AI 速递 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