周报 2026-06-15 ~ 2026-06-21
生成时间：2026/6/21 14:05:20（UTC: 2026-06-21T06:05:20.739Z）
本周自动总结未启用或调用失败，以下为原始内容合并。
2026-06-15
Section titled “2026-06-15”生成时间：2026/6/15 10:34:41（UTC: 2026-06-15T02:34:41.644Z）
EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments
Section titled “EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments”👍 129 · arXiv
Large language model (LLM) agents have achieved strong performance on a wide range of benchmarks, yet most evaluations assume static environments. In contrast, real-world deployment is inherently dyna…
MiniMax Sparse Attention
Section titled “MiniMax Sparse Attention”👍 120 · arXiv
Ultra-long-context capability is becoming indispensable for frontier LLMs: agentic workflows, repository-scale code reasoning, and persistent memory all require the model to jointly attend over hundre…
WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces
Section titled “WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces”👍 97 · arXiv
Computer-use agents (CUAs) increasingly operate in runtimes that combine visual desktop control, command-line execution, code editing, browsers, and external tools. Existing benchmarks, however, often…
SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoning
Section titled “SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoning”👍 87 · arXiv
Spatial reasoning, the ability to determine where objects are, how they relate, and how they move in 3D, remains a fundamental challenge for vision-language models (VLMs). Tool-augmented agents attemp…
InterleaveThinker: Reinforcing Agentic Interleaved Generation
Section titled “InterleaveThinker: Reinforcing Agentic Interleaved Generation”👍 77 · arXiv
Recent image generators have demonstrated impressive photorealism and instruction-following capabilities in single-image generation and editing. However, constrained by their architectures, they canno…
OpenClaw v2026.6.8-beta.1
Section titled “OpenClaw v2026.6.8-beta.1”2026.6.8
Section titled “2026.6.8”Highlights
Section titled “Highlights”- Telegram and WhatsApp channel delivery are richer and less brittle: Telegram can send structured rich text with tables, lists, expandable blockquotes, prompt-preserving …
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.8-beta.1
LangChain langchain-openai==1.3.2
Section titled “LangChain langchain-openai==1.3.2”Changes since langchain-openai==1.3.1
release(openai): 1.3.2 (#38130)…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain-openai%3D%3D1.3.2
OpenAI Codex CLI rust-v0.140.0-alpha.19
Section titled “OpenAI Codex CLI rust-v0.140.0-alpha.19”Release 0.140.0-alpha.19
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.140.0-alpha.19
As AI companies race to go public, who else is along for the ride?
Section titled “As AI companies race to go public, who else is along for the ride?”Startups are trying to “ride that SpaceX IPO wave.”
As Anthropic suspends access to new models, India debates its AI future
Section titled “As Anthropic suspends access to new models, India debates its AI future”Tech leaders debate whether the Anthropic episode is a wake-up call for India’s AI ambitions.
Meta reportedly moves to unwind $2B Manus deal after Beijing’s demand
Section titled “Meta reportedly moves to unwind $2B Manus deal after Beijing’s demand”Meta starts dismantling its $2 billion Manus acquisition after Beijing ordered the deal reversed.
KPMG pulls report on AI usage due to apparent hallucinations
Section titled “KPMG pulls report on AI usage due to apparent hallucinations”Once again, AI proves to be an unreliable source of information about AI.
Amazon CEO reportedly raised Anthropic model concerns before government crackdown
Section titled “Amazon CEO reportedly raised Anthropic model concerns before government crackdown”Amazon CEO Andy Jassy may have been the source of security concerns that led Anthropic to cut off worldwide access to two models on Friday.
OpenAI faces investigation from state attorneys general
Section titled “OpenAI faces investigation from state attorneys general”It’s not clear which states are involved, but they’re asking about everything from OpenAI’s ad policies to its handling of health data.
Andrew Yang thinks the next big startup opportunity is lowering the cost of living
Section titled “Andrew Yang thinks the next big startup opportunity is lowering the cost of living”Andrew Yang made a list of everything Americans overpay for — housing, food, wireless — and thinks the next startup gold rush is giving that money back.
Anthropic’s safety warnings may have just backfired — the government has pulled the plug on its most powerful AI
Section titled “Anthropic’s safety warnings may have just backfired — the government has pulled the plug on its most powerful AI”Anthropic isn’t hiding its frustration. “We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” the company wrote in a blog post.
2026-06-16
Section titled “2026-06-16”生成时间：2026/6/16 10:37:18（UTC: 2026-06-16T02:37:18.884Z）
OmniDirector: General Multi-Shot Camera Cloning without Cross-Paired Data
Section titled “OmniDirector: General Multi-Shot Camera Cloning without Cross-Paired Data”👍 95 · arXiv
Cloning camera motion from reference videos is an important task in video generation, as videos provide intuitive and precise control. Existing methods either directly use parametric representations t…
APPO: Agentic Procedural Policy Optimization
Section titled “APPO: Agentic Procedural Policy Optimization”👍 66 · arXiv
Recent advances in agentic Reinforcement Learning (RL) have substantially improved the multi-turn tool-use capabilities of large language model agents. However, most existing methods assign credit ove…
Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents
Section titled “Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents”👍 60 · arXiv
Despite recent progress, LLM agents still struggle with reasoning over long interaction histories. While current memory-augmented agents rely on a static retrieve-then-reason paradigm, this rigid pipe…
From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI
Section titled “From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI”👍 42 · arXiv
Large Language Models (LLMs) are undergoing a fundamental transformation from conversational generators into integrated AI systems capable of reasoning, action, memory, and self-improvement. We concep…
Measuring Epistemic Resilience of LLMs Under Misleading Medical Context
Section titled “Measuring Epistemic Resilience of LLMs Under Misleading Medical Context”👍 40 · arXiv
Large language models (LLMs) now reach expert-level scores on medical licensing exams, encouraging the assumption that high scores imply safe medical judgment while patients increasingly use them for …
OpenClaw v2026.6.8-beta.2
Section titled “OpenClaw v2026.6.8-beta.2”Highlights
Section titled “Highlights”- Telegram and WhatsApp channel delivery are richer and less brittle: Telegram can send structured rich text with tables, lists, expandable blockquotes, preserved intentional line bre…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.8-beta.2
vLLM v0.23.0
Section titled “vLLM v0.23.0”vLLM v0.23.0 Release Notes
Section titled “vLLM v0.23.0 Release Notes”Please note that Minimax M3 is not yet supported in this version. Please follow vLLM recipe for usage guides for M3.
…
链接：https://github.com/vllm-project/vllm/releases/tag/v0.23.0
Ollama v0.30.9-rc1
Section titled “Ollama v0.30.9-rc1”What’s Changed
Section titled “What’s Changed”- llama: update llama.cpp to b9637 by @jmorganca in https://github.com/ollama/ollama/pull/16609
Full Changelog: https://github.com/ollama/ollama/compare/v0.30.8…v0.30.9-r…
链接：https://github.com/ollama/ollama/releases/tag/v0.30.9-rc1
Continue v1.3.40-vscode
Section titled “Continue v1.3.40-vscode”Pre-release of the Continue VS Code extension (final-release cleanup)…
链接：https://github.com/continuedev/continue/releases/tag/v1.3.40-vscode
OpenAI Codex CLI rust-v0.141.0-alpha.2
Section titled “OpenAI Codex CLI rust-v0.141.0-alpha.2”Release 0.141.0-alpha.2
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.141.0-alpha.2
Sundar Pichai faces boos, walkout at Stanford graduation ceremony over Google’s Israel, ICE ties
Section titled “Sundar Pichai faces boos, walkout at Stanford graduation ceremony over Google’s Israel, ICE ties”AI is once again at the heart of a college graduation protest — this time for the technology’s use in Google’s defense contracts.
The US government’s Anthropic models ban was never about an AI jailbreak
Section titled “The US government’s Anthropic models ban was never about an AI jailbreak”The Trump administration’s decision that forced Anthropic to pull its latest cybersecurity models could be reactionary, retaliatory, or both, but the message is clear: The AI industry isn’t immune from U.S. government interference.
Meta’s new ‘AI Mode’ on Facebook pulls from public info across its platforms
Section titled “Meta’s new ‘AI Mode’ on Facebook pulls from public info across its platforms”Meta announced Monday that it’s rolling out a wave of new AI features on Facebook, the latest sign of the company’s effort to catch up in the AI race and keep users more engaged on the platform.
SpaceX is public: Everything you need to know post-IPO
Section titled “SpaceX is public: Everything you need to know post-IPO”TechCrunch has followed SpaceX’s start, struggles, and successes from the early days. And we’re here for what happens next too. This package of SpaceX IPO coverage includes who stands to win (and maybe some who won’t), pre-IPO deals, and what’s tucked inside its S-1 registration document.
Cybersecurity vets protest ‘dangerous’ US government ban on Anthropic’s most powerful models
Section titled “Cybersecurity vets protest ‘dangerous’ US government ban on Anthropic’s most powerful models”A group made up of dozens of cybersecurity experts urged the White House to remove export-control restrictions on Anthropic’s Fable and Mythos models, arguing that the order is going to limit the ability of cybersecurity defenders to secure their software and products.
Salesforce acquires AI customer service platform Fin for $3.6B
Section titled “Salesforce acquires AI customer service platform Fin for $3.6B”Salesforce says it wants to use Fin’s team and technology to improve Agentforce, its existing enterprise platform that businesses can use to build custom AI agents that automate tasks.
Sarvam becomes India’s newest AI unicorn with $234 million funding round led by HCLTech
Section titled “Sarvam becomes India’s newest AI unicorn with $234 million funding round led by HCLTech”Indian IT services company HCLTech is investing $150 million in the Bengaluru startup.
As AI agents become employees, NewCore emerges with $66M to give them identities
Section titled “As AI agents become employees, NewCore emerges with $66M to give them identities”NewCore argues the next challenge in enterprise security will be managing AI agents, not people.
2026-06-17
Section titled “2026-06-17”生成时间：2026/6/17 10:34:35（UTC: 2026-06-17T02:34:35.608Z）
JoyAI-VL-Interaction: Real-Time Vision-Language Interaction Intelligence
Section titled “JoyAI-VL-Interaction: Real-Time Vision-Language Interaction Intelligence”👍 159 · arXiv
Many moments in the real world do not wait for a user to ask. A fire starts on a security monitor, an expression flickers across a video call, or a product a viewer wants flashes by in a livestream. Y…
Data Journalist Agent: Transforming Data into Verifiable Multimodal Stories
Section titled “Data Journalist Agent: Transforming Data into Verifiable Multimodal Stories”👍 101 · arXiv
Data tells stories that shape society; the data journalist’s job is to turn raw information into stories non-experts can trust. A high-quality news feature takes a newsroom team weeks: hunting for con…
Geometric Action Model for Robot Policy Learning
Section titled “Geometric Action Model for Robot Policy Learning”👍 90 · arXiv
Generalist robot policies must follow user instructions while reasoning about how objects, cameras, and robot actions interact in the 3D physical world. Recent vision-language-action models (VLAs) and…
DreamX-World 1.0: A General-Purpose Interactive World Model
Section titled “DreamX-World 1.0: A General-Purpose Interactive World Model”👍 86 · arXiv
DreamX-World 1.0 is a general-purpose interactive text/image-to-video world model for controllable long-horizon generation. It supports camera navigation, revisits to previously observed regions, and …
VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models
Section titled “VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models”👍 65 · arXiv
This technical report introduces VibeThinker-3B, a compact dense model with 3B parameters developed to investigate how far verifiable reasoning can be pushed within a strictly small-model regime. Buil…
OpenClaw v2026.6.8
Section titled “OpenClaw v2026.6.8”2026.6.8
Section titled “2026.6.8”Highlights
Section titled “Highlights”- Richer channel delivery: Telegram and WhatsApp are less brittle: Telegram renders structured text with tables, lists, expandable blockquotes, preserved intentional l…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.8
vLLM v0.23.0
Section titled “vLLM v0.23.0”vLLM v0.23.0 Release Notes
Section titled “vLLM v0.23.0 Release Notes”Please note that Minimax M3 is not yet supported in this version. Please follow vLLM recipe for usage guides for M3.
…
链接：https://github.com/vllm-project/vllm/releases/tag/v0.23.0
Ollama v0.30.9
Section titled “Ollama v0.30.9”What’s Changed
Section titled “What’s Changed”- Support for Cohere2Moe architecture
- Fixed LFM2 parser/render for cases where thinking was not emitted
- Fixed issue where
ollama launch claudeand other coding agent or assi…
链接：https://github.com/ollama/ollama/releases/tag/v0.30.9
Continue v1.3.40-vscode
Section titled “Continue v1.3.40-vscode”Pre-release of the Continue VS Code extension (final-release cleanup)…
链接：https://github.com/continuedev/continue/releases/tag/v1.3.40-vscode
OpenAI Codex CLI rust-v0.141.0-alpha.4
Section titled “OpenAI Codex CLI rust-v0.141.0-alpha.4”Release 0.141.0-alpha.4
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.141.0-alpha.4
Anthropic’s latest feud with the Trump admin may actually help it, sales data suggests
Section titled “Anthropic’s latest feud with the Trump admin may actually help it, sales data suggests”Anthropic’s popularity with business users is growing so well that the latest beef with the government might actually boost it, data from Ramp suggests.
SpaceX valuation balloons to $2.6T, briefly passes Amazon
Section titled “SpaceX valuation balloons to $2.6T, briefly passes Amazon”SpaceX’s valuation has increased by $1 trillion since its shares started trading on Friday.
Android 17 launches with new multitasking tools as Google expands Gemini features
Section titled “Android 17 launches with new multitasking tools as Google expands Gemini features”Google has released Android 17 and Wear OS 7, introducing new multitasking features, parental controls, security tools, and smartwatch upgrades. The launch is also accompanied by a Pixel Drop that brings Google’s latest AI models to its devices.
Sixty percent of US consumers say ‘AI’ in brand messaging is a turnoff, survey finds
Section titled “Sixty percent of US consumers say ‘AI’ in brand messaging is a turnoff, survey finds”WordPress VIP’s latest survey suggests consumers are wary of AI-generated answers even as companies increasingly view AI search as an important referral channel.
SpaceX is public: Everything you need to know post-IPO
Section titled “SpaceX is public: Everything you need to know post-IPO”TechCrunch has followed SpaceX’s start, struggles, and successes from the early days. And we’re here for what happens next too. This package of SpaceX IPO coverage includes who stands to win (and maybe some who won’t), pre-IPO deals, and what’s tucked inside its S-1 registration document.
DOJ claims xAI’s unpermitted gas turbines are a matter of ‘national, economic, and energy security’
Section titled “DOJ claims xAI’s unpermitted gas turbines are a matter of ‘national, economic, and energy security’”The Justice department says the Pentagon needs xAI to keep using its unpermitted gas turbines.
Plaud says its software business topped $100M in ARR after shipping over 2M AI notetakers
Section titled “Plaud says its software business topped $100M in ARR after shipping over 2M AI notetakers”Plaud is trying to make a mark in a crowded market full of AI-powered meeting notetakers.
Robinhood’s note on 10% layoffs shows blaming AI isn’t cutting it
Section titled “Robinhood’s note on 10% layoffs shows blaming AI isn’t cutting it”Unlike many of his tech industry peers who have cut thousands of jobs citing the need to restructure to make the most of AI, Robinhood’s CEO Vlad Tenev conspicuously made no mention of AI in his note about layoffs.
2026-06-18
Section titled “2026-06-18”生成时间：2026/6/18 10:32:21（UTC: 2026-06-18T02:32:21.702Z）
LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling
Section titled “LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling”👍 117 · arXiv
Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count. Parallel loop Transformers (PLT) al…
Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale
Section titled “Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale”👍 70 · arXiv
Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this…
Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients
Section titled “Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients”👍 42 · arXiv
Knowledge distillation transfers a teacher’s competence to a small student but is brittle in the small-student regime: forcing the student to imitate logits from a much larger teacher concentrates it …
ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining
Section titled “ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining”👍 40 · arXiv
Vision-Language-Action (VLA) models benefit from large-scale and diverse embodied data, yet scaling robot trajectory collection is costly and labor-intensive. Recent advances show that large-scale ego…
GameCraft-Bench: Can Agents Build Playable Games End-to-End in a Real Game Engine?
Section titled “GameCraft-Bench: Can Agents Build Playable Games End-to-End in a Real Game Engine?”👍 37 · arXiv
Game generation is an emerging application of coding agents, requiring models to transform natural-language specifications into playable interactive systems. Unlike traditional coding tasks, game gene…
OpenClaw v2026.6.8
Section titled “OpenClaw v2026.6.8”2026.6.8
Section titled “2026.6.8”Highlights
Section titled “Highlights”- Richer channel delivery: Telegram and WhatsApp are less brittle: Telegram renders structured text with tables, lists, expandable blockquotes, preserved intentional l…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.8
Ollama v0.30.10
Section titled “Ollama v0.30.10”What’s Changed
Section titled “What’s Changed”- models: add Cohere2MoE model by @jmorganca in https://github.com/ollama/ollama/pull/16670
- llama: update llama.cpp to b9672 by @pdevine in https://github.com/ollama/ollama/pull…
链接：https://github.com/ollama/ollama/releases/tag/v0.30.10
Goose v1.38.0
Section titled “Goose v1.38.0”✨ Features
Section titled “✨ Features”- Opt-in ACP last message snippets for sessions #9798
- Unified OTLP logging schema for cross-tool detection [#9713](https://github.com/…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.38.0
OpenAI Codex CLI rust-v0.141.0-alpha.7
Section titled “OpenAI Codex CLI rust-v0.141.0-alpha.7”Release 0.141.0-alpha.7
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.141.0-alpha.7
A robot is sprinting towards you. Do you want it running on Claude or Grok?
Section titled “A robot is sprinting towards you. Do you want it running on Claude or Grok?”Article URL: https://openrouter.ai/blog/insights/royale-last-agent-standing/ Comments URL: https://news.ycombinator.com/item?id=48576824 Points: 177
Comments: 145
Section titled “Comments: 145”The hacker sent by Anthropic to calm the government’s nerves about AI safety
Section titled “The hacker sent by Anthropic to calm the government’s nerves about AI safety”Readable: https://www.wsj.com/tech/ai/anthropic-mythos-safety-nicholas…
Comments URL: https://news.ycombinator.com/item?id=48575451 Points: 72
Comments: 77
Section titled “Comments: 77”Using AI to improve a challenging reaction in medicinal chemistry
Section titled “Using AI to improve a challenging reaction in medicinal chemistry”Article URL: https://openai.com/index/ai-chemist-improves-reaction/ Comments URL: https://news.ycombinator.com/item?id=48573757 Points: 50
Comments: 18
Section titled “Comments: 18”The Competitive Moat That AI Can’t Replicate
Section titled “The Competitive Moat That AI Can’t Replicate”Article URL: https://ghostinthedata.info/posts/2026/2026-06-13-human-connection-moat/ Comments URL: https://news.ycombinator.com/item?id=48573435 Points: 130
Comments: 107
Section titled “Comments: 107”Only 16 Percent of Americans Think AI Will Have a Positive Impact on Society
Section titled “Only 16 Percent of Americans Think AI Will Have a Positive Impact on Society”Article URL: https://techcrunch.com/2026/06/17/only-16-percent-of-americans-think-ai-will-have-a-positive-impact-on-society-a-new-study-shows/ Comments URL: https://news.ycombinator.com/item?id=48573332 Points: 370
Comments: 451
Section titled “Comments: 451”Launch HN: Adam (YC W25) – Open-Source AI CAD
Section titled “Launch HN: Adam (YC W25) – Open-Source AI CAD”Hey HN! I’m Zach from Adam (https://adam.new/). We’re building AI agents for mechanical CAD software. We’ve built the company on two fundamental beliefs:- AI will be the primary medium for creating mechanical designs just like it is in software today.- The best paradigm for CAD generation is to gene
TREX: An AI code reviewer that runs your code
Section titled “TREX: An AI code reviewer that runs your code”Article URL: https://www.greptile.com/blog/trex-code-execution Comments URL: https://news.ycombinator.com/item?id=48571851 Points: 51
Comments: 10
Section titled “Comments: 10”Pentagon boasts of using AI to write reports mandated by Congress (1.5mil users)
Section titled “Pentagon boasts of using AI to write reports mandated by Congress (1.5mil users)”Article URL: https://arstechnica.com/ai/2026/06/pentagon-boasts-of-using-ai-to-write-reports-mandated-by-congress/ Comments URL: https://news.ycombinator.com/item?id=48571016 Points: 69
Comments: 53
Section titled “Comments: 53”2026-06-19
Section titled “2026-06-19”生成时间：2026/6/19 10:48:05（UTC: 2026-06-19T02:48:05.269Z）
Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games
Section titled “Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games”👍 38 · arXiv
Deploying multimodal foundation models as closed-loop policies increasingly requires conditioning actions on observations that are no longer visible. However, existing benchmarks either expose the ful…
MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction
Section titled “MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction”👍 35 · 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…
Kairos: A Native World Model Stack for Physical AI
Section titled “Kairos: A Native World Model Stack for Physical AI”👍 29 · arXiv
World models are transitioning from passive visual generators to foundational, operational infrastructure for Physical AI: they must natively acquire world knowledge from heterogeneous experience, mai…
Guava: An Effective and Universal Harness for Embodied Manipulation
Section titled “Guava: An Effective and Universal Harness for Embodied Manipulation”👍 23 · arXiv
Language models trained on large-scale vision-language data have demonstrated strong potential for embodied agents. Harnessing models through embodied tools use offers a promising alternative to end-t…
EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts
Section titled “EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts”👍 18 · arXiv
Reinforcement learning (RL) has become a representative post-training paradigm for LLMs, enabling strong reasoning and agentic capabilities. However, rollout generation remains a dominant latency bott…
LangChain langchain==1.3.10
Section titled “LangChain langchain==1.3.10”Changes since langchain==1.3.9
release(langchain): 1.3.10 (#38255) chore: bump cryptography from 46.0.7 to 48.0.1 in /libs/langchain_v1 (#38176) chore: bump aiohttp from 3.14.0 to 3.14.1 in /libs/lan…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain%3D%3D1.3.10
Ollama v0.30.10
Section titled “Ollama v0.30.10”What’s Changed
Section titled “What’s Changed”- Command A and North family models now run on Apple Silicon with the MLX engine
- Updated the underlying llama.cpp engine to build 9672
- Fixed build artifacts for MLX
**Ful…
链接：https://github.com/ollama/ollama/releases/tag/v0.30.10
CrewAI 1.14.8a2
Section titled “CrewAI 1.14.8a2”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Add single agent action to Flow definitions
- Validate flow CEL expressions at definition load time
Documentation
Section titled “Documentation”- Add Datadog integration guide with importable…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.14.8a2
Continue v2.1.0-vscode
Section titled “Continue v2.1.0-vscode”…
链接：https://github.com/continuedev/continue/releases/tag/v2.1.0-vscode
Goose v1.38.0
Section titled “Goose v1.38.0”✨ Features
Section titled “✨ Features”- Opt-in ACP last message snippets for sessions #9798
- Unified OTLP logging schema for cross-tool detection [#9713](https://github.com/…
链接：https://github.com/aaif-goose/goose/releases/tag/v1.38.0
OpenAI Codex CLI rust-v0.141.0
Section titled “OpenAI Codex CLI rust-v0.141.0”New Features
Section titled “New Features”- Remote executors now use authenticated, end-to-end encrypted Noise relay channels. (#26242, #26245)
- Cross-platform remote execution now preserves executor-native working directori…
链接：https://github.com/openai/codex/releases/tag/rust-v0.141.0
Source: Elastic agrees to buy CRV-backed DeductiveAI for up to $85M
Section titled “Source: Elastic agrees to buy CRV-backed DeductiveAI for up to $85M”DeductiveAI, a startup that uses AI to catch and resolve bugs in software, was founded just three years ago.
AI inference startup Baseten reportedly raising $1.5B months after its last mega-round
Section titled “AI inference startup Baseten reportedly raising $1.5B months after its last mega-round”Startup Baseten is reportedly close to finalizing a $1.5 billion round at a $13 billion as the “inference gold rush” marches on.
Snap spins off AI video team into new company, Dotmo, due to costs
Section titled “Snap spins off AI video team into new company, Dotmo, due to costs”The Snapchat maker is spinning off yet another internal unit. Dotmo will be composed of current Snap staff who are leaving the social media company to focus on AI video development.
OpenAI is bringing on some big guns in the lead-up to its IPO
Section titled “OpenAI is bringing on some big guns in the lead-up to its IPO”OpenAI is bulking up before its IPO, landing Transformer co-inventor Noam Shazeer from Google DeepMind and former Trump AI policy official Dean Ball in the same week.
Almost half of US singles feel negatively about AI in dating, Match says
Section titled “Almost half of US singles feel negatively about AI in dating, Match says”About 47% of singles look negatively at the use of AI in dating — but many dating app users are open to AI helping with profile punch-ups and conversation starters.
Amazon hopes to challenge Nvidia more directly by selling its AI chips
Section titled “Amazon hopes to challenge Nvidia more directly by selling its AI chips”AWS is in talks to sell its chips to other data centers. CEO Andy Jassy has said this represents a $50 billion opportunity for the company.
AI data centers just got a government-mandated fast lane to the grid
Section titled “AI data centers just got a government-mandated fast lane to the grid”FERC told grid operators to give data centers a fast lane for interconnections, but it failed to address electricity supply shortages.
The smartphone era created an attention crisis. Slowtech is fixing it
Section titled “The smartphone era created an attention crisis. Slowtech is fixing it”“People just really want to take back control of their time, their lives, their attention… They’re down for whatever helps them do that.”
2026-06-20
Section titled “2026-06-20”生成时间：2026/6/20 10:08:38（UTC: 2026-06-20T02:08:38.398Z）
Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance
Section titled “Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance”👍 98 · 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”👍 61 · 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…
MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction
Section titled “MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction”👍 41 · 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”👍 35 · 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…
S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence
Section titled “S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence”👍 28 · arXiv
Real-world spatial intelligence requires reasoning over a continuous and evolving 3D world, yet existing VLMs and tool-augmented agents largely remain tied to static, stateless inference from isolated…
OpenClaw v2026.6.9-beta.1
Section titled “OpenClaw v2026.6.9-beta.1”2026.6.9
Section titled “2026.6.9”Highlights
Section titled “Highlights”- Richer Telegram delivery: Telegram now sends rich HTML, preserves rich markdown and sticker paths, renders progress drafts and command output more faithfully, and ke…
链接：https://github.com/openclaw/openclaw/releases/tag/v2026.6.9-beta.1
LangChain langchain==1.3.10
Section titled “LangChain langchain==1.3.10”Changes since langchain==1.3.9
release(langchain): 1.3.10 (#38255) chore: bump cryptography from 46.0.7 to 48.0.1 in /libs/langchain_v1 (#38176) chore: bump aiohttp from 3.14.0 to 3.14.1 in /libs/lan…
链接：https://github.com/langchain-ai/langchain/releases/tag/langchain%3D%3D1.3.10
CrewAI 1.14.8a2
Section titled “CrewAI 1.14.8a2”What’s Changed
Section titled “What’s Changed”Features
Section titled “Features”- Add single agent action to Flow definitions
- Validate flow CEL expressions at definition load time
Documentation
Section titled “Documentation”- Add Datadog integration guide with importable…
链接：https://github.com/crewAIInc/crewAI/releases/tag/1.14.8a2
Continue v2.1.0-vscode
Section titled “Continue v2.1.0-vscode”…
链接：https://github.com/continuedev/continue/releases/tag/v2.1.0-vscode
OpenAI Codex CLI rust-v0.142.0-alpha.7
Section titled “OpenAI Codex CLI rust-v0.142.0-alpha.7”Release 0.142.0-alpha.7
…
链接：https://github.com/openai/codex/releases/tag/rust-v0.142.0-alpha.7
Encryption, spyware, and now Mythos: History shows why cyber export control doesn’t work
Section titled “Encryption, spyware, and now Mythos: History shows why cyber export control doesn’t work”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
Billionaire Ambani wants AI in every call, app, and home
Section titled “Billionaire Ambani wants AI in every call, app, and home”Reliance is weaving AI into telecom services used by more than 500 million people.
The CEO of Allbirds’ new AI biz has a plan, but no team
Section titled “The CEO of Allbirds’ new AI biz has a plan, but no team”Call it a startup with a sole founder and a very large seed round, but what’s next is less clear.
The US says ASML’s top chip tool may be in China, but how?
Section titled “The US says ASML’s top chip tool may be in China, but how?”There’s a commercial logic that cuts against the idea that ASML would risk its export license to arm a Chinese customer.
Source: Elastic agrees to buy CRV-backed Deductive AI for up to $85M
Section titled “Source: Elastic agrees to buy CRV-backed Deductive AI for up to $85M”Deductive AI, a startup that uses AI to catch and resolve bugs in software, was founded just three years ago.
AI inference startup Baseten reportedly raising $1.5B months after its last mega-round
Section titled “AI inference startup Baseten reportedly raising $1.5B months after its last mega-round”Startup Baseten is reportedly close to finalizing a $1.5 billion round at a $13 billion as the “inference gold rush” marches on.