<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>算力浪费 on AI内参</title>
    <link>https://www.neican.ai/tags/%E7%AE%97%E5%8A%9B%E6%B5%AA%E8%B4%B9/</link>
    <description>Recent content in 算力浪费 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Mon, 06 Jul 2026 08:10:09 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/%E7%AE%97%E5%8A%9B%E6%B5%AA%E8%B4%B9/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>万亿算力白烧了？OpenAI Scaling Law论文被曝致命bug，整个AI圈被“带偏”好几年！</title>
      <link>https://www.neican.ai/insights/openai-scaling-lawbugai-20260706081009171-11/</link>
      <pubDate>Mon, 06 Jul 2026 08:10:09 +0800</pubDate>
      <guid>https://www.neican.ai/insights/openai-scaling-lawbugai-20260706081009171-11/</guid>
      <description>OpenAI的Scaling Law原始论文被发现存在致命bug，导致全球AI行业在“体量过大、训练不足”的模型上白白烧掉了万亿算力。前研究员Diogo Almeida揭示了论文中数据分配不公、学习率人为掐断等三招“误导”，并引发对AI行业盲目堆参数的反思。文章以轻松调侃的方式剖析了这一科技“塌房”事件。</description>
    </item>
  </channel>
</rss>
