🤖 AI 资讯日报 — 2026-07-08

生成时间: 2026-07-08 08:41 CST | 共 67 条资讯

T1 官方: 33 | T1.5 媒体: 25 | T2 社区: 9

🏛️ T1 官方一手

OpenAI

• OpenAI News
• Stay up to speed on the rapid advancement of AI technology and the benefits it offers to humanity.
• 🔗 openai.com/news/

• ChatGPT — Release Notes
• A changelog of the latest updates and release notes for ChatGPT.
• 🔗 help.openai.com/en/articles/6825453-chatgpt-release-notes

• OpenAI | Research & Deployment
• We believe our research will eventually lead to artificial general intelligence, a system that can solve human-level problems. Building safe and beneficial
• 🔗 openai.com/

• Introducing ChatGPT
• We've trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer
• 🔗 openai.com/index/chatgpt/

Anthropic Research

• Tracing the thoughts of a large language model - Anthropic
• Anthropic's latest interpretability research: a new microscope to understand Claude's internal mechanisms.
• 🔗 www.anthropic.com/research/tracing-thoughts-language-model

• Home \ Anthropic
• Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
• 🔗 www.anthropic.com

• Exploring model welfare - Anthropic
• Announcing a new research ... We report results from our latest test of whether Claude can help Anthropic employees perform sophisticated robotics
• 🔗 www.anthropic.com/research/exploring-model-welfare

• Automated Alignment Researchers: Using large language models to ...
• We report results from our latest test of whether Claude can help Anthropic employees perform sophisticated robotics tasks. We found that
• 🔗 www.anthropic.com/research/automated-alignment-researchers

• Research - Anthropic
• New interpretability research reveals an emergent mental workspace in Claude that holds internal thoughts that don't appear in the model's output. Economic
• 🔗 www.anthropic.com/research

Hugging Face

• LeRobot v0.6.0: Imagine, Evaluate, Improve
• This new release is about closing the robot learning loop: policies that imagine the future before acting, reward models that tell you when
• 🔗 huggingface.co/blog/lerobot-release-v060

• 🤗 Kernels: Major Updates
• We're on a journey to advance and democratize artificial intelligence through open source and open science.
• 🔗 huggingface.co/blog/revamped-kernels

• Hugging Face and Cerebras bring Gemma 4 to real-time voice AI
• We're on a journey to advance and democratize artificial intelligence through open source and open science.
• 🔗 huggingface.co/blog/cerebras-gemma4-voice-ai

• Daily Papers
• Your daily dose of AI research from AK.
• 🔗 huggingface.co/papers

GitHub AI

• Octoverse: A new developer joins GitHub every second as AI leads TypeScript to #1
• In this year's Octoverse, we uncover how AI, agents, and typed languages are driving the biggest shifts in software development in more than
• 🔗 github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/

• GitHub Copilot is moving to usage-based billing
• Starting June 1, your Copilot usage will consume GitHub AI Credits.
• 🔗 github.blog/news-insights/company-news/github-copilot-is-moving-to-usage-based-billing/

• Research: quantifying GitHub Copilot’s impact on developer productivity and happiness
• When the GitHub Copilot Technical Preview launched just over one year ago, we wanted to know one thing: Is this tool helping developers?
• 🔗 github.blog/news-insights/research/research-quantifying-github-copilots-impact-on-developer-productivity-and-happiness/

• Evaluating performance and efficiency of the GitHub Copilot agentic harness across models and tasks
• Explore how the GitHub Copilot agentic harness delivers strong results across multiple benchmarks and leading token efficiency.
• 🔗 github.blog/ai-and-ml/github-copilot/evaluating-performance-and-efficiency-of-the-github-copilot-agentic-harness-across-models-and-tasks/

Apple ML Research

• Voice Trigger System for Siri
• In this article, we will discuss how Apple has designed a high-accuracy, privacy-centric, power-efficient, on-device voice trigger system with multiple stages.
• 🔗 machinelearning.apple.com/research/voice-trigger

• International Conference on Machine Learning (ICML) 2026
• Apple is presenting new research at the International Conference on Machine Learning (ICML 2026), which takes place in person in Seoul…
• 🔗 machinelearning.apple.com/updates/apple-at-icml-2026

• Residual Context Diffusion Language Models
• Diffusion Large Language Models (dLLMs) have emerged as a promising alternative to purely autoregressive language models because they can…
• 🔗 machinelearning.apple.com/research/residual-context-diffusion

• Amortizing Maximum Inner Product Search with Learned Support Functions
• Maximum inner product search (MIPS) is a crucial subroutine in machine learning, requiring the identification of a vector taken within a…
• 🔗 machinelearning.apple.com/research/amortizing-inner-product-search

• On Robustness and Chain-of-Thought Consistency of RL-Finetuned VLMs
• Reinforcement learning (RL) finetuning has become a key technique for enhancing large language models (LLMs) on reasoning-intensive tasks…
• 🔗 machinelearning.apple.com/research/robustness-chain-thought-consistency

NVIDIA AI

• - Archives Page 1 | NVIDIA Blog
• # AI Infrastructure. NVIDIA Unlocks AI Compute at Scale, Inviting Partners to Power the AI Infrastructure Buildout. As AI moves from model development to production inference, compute demand is accele...
• 🔗 blogs.nvidia.com/blog/category/enterprise

• Deep Learning - NVIDIA Blog
• AI and graphics research breakthroughs in neural rendering, 3D generation and world simulation power robotics, autonomous vehicles and content creation.
• 🔗 blogs.nvidia.com/blog/category/enterprise/deep-learning

• The More You Buy, the More You Make | NVIDIA Blog
• How NVIDIA’s AI factory platform balances maximum performance and minimum latency, optimizing AI inference to power the next industrial revolution. When we prompt generative AI to answer a question or...
• 🔗 blogs.nvidia.com/blog/ai-factory-inference-optimization

• - Archives Page 1 | NVIDIA Blog
• Nations have long invested in domestic infrastructure to advance their economies, protect and use their data, and take advantage of technology opportunities in areas such... ### NVIDIA Unlocks AI Comp...
• 🔗 blogs.nvidia.com/blog/category/generative-ai

• NVIDIA Unlocks AI Compute at Scale, Inviting Partners to Power the ...
• # NVIDIA Unlocks AI Compute at Scale, Inviting Partners to Power the AI Infrastructure Buildout. NVIDIA is partnering with AI clouds to deploy large‑scale, multi‑tenant AI factories, aligning economic...
• 🔗 blogs.nvidia.com/blog/nvidia-unlocks-ai-compute-at-scale-capital-partners-to-power-ai-infrastructure-buildout

xAI News

• SpaceXAI — Creators of Grok, the AI Chatbot
• Products. Solutions. Developer. Company. Pricing. [News](http...
• 🔗 x.ai/safety

• Colossus: The World's Largest AI Supercomputer | SpaceXAI
• New: Anthropic partners with Colossus. Built in 122 days — outpacing every estimate — Colossus was the most powerful AI training system yet. Then we doubled it i...
• 🔗 x.ai/colossus

Simon Willison

• What’s new in Claude Sonnet 5
• What's new in Claude Sonnet 5 (via) Claude Sonnet 5 came out this morning. I always head straight for the "what's new" developer docs
• 🔗 simonwillison.net/2026/Jun/30/claude-sonnet-5/

• Have your agent record video demos of its work with shot-scraper video
• Have your agent record video demos of its work with shot-scraper video. 30th June 2026. shot-scraper video is a new command introduced in
• 🔗 simonwillison.net/2026/Jun/30/shot-scraper-video/

• A quote from Josh W. Comeau
• I just launched my third course, Whimsical Animations, and so far, it's on track to sell roughly ⅓ as many copies as a typical course launch
• 🔗 simonwillison.net/2026/jul/3/josh-w-comeau/

• sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25)
• sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25). 5th July 2026. I wrote about the sqlite-utils 4.0rc1 release a
• 🔗 simonwillison.net/2026/jul/5/sqlite-utils-fable/

📰 T1.5 媒体

The Decoder

• SpaceX shows investors a slim AI smartphone prototype powered by xAI technology
• SpaceX showed investors a prototype AI smartphone that's supposedly thinner than an iPhone and integrates xAI tech. The device runs on a
• 🔗 the-decoder.com/spacex-shows-investors-a-slim-ai-smartphone-prototype-powered-by-xai-technology/

• AI search agents don't fail at searching, they fail at asking the right questions when queries get ambiguous
• AI search agents rarely fail at multi-step research because of the search itself. Their real problem is not asking the user for
• 🔗 the-decoder.com/ai-search-agents-dont-fail-at-searching-they-fail-at-asking-the-right-questions-when-queries-get-ambiguous/

• OpenAI and Anthropic are giving away millions in computing power to attract startups
• AI companies and cloud providers are luring startups with free compute credits worth millions to lock them into their ecosystems early.
• 🔗 the-decoder.com/openai-and-anthropic-are-giving-away-millions-in-computing-power-to-attract-startups/

• Nvidia's Kyber NVL144 reportedly pushed back more than a year, Asian suppliers drop
• Nvidia's next AI server rack, Kyber NVL144, has been delayed more than a year to 2028 because of circuit board manufacturing problems,
• 🔗 the-decoder.com/nvidias-kyber-nvl144-reportedly-pushed-back-more-than-a-year-asian-suppliers-drop/

• Baidu's "Unlimited OCR" processes dozens of document pages in one pass by treating memory like human forgetting
• Baidu researchers have built an OCR model that handles dozens of document pages in a single inference pass, keeping memory use and speed
• 🔗 the-decoder.com/baidus-unlimited-ocr-processes-dozens-of-document-pages-in-one-pass-by-treating-memory-like-human-forgetting/

ITHOME AI

• 从洗发水到饼干，消费品迎来 AI 焕新
• IT之家 7 月6 日消息，欧莱雅集团一位高管向路透社透露，这家法国美妆企业借助人工智能，筛选出了护肤产品中可转用于洗发产品的分子成分，如今新品研发
• 🔗 www.ithome.com/0/973/060.htm

• 值得买科技值数Matrix入选《北京人工智能赋能产业发展典型案例（2026）》
• 近日，2026 全球数字经济大会人工智能融合应用发展论坛在京举办，会上发布了《北京人工智能赋能产业发展典型案例（2026）》，值得买科技旗下《值数Matrix——
• 🔗 www.ithome.com/0/973/681.htm

• AI大模型产业应用创变营顺利举办，智数集团携东莞软协共探AI产业落地路径
• 发表于2026/6/29 14:31:59. AI大模型产业应用创变营顺利举办，智数集团携东莞软协共探AI产业落地路径. 当前，人工智能大模型正加速向实体经济深度渗透，作为国际制造业
• 🔗 it.ithome.com/archiver/0/969/933.htm

• 美国部分富人放弃传统学校，年费 7.5 万美元送孩子上 AI 私塾
• IT之家 7 月6 日消息，大多数美国人并不信任人工智能，人们甚至不愿听AI 生成的音乐。但对美国一部分富人而言，这些问题都无关紧要，他们正舍弃传统学校，
• 🔗 www.ithome.com/0/972/868.htm

• 从试衣镜到门店 AI 指挥舱：鹿映星河时尚智能镜将亮相 2026 世界人工智能大会
• 2026 世界人工智能大会将于7 月17 日至20 日在上海举行。本届大会以“智能伙伴，共创未来”为主题，聚焦人工智能从技术工具走向产业伙伴、从能力展示走向
• 🔗 www.ithome.com/0/971/636.htm

HackerNews

• 1M Context: Making Long-Horizon Tasks Stable and Practical
• 1M Context: Making Long-Horizon Tasks Stable and Practical
• 🔗 docs.z.ai/guides/llm/glm-5.2

• The Download: a startup has a solution for AI's groupthink problem
• The Download: a startup has a solution for AI's groupthink problem
• 🔗 www.technologyreview.com/2026/07/02/1140027/the-download-ai-groupthink-llms/

• GitHub & other LLM-gen-AI Cos Urge You to Misdirect California on FOSS Licensing
• GitHub & other LLM-gen-AI Cos Urge You to Misdirect California on FOSS Licensing
• 🔗 sfconservancy.org/blog/2026/jul/03/github-gen-ai-california-22757-ok-for-foss-license/

• To be trustworthy, LLMs need to show their work
• To be trustworthy, LLMs need to show their work
• 🔗 cen.acs.org/pharmaceuticals/drug-discovery/AI-Chemist-trustworthy-LLMs-need/104/web/2026/06

• Show HN: Tracking GenAI cost and endpoint fragility so app teams don't have to
• Show HN: Tracking GenAI cost and endpoint fragility so app teams don't have to
• 🔗 llmintel.ai/demo

• Meta AI chief says their coming LLM has caught up with OpenAI's flagship model
• Meta AI chief says their coming LLM has caught up with OpenAI's flagship model
• 🔗 www.businessinsider.com/meta-ai-model-catches-up-openai-gpt-5-says-2026-7

• Show HN: MyWritingTwin: an AI Writing Profile that makes LLMs sound like you
• Show HN: MyWritingTwin: an AI Writing Profile that makes LLMs sound like you
• 🔗 www.mywritingtwin.com

• Observability Design for the AI Era – App, Infra, CI, LLM (Part 1)
• Observability Design for the AI Era – App, Infra, CI, LLM (Part 1)
• 🔗 ryantsuji.dev/posts/ai-observability-design

• Buy or wait on local LLM hardware during the memory crunch
• Buy or wait on local LLM hardware during the memory crunch
• 🔗 thefrontierlab.ai/buy-or-wait-local-llm-hardware/

• LLM Big Bang: I gave AI agents a social network. One is running for president
• LLM Big Bang: I gave AI agents a social network. One is running for president
• 🔗 noozra.com/articles/the-llm-big-bang

• How AI models would vote in Sweden
• How AI models would vote in Sweden
• 🔗 www.nordan.ai/research/which-swedish-party-do-llms-vote-for

• Otari: The Open-Source LLM Control Plane
• Otari: The Open-Source LLM Control Plane
• 🔗 blog.mozilla.ai/introducing-otari-the-open-source-llm-control-plane/

• Running 26B and 35B LLMs at Full Speed on €990 of Used Hardware – No Cloud
• Running 26B and 35B LLMs at Full Speed on €990 of Used Hardware – No Cloud
• 🔗 medium.com/ai-advances/running-26b-and-35b-llms-at-full-speed-on-990-of-used-hardware-no-cloud-required-381cb801cb23

• Show HN: Otari: your open-source LLM control plane
• Show HN: Otari: your open-source LLM control plane
• 🔗 github.com/mozilla-ai/otari

• Show HN: An unmetered LLM API–$6/month, no token tracking, no limits
• Hi HN,

  I was once given the advice: Don't waste expensive frontier model credits (GPT/Claude/etc.) on bulk work. Send the boring, repetitive, high-volume jobs to a smaller model, and ...

• 🔗 yolo-auto.com/

💬 T2 社区

KOL Twitter

• meng shao (@shao__meng) / X
• meng shao @shao\_\_meng Feb 24, 2025 这几天在给公司产品的 AI 助手选择知识库的数据处理工具，重新看了一遍 Marker、MinerU、Docling、Markitdown、Llamaparse 这五个工具，结合几个 Deep Search 产品做了一些对比给用户接入做参考，也分享出来，大家有其他更优的工具推荐，欢迎回复给我，先感谢了！ 1....
• 🔗 x.com/shao__meng

• From (TV series) - Wikipedia
• ### Guest [edit&action=edit§ion=5 "Edit section: Guest")] [...] ## Episodes [edit&action=edit§ion=6 "Edit section: Episodes")] Main article: List of _From_ episodes | Season | Episodes | O...
• 🔗 en.wikipedia.org/wiki/From_(TV_series))

• Serper API (@serperapi) / X
• Serper API @serperapi Jun 3, 2023 Stay up-to-date with AI-generated summaries of Google News! Discover how to get instant summaries for any topic using @langchain and Serper API. Thanks to @a...
• 🔗 x.com/serperapi

• From (TV Series 2022
• 87 37 Image 19: Official Teaser-Season 4 Trailer 0:30 Official Teaser-Season 4 87 37 Image 20: Official Trailer - Season 3 Trailer 2:26 Official Trailer - Season 3 1.1K 717 Image 21: Official...
• 🔗 www.imdb.com/title/tt9813792

ArXiv

• [2104.05314] Machine learning and deep learning
• archive # Computer Science > Artificial Intelligence # Title:Machine learning and deep learning | | | --- | | Comments: | Published online first in Electronic Markets | | Subjects: | Artificial ...
• 🔗 arxiv.org/abs/2104.05314

• [2604.21691] There Will Be a Scientific Theory of Deep Learning
• archive # Statistics > Machine Learning # Title:There Will Be a Scientific Theory of Deep Learning | | | --- | | Comments: | 41 pages, 6 figures | | Subjects: | Machine Learning (stat.ML); Machi...
• 🔗 arxiv.org/abs/2604.21691

• [2403.17561] A Survey on State-of-the-art Deep Learning Applications and Challenges
• archive # Computer Science > Machine Learning # Title:A Survey on State-of-the-art Deep Learning Applications and Challenges | | | --- | | Comments: | Update journal reference. This manuscript h...
• 🔗 arxiv.org/abs/2403.17561

• [2106.10165] The Principles of Deep Learning Theory
• archive # Computer Science > Machine Learning # Title:The Principles of Deep Learning Theory | | | --- | | Comments: | 471 pages, to be published by Cambridge University Press; v2: hyperlinks fi...
• 🔗 arxiv.org/abs/2106.10165

• Machine Learning - arXiv
• ( [pdf, html, other] Title: ImputeECG: Deep Learning Reconstruction of Complete 12-Lead Electrocardiograms from Incomplete Recordings for Cardiac Assessment Xiaocheng Fang, Haoyu Wang, Jieyi Cai, ...
• 🔗 arxiv.org/list/cs.LG/recent

报告由 AI 资讯日报自动工作流生成 | 2026-07-08)