan-Streamer v0.1: End-to-end Real-time Interactive Foundation Models
Section titled “Wan-Streamer v0.1: End-to-end Real-time Interactive Foundation Models”👍 46 · arXiv
We present Wan-Streamer, a native-streaming, end-to-end interactive foundation model designed from the ground up for real-time, low-latency, full-duplex audio-visual interaction. Wan-Streamer seamless…
ShutterMuse: Capture-Time Photography Guidance with MLLMs
Section titled “ShutterMuse: Capture-Time Photography Guidance with MLLMs”👍 38 · arXiv
Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject…
Improved Large Language Diffusion Models
Section titled “Improved Large Language Diffusion Models”👍 32 · arXiv
Modern large language models are predominantly trained with autoregressive factorization and causal attention. We present iLLaDA, an 8B masked diffusion language model trained from scratch with fully …
OpenClaw v2026.6.11-beta.1
Section titled “OpenClaw v2026.6.11-beta.1”2026.6.11
Section titled “2026.6.11”Highlights
Section titled “Highlights”- More capable channel control: Slack relay mode, native Mattermost
/oc_queue, and per-DM model overrides make channel operations easier to automate and tune. (#947…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.11-beta.1
Ollama v0.30.11-rc0
Section titled “Ollama v0.30.11-rc0”What’s Changed
Section titled “What’s Changed”- launch: add thinking capability detection to opencode by @hoyyeva in https://github.com/ollama/ollama/pull/15434
- launch: auto-install Claude Code by @hoyyeva in https://github…
链接：https://github.com/ollama/ollama/releases/tag/v0.30.11-rc0
LlamaIndex v0.14.23
Section titled “LlamaIndex v0.14.23”Release Notes
Section titled “Release Notes”[2026-06-24]
Section titled “[2026-06-24]”llama-index-callbacks-argilla [0.5.0]
Section titled “llama-index-callbacks-argilla [0.5.0]”- chore(deps): bump the uv group across 32 directories with 3 updates ([#21664](https://github.com/run-llama/llama_index/pu…
链接：https://github.com/run-llama/llama_index/releases/tag/v0.14.23
CrewAI 1.15.0
Section titled “CrewAI 1.15.0”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Track conversational flow turn usage in telemetry
- Support conversational flows in the CLI TUI
- Add unified declarative flow loading
- Add declarative Flow CLI supp…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.0
Goose v1.39.0
Section titled “Goose v1.39.0”✨ Features
Section titled “✨ Features”- ACP method for managing recipes and UI changes #9951
- Use ACP to manage global config and session extensions [#9948](https://github.c…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.39.0
OpenAI Codex CLI rust-v0.142.2
Section titled “OpenAI Codex CLI rust-v0.142.2”New Features
Section titled “New Features”- MCP tools now use tool search by default when supported, improving tool discovery while preserving compatibility with older models and providers. (#29486)
- macOS authentication cli…
链接：https://github.com/openai/codex/releases/tag/rust-v0.142.2
Show HN: OpenKnowledge – open source AI-first alternative to Obsidian/Notion
Section titled “Show HN: OpenKnowledge – open source AI-first alternative to Obsidian/Notion”Hi HN, Nick here. We’re launching OpenKnowledge (https://openknowledge.ai/), a “what you see is what you get” markdown editor that has direct integrations with Claude, Codex, and other agents. Available as MacOS app or Web UI+CLI. Fully free/local and OSS.We built this because we wanted a Notion-lik
Ford AI hiccups push carmaker to rehire ‘gray beard’ inspectors
Section titled “Ford AI hiccups push carmaker to rehire ‘gray beard’ inspectors”Comments URL: https://news.ycombinator.com/item?id=48674446 Points: 574
Comments: 304
Section titled “Comments: 304”Political bias in AI: Where the AI models stand
Section titled “Political bias in AI: Where the AI models stand”Article URL: https://trakkr.ai/bias Comments URL: https://news.ycombinator.com/item?id=48672779 Points: 120
Comments: 238
Section titled “Comments: 238”Why Does Everyone Hate AI?
Section titled “Why Does Everyone Hate AI?”Article URL: https://paulkrugman.substack.com/p/why-does-everyone-hate-ai Comments URL: https://news.ycombinator.com/item?id=48672694 Points: 81
Comments: 141
Section titled “Comments: 141”Anthropic says Alibaba illicitly extracted Claude AI model capabilities
Section titled “Anthropic says Alibaba illicitly extracted Claude AI model capabilities”Article URL: https://www.reuters.com/world/china/anthropic-says-alibaba-illicitly-extracted-claude-ai-model-capabilities-2026-06-24/ Comments URL: https://news.ycombinator.com/item?id=48664814 Points: 750
Comments: 1210
Section titled “Comments: 1210”Big AI labs are hiring philosophers
Section titled “Big AI labs are hiring philosophers”Comments URL: https://news.ycombinator.com/item?id=48662452 Points: 147
Comments: 133
Section titled “Comments: 133”For most of the world, open-source AI is the only way forward
Section titled “For most of the world, open-source AI is the only way forward”Article URL: https://techstrong.ai/articles/for-most-of-the-world-open-source-ai-is-the-only-way-forward/ Comments URL: https://news.ycombinator.com/item?id=48660839 Points: 231
Comments: 144
Section titled “Comments: 144”RubyLLM: A Ruby framework for all major AI providers
Section titled “RubyLLM: A Ruby framework for all major AI providers”Article URL: https://rubyllm.com/ Comments URL: https://news.ycombinator.com/item?id=48660711 Points: 441
Comments: 78
Section titled “Comments: 78”2026-06-27
Section titled “2026-06-27”生成时间：2026/6/27 10:01:43（UTC: 2026-06-27T02:01:43.094Z）
Are We Ready For An Agent-Native Memory System?
Section titled “Are We Ready For An Agent-Native Memory System?”👍 102 · arXiv
Memory for large language model (LLM) agents has rapidly evolved from simple retrieval-augmented mechanisms into a data management system that supports persistent information storage, retrieval, updat…
DomainShuttle: Freeform Open Domain Subject-driven Text-to-video Generation
Section titled “DomainShuttle: Freeform Open Domain Subject-driven Text-to-video Generation”👍 61 · arXiv
Open domain subject-driven text-to-video (S2V) generation has drawn significant interest in academia and industry. Open domain S2V mainly involves two scenarios: in-domain, which requires retaining th…
DanceOPD: On-Policy Generative Field Distillation
Section titled “DanceOPD: On-Policy Generative Field Distillation”👍 55 · arXiv
Modern image generation demands a single model that unifies diverse capabilities, including text-to-image (T2I), local editing, and global editing. However, these capabilities are rarely naturally ali…
ShutterMuse: Capture-Time Photography Guidance with MLLMs
Section titled “ShutterMuse: Capture-Time Photography Guidance with MLLMs”👍 42 · arXiv
Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject…
ViQ: Text-Aligned Visual Quantized Representations at Any Resolution
Section titled “ViQ: Text-Aligned Visual Quantized Representations at Any Resolution”👍 36 · arXiv
A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same w…
LangChain langchain-fireworks==1.4.3
Section titled “LangChain langchain-fireworks==1.4.3”Changes since langchain-fireworks==1.4.2
release(fireworks): 1.4.3 chore: bump vcrpy from 8.1.1 to 8.2.1 in /libs/partners/fireworks (#38314) chore: bump langsmith from 0.8.16 to 0.8.18 in /libs/part…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-fireworks%3D%3D1.4.3
CrewAI 1.15.1a1
Section titled “CrewAI 1.15.1a1”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Track TUI button telemetry
- Require explicit CrewAI project definitions
- Open deployment page after CLI deploy
Bug Fixes
Section titled “Bug Fixes”- Fix JSON crew template rendering
- F…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.15.1a1
Goose v1.39.0
Section titled “Goose v1.39.0”✨ Features
Section titled “✨ Features”- ACP method for managing recipes and UI changes #9951
- Use ACP to manage global config and session extensions [#9948](https://github.c…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.39.0
OpenAI Codex CLI rust-v0.142.3
Section titled “OpenAI Codex CLI rust-v0.142.3”Chores
Section titled “Chores”- Maintenance-only patch release with no user-facing changes since 0.142.2.
Changelog
Section titled “Changelog”Full Changelog: https://github.com/openai/codex/compare/rust-v0.142.2…rust-v0.142.3
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.142.3
Trump Admin releases Anthropic Mythos to be used by more than 100 US companies, agencies
Section titled “Trump Admin releases Anthropic Mythos to be used by more than 100 US companies, agencies”Over 100 companies and government agencies are reportedly authorized to use Mythos 5, including their non-American employees.
OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn’t be the norm
Section titled “OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn’t be the norm”“We don’t believe this kind of government access process should become the long-term default,” says OpenAI. “It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.”
OpenAI poaches Uber India chief to lead its biggest market outside the US
Section titled “OpenAI poaches Uber India chief to lead its biggest market outside the US”The hire marks OpenAI’s latest push into India, expanding offices, partnerships and hiring.
Why everyone from OpenAI to SpaceX is building their own chips (and turning up the heat on Nvidia)
Section titled “Why everyone from OpenAI to SpaceX is building their own chips (and turning up the heat on Nvidia)”Nvidia has dominated the AI chip market for years, but the era of total dependence might be ending. OpenAI just shared its plans to spice things up with Jalapeño, its custom inference chip built with Broadcom, joining Google, Apple, and SpaceX in a growing list of companies building
It’s not about Anthropic vs. OpenAI anymore
Section titled “It’s not about Anthropic vs. OpenAI anymore”AI models have progressed to the point where their capabilities have real political consequences. Dealing with those consequences will require collective action.
OpenAI’s Jalapeño chip is Big Tech’s spiciest move away from Nvidia
Section titled “OpenAI’s Jalapeño chip is Big Tech’s spiciest move away from Nvidia”Nvidia has dominated the AI chip market for years, but the era of total dependence might be ending. OpenAI just shared its plans to spice things up with Jalapeño, its custom inference chip built with Broadcom, joining Google, Apple, and SpaceX in a growing list of companies building
Early Bird pricing ends tonight for TechCrunch Founder Summit
Section titled “Early Bird pricing ends tonight for TechCrunch Founder Summit”Save up to $190 on your pass to TechCrunch Founder Summit 2026. Early Bird pricing ends today, at 11:59 p.m. PT, after which rates increase. Register now.
The White House is asking OpenAI to slow roll the release of its new model over safety concerns
Section titled “The White House is asking OpenAI to slow roll the release of its new model over safety concerns”OpenAI reportedly plans to share its newest model, GPT 5.6, with a select group of partners instead of with the broader public. The reason: the Trump administration told it to.