4 ms·

Perfect use case for chatgpt or other llm

pseudalopex 2y ago
Are you serious? It's the primary use for a dictionary. And did you try ChatGPT? It suggested bananas heavy meant bananas are physically heavy because they are ripe.
- jondwillis 2y ago
  dictionaries do not provide much context that would un-confuse someone who has not come into contact with a colloquialism.
  - pseudalopex 2y ago
    The context is bananas means crazy. Dictionaries provide this context.
    - donatj 2y ago
      It would not have helped me in this situation. My assumption was "bananas heavy" was a phrase I was unfamiliar with, so I was trying to look that up. "Bananas heavy" is not in the dictionary, I fully understood that bananas can mean crazy, but what I do not understand is the sentence and the context. I never would have used the word with that meaning like this. Who would say something like this? I still don't understand why a person would say "bananas heavy" to mean "YouTube embeds are bloated". It just sounds so awkward and bizarre.
      - popcalc 2y ago
        bananas == crazy; (x heavy) == crazy heavy
- marcosdumay 2y ago
  Yet, when I put "What's the meaning of "bananas heavy"?" in perplexity, it explains it perfectly well. One of the easiest things for a LLM to "memorize" is the dictionary. And it makes for a much more flexible one. It even suggest additional queries where it explains why "bananas" is used on that phrase. The only problem I found on the query is that perplexity seems to be de-emphasizing its sources, what IMO is bananas. It just pushes the site into direct competition with ChatGPT and removes most of its value.
- spencerchubb 2y ago
  Yes I'm serious. Here's the prompt I would use: "what does bananas mean in this sentence: YouTube embeds are bananas heavy and it’s fixable" Put it in google, and it doesn't give anything useful Put it in an LLM, and you get a correct answer because LLMs can understand full sentences
  - kmoser 2y ago
    LLMs don't actually "understand" anything. They're just better at parsing and manipulating language.

Are you serious? It's the primary use for a dictionary. And did you try ChatGPT? It suggested bananas heavy meant bananas are physically heavy because they are ripe.

jondwillis 2y ago
dictionaries do not provide much context that would un-confuse someone who has not come into contact with a colloquialism.
- pseudalopex 2y ago
  The context is bananas means crazy. Dictionaries provide this context.
  - donatj 2y ago
    It would not have helped me in this situation. My assumption was "bananas heavy" was a phrase I was unfamiliar with, so I was trying to look that up. "Bananas heavy" is not in the dictionary, I fully understood that bananas can mean crazy, but what I do not understand is the sentence and the context. I never would have used the word with that meaning like this. Who would say something like this? I still don't understand why a person would say "bananas heavy" to mean "YouTube embeds are bloated". It just sounds so awkward and bizarre.
    - popcalc 2y ago
      bananas == crazy; (x heavy) == crazy heavy
marcosdumay 2y ago
Yet, when I put "What's the meaning of "bananas heavy"?" in perplexity, it explains it perfectly well. One of the easiest things for a LLM to "memorize" is the dictionary. And it makes for a much more flexible one. It even suggest additional queries where it explains why "bananas" is used on that phrase. The only problem I found on the query is that perplexity seems to be de-emphasizing its sources, what IMO is bananas. It just pushes the site into direct competition with ChatGPT and removes most of its value.
spencerchubb 2y ago
Yes I'm serious. Here's the prompt I would use: "what does bananas mean in this sentence: YouTube embeds are bananas heavy and it’s fixable" Put it in google, and it doesn't give anything useful Put it in an LLM, and you get a correct answer because LLMs can understand full sentences
- kmoser 2y ago
  LLMs don't actually "understand" anything. They're just better at parsing and manipulating language.