- cross-posted to:
- hackernews@lemmy.smeargle.fans
- cross-posted to:
- hackernews@lemmy.smeargle.fans
There is a discussion on Hacker News, but feel free to comment here as well.
There is a discussion on Hacker News, but feel free to comment here as well.
I agree too much with the text to comment anything meaningful about it. So let’s see the comments…
Both comments reminded me a blogpost that I wrote more than a year ago, regarding chatGPT-3. It still applies rather well to 2024 LLMs, and it shows what those two tech bros are missing, so I’ll copypaste it here.
###The problem with GPT3.
Consider the following two examples.
Example A.
GPT3 bots trained on the arsehole of the internet (Reddit), chatting among themselves:
The grammar is fine, and yet those messages don’t say jack shit.
Example B.
Human translation made by someone with not-so-good grasp of the target language.
The grammar is so broken that this excerpt became a meme. And yet you can still retrieve meaning from it:
What’s the difference? It’s purpose. In (B) we can give each utterance a purpose, even if the characters are fictional - because they were written by a human being. However, we cannot do the same in (A), because the current AI-generated text does not model that purpose.
And yes, assigning purpose to your utterances is part of the language. Not just what tech bros are able to see, namely: syntax, morphology, and spelling.
If reddit is the arsehole of the Internet, what is 4chan?
4chan was always called the asshole of the internet, but it’s more like the mouth of an extremely drunk internet ready to vomit on you.
No, it’s far worse than that.