发布日期：2026-08-26
收录条目：20
1. Liquid AI Open-Sources Pipette: A Reproducible Benchmarking Suite That Measures On-Device Models, Quantization, Runtime and Hardware Together
- 来源：MarkTechPost
- 发布时间：2026-08-25 23:48 UTC
- 链接：https://www.marktechpost.com/2026/08/25/liquid-ai-open-sources-pipette-a-reproducible-benchmarking-suite-that-measures-on-device-models-quantization-runtime-and-hardware-together/
摘要：Model cards report quality under server-class, full-precision conditions. Those numbers rarely predict how the same model behaves on a phone. This week, Liquid AI released Pipette. It is an open-source platform for bench
2. Perplexity Ships Portable Computer on NVIDIA DGX Spark: Local Harness, OS-Enforced Sandbox, and Zero Per-Token Cost for Local Steps
- 来源：MarkTechPost
- 发布时间：2026-08-25 19:12 UTC
- 链接：https://www.marktechpost.com/2026/08/25/perplexity-ships-portable-computer-on-nvidia-dgx-spark-local-harness-os-enforced-sandbox-and-zero-per-token-cost-for-local-steps/
摘要：Perplexity releases Portable Computer, packaging local models, harness, sandbox, and connectors into one system running on NVIDIA DGX Spark. The post Perplexity Ships Portable Computer on NVIDIA DGX Spark: Local Harness,
3. Meta AI Introduces MetaRoCE: A Clean-Sheet RDMA Transport Built for AI-Scale Ethernet
- 来源：MarkTechPost
- 发布时间：2026-08-25 17:25 UTC
- 链接：https://www.marktechpost.com/2026/08/25/meta-ai-introduces-metaroce-a-clean-sheet-rdma-transport-built-for-ai-scale-ethernet/
摘要：Training and serving frontier models is now a networking problem as much as a compute problem. Collective operations like all-reduce and all-to-all synchronize thousands of accelerators during training, and the slowest t
4. 5 ways to upgrade your home decor with Google Search
- 来源：Google AI Blog
- 发布时间：2026-08-25 16:00 UTC
- 链接：https://blog.google/products-and-platforms/products/search/home-decor-tips/
摘要：Learn how to use Google Search tools to find home decor inspiration, shop for furniture, and tackle DIY projects.
5. OpenAI says its Jalapeño chip can power faster AI responses than the competition
- 来源：The Verge AI
- 发布时间：2026-08-25 14:00 UTC
- 链接：https://www.theverge.com/ai-artificial-intelligence/984290/openai-jalapeno-ai-chip-benchmarks
摘要：OpenAI says its new AI chip, Jalapeño, completes tasks more efficiently and returns responses faster than other AI systems, according to a blog post published on Tuesday. During a briefing with reporters, OpenAI hardware
6. OpenAI subpoenaed by Alabama AG over Hugging Face hack
- 来源：The Verge AI
- 发布时间：2026-08-25 09:15 UTC
- 链接：https://www.theverge.com/ai-artificial-intelligence/984239/alabama-attorney-general-subpoena-openai-hugging-face-hack
摘要：Alabama's attorney general issued a subpoena to OpenAI on Monday as part of an investigation into how one of its AI agents escaped a supposedly secure testing environment and autonomously hacked another company last mont
7. The full stack behind abundant intelligence
- 来源：OpenAI News
- 发布时间：2026-08-25 07:05 UTC
- 链接：https://openai.com/index/the-full-stack-behind-abundant-intelligence
摘要：OpenAI CFO Sarah Friar explains how advances across chips, compute, models, and products compound to deliver more useful intelligence at greater scale and lower cost.
8. Jalapeño’s first results show industry-leading speed and efficiency in AI inference
- 来源：OpenAI News
- 发布时间：2026-08-25 07:00 UTC
- 链接：https://openai.com/index/jalapeno-first-results
摘要：Jalapeño is a custom inference chip from OpenAI that delivers faster, more power-efficient AI inference, with higher throughput and lower latency for modern models.
9. KVBoost: Chunk-Level Key-Value Cache Reuse with Deviation-Guided Recomputation for Efficient Large Language Model Inference
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21362
摘要：arXiv:2608.21362v1 Announce Type: new Abstract: Transformer-based large language models (LLMs) incur high prefill latency because key-value (KV) tensors must be recomputed for each request. Existing prefix-caching system
10. AIREP: A Protocol for Per-Decision Evidence in AI Runtime Governance
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21363
摘要：arXiv:2608.21363v1 Announce Type: new Abstract: A protocol is presented for recording the governance decisions of automated AI runtimes. When a runtime releases, blocks, defers, redacts, or escalates an individual output
11. Reviewing Model Collapse and Countermeasures
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21366
摘要：arXiv:2608.21366v1 Announce Type: new Abstract: Driven by massive amounts of web-scale data, generative AI (GenAI) has achieved remarkable progress, enabling various applications in diverse sectors. The advances of GenAI
12. AI Learning and Conceptual Transfer in the Game of Hidden Rules
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21372
摘要：arXiv:2608.21372v1 Announce Type: new Abstract: This report summarizes the work conducted on the Game of Hidden Rules (GOHR), focusing on reinforcement learning agents trained to infer hidden rules from trial-and-error f
13. LitReview Arena: Evaluating Literature Review Agents with Battle-Style Peer Review Platform
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21374
摘要：arXiv:2608.21374v1 Announce Type: new Abstract: Literature reviews are essential to scientific progress, but rigorously evaluating automatically generated reviews remains difficult because many aspects of research utilit
14. SchemaRouter: Field-Aware Tool Routing for Efficient Heterogeneous Agentic RAG
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21375
摘要：arXiv:2608.21375v1 Announce Type: new Abstract: Heterogeneous agentic retrieval-augmented generation (RAG) systems increasingly orchestrate external APIs, internal databases, vector stores, and graph stores. Exposing all
15. RIACT: A Responsible AI System for Personalized Study Habit Tracking and Early Burnout Signal Detection in University Students
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21379
摘要：arXiv:2608.21379v1 Announce Type: new Abstract: Student burnout is highly prevalent in higher education, with reported rates ranging from 12% to over 70% and consistently exceeding those of the working population - yet i
16. There Is No Neutral Harness: Modern LLM Leaderboards Are Manufactured by Config-Fragile Items
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21382
摘要：arXiv:2608.21382v1 Announce Type: new Abstract: Multiple-choice benchmarks fix the questions and the correct answers, but not the harness: the order of the options, the wording of the prompt, and whether a language model
17. Spyre-Accelerated Retrieval-Augmented Generation on IBM LinuxONE: A Cloud-Native Architecture for Secure, High-Throughput Enterprise AI Inference
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21393
摘要：arXiv:2608.21393v1 Announce Type: new Abstract: Running large language models inside enterprise environments has always bumped up against a practical wall: the data lives in one place, the AI horsepower sits somewhere el
18. Hate Speech Classification In Roman Urdu: A Comparative Study On Parameter Efficient Fine-Tuning And Prompt Engineering
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21408
摘要：arXiv:2608.21408v1 Announce Type: new Abstract: Due to the widespread accessibility of the internet and social media, toxic and hateful con-tent has grown exponentially, causing significant distress and negative societal
19. The Abstention Protocol: RCA for Clos Fabrics
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21412
摘要：arXiv:2608.21412v1 Announce Type: new Abstract: Root cause analysis (RCA) in large datacenter networks is challenging because telemetry is noisy, partial, and asynchronous. Score-based approaches degrade under these cond
20. Retrieval-grounded robot program generation and simulation-based correction via Model Context Protocol
- 来源：arXiv cs.AI
- 发布时间：2026-08-25 04:00 UTC
- 链接：https://arxiv.org/abs/2608.21417
摘要：arXiv:2608.21417v1 Announce Type: new Abstract: Flexible manufacturing requires industrial robots to be reprogrammed rapidly as product variants change. This paper presents a language-model-based workflow that generates,