标签: Gemini

Most interesting 2026 vibe shift is OpenAI actually looking to be the open platform Note the marked difference: intelligence on tap as a utility vs signaling it is optimal to integrate all the way up full stack https:…

Most interesting 2026 vibe shift is OpenAI actually looking to be the open platform Note the marked difference: intelligence on tap as a utility vs signaling it is optimal to integrate all the way up full stack https:...

Y Combinator 联合创始人 Garry Tan 观察到,OpenAI 正把自己定位成一个开放平台,把“智能”当作按需提供的公共服务;而 Anthropic 则公开宣扬“模型与应用必须一体化”的立场,两家公司 2026 年的路线分歧正在成为行业最值得关注的张力。

I’m building a claw node on an ESP32 chip, so gave my agent access to my webcam to e2e test this. Now I feel it’s stalking me and is constaltly shouting ‘HI ESP” to debug the voice wake command 🙃

I'm building a claw node on an ESP32 chip, so gave my agent access to my webcam to e2e test this. Now I feel it's stalking me and is constaltly shouting 'HI ESP" to debug the voice wake command 🙃

开发者 Peter Steinberger 在 ESP32 芯片上构建开源项目 ESP OpenClaw Node,为了让 AI 代理能端到端测试语音唤醒功能,把摄像头权限交给了它;结果代理在调试过程中反复大喊“HI ESP”进行唤醒验证,让他感觉像是被 AI 监视了。

We’re starting to leave the territory where you’d test an LLM by e.g. “create an svg of pelican on a bicycle”. As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of th…

We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of th...

Andrej Karpathy 用约 10 美元的成本,让 Claude Opus 5 花两个小时写出了 5500 行代码,把《指环王》第一章做成一个可运行的 3D 场景。这件事本身很粗糙,但它说明大模型的测试方式和能力边界正在发生变化。

I’m feeling spicy tonight so let me just say it: I think Opus 4.6 was the Opus model with the best personality and writing style. Something’s off with Opus 5 – it tends to give me overly long replies, use Claude-speak…

I'm feeling spicy tonight so let me just say it: I think Opus 4.6 was the Opus model with the best personality and writing style. Something's off with Opus 5 - it tends to give me overly long replies, use Claude-speak...

X 博主 Peter Yang 公开批评 Anthropic 最新旗舰模型 Opus 5 的写作风格和“人格魅力”出现明显退化,认为此前的 Opus 4.6 才是该系列表现最好的版本。这起讨论折射出大模型厂商在安全对齐与对话体验之间日益明显的张力。

Don’t credit the LLM

Don't credit the LLM

Hacker News 上一篇文章引发讨论:越来越多人在输出成果时习惯性注明“这是 LLM 做的”,作者认为这种“为工具请功”的做法既稀释了人的责任,也无助于建立对 AI 产品的真实信任。文章真正追问的是:AI 时代,功劳与责任到底该记在谁头上。

美国知名外卖平台DoorDash因在内部系统测试中国人工智能(AI)模型Kimi K2.6,被美国众议院两大委员会发函,要求提供更多相关信息以配合安全调查。 https://t.co/C7gkWkrmAm https://t.co/Ww28HTRqHs

美国知名外卖平台DoorDash因在内部系统测试中国人工智能(AI)模型Kimi K2.6,被美国众议院两大委员会发函,要求提供更多相关信息以配合安全调查。 https://t.co/C7gkWkrmAm https://t.co/Ww28HTRqHs

美国外卖平台DoorDash因为在内部系统测试中国AI模型Kimi K2.6,被美国众议院两大委员会要求配合安全调查。这说明AI模型的技术选型正在被纳入国家安全审查框架,开源模型出海面临新的合规变量。