When /ChatGPT/ would actually be so smart and help the small net

There are several approaches to operating /ChatGPT/ from a command line, and there is still much hype about /ChatGPT/ at all. Many are aware that this is still just a statistical model. It's operating on training text and it's providing a result as the most probable letters, words, and sentences for asked questions. Sometimes the result is impressive because instead of many traditional searches, we get a full-sentence answer to one of our questions in natural language. Sometimes after a while, we see the imperfections of this answer. And sometimes it is of sufficient quality.

Since today's large network is based largely on average-quality content, you can imagine how it fills up automatically based on responses from /ChatGPT/. Reportedly, legislative work is underway to [require marking of the automatic content generated].

However, I thought a truly smart /ChatGPT/ could be really helpful for a small network. In the same way as small-net proxies are providing small-net content in the big net, we could extract by AI the true content of the big net and automatically translate it to the gemtext.

AI would remove decorations, interactive scripts, adverts, cookie pop-ups, affiliate texts, and so on. AI would merge all scattered information into a single content. AI could detect the real sources of copy-pasted text, and show the original and nondistracted content. AI finally will lay out that data into a gemtext form. It would be a kind of magic.

[require marking of the automatic content generated]


szczezuja.space CC BY-SA

@ Thu 08 Jun 2023 11:25:45 AM CEST

tags: #chatgpt, #ai