人脑在20瓦下处理信息的高效率与当前人工智能系统所需的巨大功率形成对比，后者需要核反应堆才能运行。最近的进展包括与活体脑细胞接口的打印技术，这一发展凸显了生物智能和人工智能在能耗和能力方面的巨大差距。 AI
影响 强调了当前人工智能巨大的能源需求，突显了对更高效硬件和算法的必要性。
排序理由 该集群讨论了人脑与人工智能的能效对比，并以人工智能基础设施投资为例，但并未发布新的模型、产品或研究突破。
在 Mastodon — fosstodon.org 阅读 →
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
🤖 Your brain does on 20 watts what AI needs a nuclear reactor to attempt. Last week a team figured out how to print something that actually speaks to living brain cells. Amazon bought a 960 megawatt nuclear reactor for AI servers. Microsoft restarted Three Mile Island. Stargate i…
📰 A Fume-Control Cabinet for Resin 3D Printing For a certain kind of intricate, highly-detailed manufacturing, there’s really no substitute for a resin 3D printer, and it’s therefore unfortunate that they require so many poisonous chemicals. Th... 📰 Source: Hackaday 🔗 Link: https…