Brett Winton: 大型语言模型目前仅生成和摄取文本但很快就会摄取并生成带有内联图像的文本到 1992 年与 Web 的韵律成为纯文本，然后在 93/94 年通过 Mosaic/Netscape 成为带有内联图像的文本

Posted on 2023-01-18

原推：That large language models currently generate and ingest text-only

but will soon ingest and generate text with inline images

Rhymes with the Web as text-only through 1992 then becoming text with inline images via Mosaic/Netscape in 93/94