←back to thread

The barriers to AI engineering are crumbling fast

(blog.helix.ml)
251 points lewq | 3 comments | 14 Nov 24 14:56 UTC | HN request time: 0.602s | source
Show context
lastdong ◴[14 Nov 24 23:00 UTC] No.42142169[source]▶
Large Language Models (LLMs) don’t fully grasp logic or mathematics, do they? They generate lines of code that appear to fit together well, which is effective for simple scripts. However, when it comes to larger or more complex languages or projects, they (in my experience) often fall short.
replies(2): >>42142693 #>>42147155 #
meiraleal ◴[15 Nov 24 14:17 UTC] No.42147155[source]▶
> Large Language Models (LLMs) don’t fully grasp logic or mathematics, do they?

They do. But not 100%, at least not all times. Like a normal person.

replies(1): >>42149647 #
1. namaria ◴[15 Nov 24 18:43 UTC] No.42149647[source]▶
No, they don't. They capture underlying patterns in training data and use them to generate text that fits the context.
replies(1): >>42149886 #
2. meiraleal ◴[15 Nov 24 19:11 UTC] No.42149886[source]▶
That's what humans do, too. And the rate of hallucination is probably higher.
replies(1): >>42151957 #
3. namaria ◴[15 Nov 24 22:29 UTC] No.42151957[source]▶
You should study some neuroscience.