this post was submitted on 21 Mar 2025
14 points (79.2% liked)
Technology
67422 readers
4158 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- 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, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
i suspect we are going to have a semantic disagreement on what "understanding" means here.
ChatGPT is absolutely trained on the concept of calendars, that iPhones indeed have calendars, and how they work. It doesn't need to "understand" what a calendar is on a deep epistemological level to process requests about them. If you ask chatGPT a question about calendars, it'll answer you. So in that shallower sense LLMs absolutely "understand" what you mean, and that's enough for chatGPT to help siri.
No, they really don't. They answer you with words that are commonly found in conjunction with calendars. There is code in Siri that understands calendars, but the LLM part ain't it. I have, ahem, firsthand knowledge of how Siri does this.
yes i know, you are getting hung up on my colloquial use of "understand". the LLM doesn't need to "understand" it on that level because siri does. the LLM is there to parse the language and hand off to Siri. that's all i'm saying.
Fair. I think it's important from an "Is this True AGI?" sense to distinguish these, but yes, in the colloquial sense I guess the system could be said to understand, even if it's not strictly the actual LLM part that does it.