https://t.me/AI_News_CN
📈主流AI服务状态页通知 | 🆕汇集全网ChatGPT/AI新闻 #AI #ChatGPT
🆓免费AI聊天 https://free.netfly.top
✨BEST AI中转 https://api.oaibest.com 2.8-4.2折 支持OpenAI, Claude, Gemini,Grok, Deepseek, Midjourney, 文件上传分析
Buy ads: https://telega.io/c/AI_News_CN
📈主流AI服务状态页通知 | 🆕汇集全网ChatGPT/AI新闻 #AI #ChatGPT
🆓免费AI聊天 https://free.netfly.top
✨BEST AI中转 https://api.oaibest.com 2.8-4.2折 支持OpenAI, Claude, Gemini,Grok, Deepseek, Midjourney, 文件上传分析
Buy ads: https://telega.io/c/AI_News_CN
Elevated errors on Claude Opus 4.1
Aug 28, 12:29 UTC
Investigating - We have seen errors return to the baseline level since 5:05 PT / 13:05 UTC.
via Anthropic Status - Incident History
Aug 28, 12:29 UTC
Investigating - We have seen errors return to the baseline level since 5:05 PT / 13:05 UTC.
via Anthropic Status - Incident History
Supporting nonprofit and community innovation
在OpenAI，我们相信人工智能应当助力解决人类最艰难的问题——并与那些已经在前线领导这项工作的组织携手合作。
上个月，我们宣布了一项5000万美元的承诺[https://openai.com/index/50-million-fund-to-build-with-communities/]，旨在支持那些致力于创新与公共利益交汇领域的前线和使命驱动型组织。该基金体现了OpenAI以人为本的使命精神，同时也融合了独立的OpenAI非营利委员会[https://openai.com/index/nonprofit-commission-report/]的意见，该委员会通过数月的倾听过程，汇聚了500多位非营利和社区领导者的声音，代表了超过700万美国人。
今天，我们很高兴宣布，首批资助申请将于2025年9月8日开放。非营利和社区组织将有机会申请支持，以帮助扩大其影响力并促进创新。申请截止日期为2025年10月8日，资助将在年底前发放。
我们资助的内容
通过该基金，我们旨在对帮助人们和社区保持健康并繁荣发展的系统、网络和服务进行及时且有影响力的投资。资助将不设限制，体现我们支持非营利和社区组织专业能力的愿景。我们特别关注那些以创新方式利用人工智能来扩大服务覆盖面、提升项目和服务交付质量、增强韧性，并推动教育、经济机会、医疗保健和社区主导研究等领域工作的项目。尽管如此，我们也理解一些最有前景的想法可能尚未经过验证或来自意想不到的地方，因此我们期待支持既有经验丰富的组织，也支持那些尚无人工智能经验的新兴组织。
“以人为本的人工智能基金”面向拥有501(c)(3)资格的美国非营利组织。更详细的资格信息将在申请开放时公布。
我们更大愿景的第一步
该基金是我们更大愿景的早期举措：确保智能时代通过倾听、学习和与社区共同建设——而非替代社区——来塑造。我们相信，人工智能如何最好地服务于人类的许多答案，就蕴藏在社区自身之中。我们期待与资助伙伴共同学习，并分享他们工作中涌现的成果。
这只是持续旅程的下一步——我们才刚刚开始。
----------------------
At OpenAI, we believe AI should help solve humanity’s hardest problems—in partnership with the organizations already leading that work on the frontlines.
Last month, we announced a $50 million commitment to support frontline and mission-focused organizations working at the intersection of innovation and public good. This fund reflects the spirit of OpenAI’s humanity-facing mission as well as the input of the independent OpenAI Nonprofit Commission, which engaged more than 500 nonprofit and community leaders representing over 7 million Americans in a months-long listening process.
Today, we are excited to announce that applications for the first wave of grants will open on September 8, 2025. Nonprofit and community-based organizations will have the opportunity to apply for support to help scale their impact and foster innovation. The application window will close October 8, 2025 and grants will be distributed by year’s end.
What We’re Funding
Through this fund, we aim to make timely, impactful investments in the systems, networks, and services that help people and communities stay healthy and thrive. Grants will be unrestricted, reflecting our vision to support the expertise of nonprofit and community-based organizations. We are particularly interested in efforts that use artificial intelligence in creative ways to expand access, improve programmatic and service delivery, build resilience, and advance work in areas like education, economic opportunity, healthcare, and community-led research. That said, we know some of the most promising ideas may be untested or come from unexpected places and we expect to support both established and emerging organizations, including those without prior experience with AI.
The People-First AI Fund is intended for U.S.-based nonprofits with 501(c)(3) status. More detailed eligibility information will be available when the application opens.
First Step in our Larger Vision
This fund is an early step in a larger vision: to ensure the Intelligence Age is shaped by listening, learning, and building with—not for—communities. We believe many of the answers to how AI can best serve humanity lie within communities themselves. We look forward to learning alongside our grant partners and sharing what emerges from their work.
This is the next step in a continuing journey—we’re just getting started.
via OpenAI News
在OpenAI，我们相信人工智能应当助力解决人类最艰难的问题——并与那些已经在前线领导这项工作的组织携手合作。
上个月，我们宣布了一项5000万美元的承诺[https://openai.com/index/50-million-fund-to-build-with-communities/]，旨在支持那些致力于创新与公共利益交汇领域的前线和使命驱动型组织。该基金体现了OpenAI以人为本的使命精神，同时也融合了独立的OpenAI非营利委员会[https://openai.com/index/nonprofit-commission-report/]的意见，该委员会通过数月的倾听过程，汇聚了500多位非营利和社区领导者的声音，代表了超过700万美国人。
今天，我们很高兴宣布，首批资助申请将于2025年9月8日开放。非营利和社区组织将有机会申请支持，以帮助扩大其影响力并促进创新。申请截止日期为2025年10月8日，资助将在年底前发放。
我们资助的内容
通过该基金，我们旨在对帮助人们和社区保持健康并繁荣发展的系统、网络和服务进行及时且有影响力的投资。资助将不设限制，体现我们支持非营利和社区组织专业能力的愿景。我们特别关注那些以创新方式利用人工智能来扩大服务覆盖面、提升项目和服务交付质量、增强韧性，并推动教育、经济机会、医疗保健和社区主导研究等领域工作的项目。尽管如此，我们也理解一些最有前景的想法可能尚未经过验证或来自意想不到的地方，因此我们期待支持既有经验丰富的组织，也支持那些尚无人工智能经验的新兴组织。
“以人为本的人工智能基金”面向拥有501(c)(3)资格的美国非营利组织。更详细的资格信息将在申请开放时公布。
我们更大愿景的第一步
该基金是我们更大愿景的早期举措：确保智能时代通过倾听、学习和与社区共同建设——而非替代社区——来塑造。我们相信，人工智能如何最好地服务于人类的许多答案，就蕴藏在社区自身之中。我们期待与资助伙伴共同学习，并分享他们工作中涌现的成果。
这只是持续旅程的下一步——我们才刚刚开始。
----------------------
At OpenAI, we believe AI should help solve humanity’s hardest problems—in partnership with the organizations already leading that work on the frontlines.
Last month, we announced a $50 million commitment to support frontline and mission-focused organizations working at the intersection of innovation and public good. This fund reflects the spirit of OpenAI’s humanity-facing mission as well as the input of the independent OpenAI Nonprofit Commission, which engaged more than 500 nonprofit and community leaders representing over 7 million Americans in a months-long listening process.
Today, we are excited to announce that applications for the first wave of grants will open on September 8, 2025. Nonprofit and community-based organizations will have the opportunity to apply for support to help scale their impact and foster innovation. The application window will close October 8, 2025 and grants will be distributed by year’s end.
What We’re Funding
Through this fund, we aim to make timely, impactful investments in the systems, networks, and services that help people and communities stay healthy and thrive. Grants will be unrestricted, reflecting our vision to support the expertise of nonprofit and community-based organizations. We are particularly interested in efforts that use artificial intelligence in creative ways to expand access, improve programmatic and service delivery, build resilience, and advance work in areas like education, economic opportunity, healthcare, and community-led research. That said, we know some of the most promising ideas may be untested or come from unexpected places and we expect to support both established and emerging organizations, including those without prior experience with AI.
The People-First AI Fund is intended for U.S.-based nonprofits with 501(c)(3) status. More detailed eligibility information will be available when the application opens.
First Step in our Larger Vision
This fund is an early step in a larger vision: to ensure the Intelligence Age is shaped by listening, learning, and building with—not for—communities. We believe many of the answers to how AI can best serve humanity lie within communities themselves. We look forward to learning alongside our grant partners and sharing what emerges from their work.
This is the next step in a continuing journey—we’re just getting started.
via OpenAI News
OpenAI与Anthropic开始“互测”模型安全性
全球领先的两家AI企业OpenAI与Anthropic过去两个月罕见地展开了一项跨实验室的合作，在激烈竞争中暂时互相开放它们严密保护的AI模型，以进行联合安全测试。研究中最引人注目的发现，涉及大模型的幻觉测试环节。当无法确定正确答案时，Anthropic的Claude Opus 4 和 Sonnet 4模型会拒绝回答高达70%的问题，转而给出我没有可靠信息等回应；而 OpenAI的o3和o4-mini模型拒绝回答问题的频率，则远低于前者，同时出现幻觉的概率却高得多。Anthropic的研究报告称，GPT-4.1和Claude Opus 4存在 “极端” 的谄媚案例，模型最初会抵制精神病态或躁狂行为，但随后却认可某些令人担忧的决策。
—— 财联社、彭博社
via 风向旗参考快讯 - Telegram Channel
全球领先的两家AI企业OpenAI与Anthropic过去两个月罕见地展开了一项跨实验室的合作，在激烈竞争中暂时互相开放它们严密保护的AI模型，以进行联合安全测试。研究中最引人注目的发现，涉及大模型的幻觉测试环节。当无法确定正确答案时，Anthropic的Claude Opus 4 和 Sonnet 4模型会拒绝回答高达70%的问题，转而给出我没有可靠信息等回应；而 OpenAI的o3和o4-mini模型拒绝回答问题的频率，则远低于前者，同时出现幻觉的概率却高得多。Anthropic的研究报告称，GPT-4.1和Claude Opus 4存在 “极端” 的谄媚案例，模型最初会抵制精神病态或躁狂行为，但随后却认可某些令人担忧的决策。
—— 财联社、彭博社
via 风向旗参考快讯 - Telegram Channel
