773
this post was submitted on 23 Apr 2024
773 points (98.7% liked)
Technology
59596 readers
5292 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I’ll agree there’s huge potential for ‘assistant’ roles (exactly like you’re using) to give a concise summary for quick understanding. But LLMs aren’t knowledgeable like an accredited professor or tutor is, understanding the context and nuance of the topic. LLMs are very good at scraping together data and presenting the shallowest of information, but their limits get exposed quickly when you try to go into a topic.
For instance, I was working a project that required very long term storage (+10 years) with intermittent exposure to open air, and was concerned about oxidation and rust. ChatGPT was very adamant that desiccant alone was sufficient (wrong) and that VCI packs would last (also wrong). It did a great job of repackaging corporate ad-copy and industrial white papers written by humans, but not of providing an objective answer to a semi complex question.
I guess it's not great for things requiring domain knowledge. Programming seems to be easy for it, as programs are very structured, predictable, and logical. That's where its pattern-matching-and-prediction abilities shine.