TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

OpenAI still not training GPT-5, Sam Altman says

3 点作者 diwakar2008将近 2 年前

1 comment

colesantiago将近 2 年前
Why even mention &#x27;GPT-5&#x27; to begin with?<p>&gt; &quot;We have a lot of work to do before we start that model,&quot; Altman, the chief executive of OpenAI, said at a conference hosted by Indian newspaper Economic Times. “We’re working on the new ideas that we think we need for it, but we are certainly not close to it to start.&quot;<p>I&#x27;m not buying this, either they:<p>Completely finished training GPT-5 already (GPT-3.5 &#x2F; 4&#x27;s knowledge cut off is in Sept 2021)<p>Finished training GPT-5 for now but not training it further.<p>Finished training GPT-5 already and are fine tuning it to comply with regulations and updating old data.<p>Didn&#x27;t start training GPT-5 and are waiting for regulations.<p>Even if either of these are true, it seems that OpenAI knows (or are marketing that) this will be possibly a dangerous model, so only enterprises may have access to this at first.<p>(Remember, GPT-4 has a image input model that is only available to certain organisations)