this post was submitted on 27 Jun 2023
393 points (99.5% liked)
Technology
59587 readers
3197 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
Not gonna happen at it's current state. I really can't wait for the "AI" trend to die. People think it's some magical resolution when it's really "As an AI language model 2+2=5".
You have to have some idea as to what the answer might already be
it doesnt provide sources etc. which is a bit annoying.
the chatgpt "bot" in the edge browser is actually decent at providing sources, but its terrible at finding specific info. i tried finding information about what TIME a certain game would be available to play, and it just kept giving me its release date. (which i also gave it, in my query)
but they're still very new.
google, and such, have had over 2 decades to refine their search etc. and to be honest i think the issue is its not giving you "generic" results. its trying to specify the results based on your previous searches etc. which means it can be difficult to find new info...
as for chatgpt ..
i use chatgpt quite often to summarize a large blob of text, in a simple manner, or give me code snippets for generic stuff im too lazy to write. test-data as well. or just "facts" about some topic. simple stuff.
chatgpt works by looking at your query, and then based on that, it tries to find the queries "key words" and fetches some result based on that. It only gives you one result. and as we've all tried when searching the internet, often times the list of results will show stuff that is clearly not what we're looking for.
For simple coding it is a dream. Or like, shitty DNS errors that need to be sorted out because apparently you can't have 2 SPF records lol. I copy and pasted all of the records over and said WHAT IS WRONG lol, and it figured it out for me.
I get that some people don't like it, but... its not going anywhere.
I second that, it's been very useful for coding/debugging for me too. And the cool part is that it's only going to get better.
Exactly - this is the worst it will ever be.
And no matter how well GPT will or won't work, we shouldn't forget who we're dealing with here. I wholehardly trust Microsoft to completely screw up and shit the bed once they got a (semi) monopoly on AI assistants. Google worked pretty well for a very long time until they got too cocky, same with Internet Explorer from Microsoft. History will repeat itself and GPT will become shit one way or antoher.
Machine learning is here to stay. This is really just the beginning of mass market adoption for it, there's still a lot of room for the tech to grow.
I really don't think your representation is fair. For Chat gpt at least, it will sometimes be wrong, it will sometimes make things up, but is an extremely useful tool for getting quick answers and meaningful insight into questions.
But if we know that it makes things up and gets things wrong, how can we trust any information it gives us? Fact-checking is one thing, but at that point, you might as well skip the LLM and just look the information up yourself.
At the end of the day you can't 100% trust anything you see on the internet. You have to think critically about the answers it gives you and cross reference it against other sources. No different than when evaluating search results, which can also be wrong. But it's a great starting point.
It's a lot easier to get a thorough and concise answer from chat gpt and double check it than it is to wade through a search engine.
Its more like "As an ai llm, 2+2=4. Now invest 500 billion in my creators company."
I think you vastly underestimate the impact machine learning and large language models are going to have on our society. It's like saying "I really can’t wait for the Smartphone trend to die" in 2007. Or "I really can’t wait for the Google trend to die" in 2000.
All aspects of our lives are going to be infiltrated by machine learning and large language models. Personal organization, work, grocery shopping, entertainment... Everything!
The hype around it is pretty insufferable though, in a way neither of the other examples you gave had.
The closest example I can think of is NFTs.
I don't think it'll go the way of NFTs, but it's also going to disappoint people because it's promising to be everything for everyone.
As far as I'm concern it's a very powerful search assistant and especially for bridging the gap between regular and power users - being able to use natural language is a game changer.
I also found it great when getting set up with a new piece of SW, and rephrasing or summarising text on general topics. It's not so good for parsing specialist information even when asked for specific items.
I'm looking forward to seeing what other tools people build with it but thus far I've been thoroughly... whelmed.
Speaking from personal experience, it was obvious NFTs would go nowhere and large language models would succeed. They've both been hyped by the public, but one of them has the utility to back up the hype and the other doesn't.
I use chat gpt all the time. I use it at work, i use it for looking up recipes, I use it to help with DIY projects around the house, and I use it to just get more information about a niche topic. The results are catered specifically to me and my question, and they're better than a search engine. This tech is only going to get more common from here.