this post was submitted on 05 Jul 2023
579 points (90.7% liked)
Technology
59735 readers
3471 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 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Google is just working hard to become less useful every day. I tend to use ChatGPT a lot instead of Google.
That seems rather risky, considering that they don't really check that they output accurate information, and OpenAI specifically recommends against using it for that due to the possibility of their GPT models outputting falsehoods as fact.
As opposed to Google searching manually, which always has accurate outputs and never outputs falsehoods as fact. π
As long as you double check the source of an answer I don't see an issue.
If you're double-checking the sources, both to make sure that they exist, and they are accurate, you may as well do the research without using an LLM in the first place.
You're just adding to your workload unnecessarily in that case.
I'm thinking that double checking is slightly faster than doing the research yourself.
Thatβs not necessarily true, LLMs are a very useful research aid because they give you a place to start. They give you a high level summary and may cite sources which may/may not exist, and itβs up to you to fact check and develop your own opinions. They can also summarize complex texts and filter out SEO garbage that would otherwise clutter google search results. Research still works the same as before, LLMs are just an accelerator if used correctly.
ChatGPT is not a search engine, it's a language model. It does not provide correct answers, it imitates what correct answers sound like.
The issue with LLMs that I have is that while they are great at certain tasks, they are bad at anything, let's call it factual, due to their nature.
I can for example use it to quickly draft up a email or a piece of python code, and I can immediately see whether or not the response it generated is actually what I want.
If I go ask it what the hottest day in a given country was or ask it to explain something, I have absolutely no idea whether it's bullshit or not and I will have to double check it anways.
I think the learning curve with LLMs as a tool is to be able to know when to use it and when to rely on other sources instead.