https://news.945688.xyz/tags/ai/ Recent content in AI on GNEWS Hugo en Sat, 18 Apr 2026 11:55:00 +0800 https://news.945688.xyz/posts/why-benchmark-infrastructure-now-matters/ Sat, 18 Apr 2026 11:55:00 +0800 https://news.945688.xyz/posts/why-benchmark-infrastructure-now-matters/ A hiring post from ARC Prize Foundation may look routine on the surface. In reality, it points to a larger shift: frontier AI competition is no longer just about training bigger models, but about building better evaluation systems to measure whether those models can actually generalize. https://news.945688.xyz/posts/even-cat-readme-is-not-safe/ Sat, 18 Apr 2026 11:50:00 +0800 https://news.945688.xyz/posts/even-cat-readme-is-not-safe/ A new wave of security research argues that even reading a local README file may no longer be a harmless action in AI-assisted development environments. The reason is simple: modern coding agents do not just read text—they interpret, plan, and act on it. https://news.945688.xyz/posts/claude-opus-4-7-released/ Thu, 16 Apr 2026 22:30:00 +0800 https://news.945688.xyz/posts/claude-opus-4-7-released/ Anthropic 正式发布 Claude Opus 4.7。新模型在复杂软件工程任务、长时推理、多步代理工作流和高分辨率视觉理解方面进一步强化，同时维持与 Opus 4.6 相同的 API 定价。 https://news.945688.xyz/posts/ai-agents-future/ Thu, 16 Apr 2026 18:00:00 +0800 https://news.945688.xyz/posts/ai-agents-future/ 传统软件长期依赖图形界面与人工操作流程，而具备规划、记忆与执行能力的 AI Agent 正在推动人机交互向“意图驱动”迁移。这一变化不仅影响产品形态，也在重构企业软件的组织方式。 https://news.945688.xyz/posts/llm-under-the-hood-1-transformer/ Wed, 15 Apr 2026 20:00:00 +0800 https://news.945688.xyz/posts/llm-under-the-hood-1-transformer/ 大多数现代大语言模型都建立在 Transformer 架构之上。理解自注意力机制、位置编码和多头注意力，是理解 LLM 为什么能处理长上下文、建立语义关联和进行复杂生成的基础。 https://news.945688.xyz/posts/llm-under-the-hood-2-training/ Wed, 15 Apr 2026 19:00:00 +0800 https://news.945688.xyz/posts/llm-under-the-hood-2-training/ 用户看到的 ChatGPT 或 Claude，往往表现为一个可以对话、解释和执行任务的助手。但在成为助手之前，它首先是一个通过海量语料进行预训练的基础模型，随后再经过监督微调与人类反馈强化学习被塑造成今天的产品形态。 https://news.945688.xyz/posts/openai-gpt6-rumors/ Wed, 15 Apr 2026 09:30:00 +0800 https://news.945688.xyz/posts/openai-gpt6-rumors/ 围绕 OpenAI 下一代基础模型 GPT-6 的讨论再次升温。尽管官方尚未披露具体产品路线，但从行业动态与高层表态来看，长期记忆、规划能力与更稳定的推理过程，可能成为下一阶段竞争焦点。