I’m going to limit to LLMs as that’s the generally accepted term and there’s so many uses for AI in other fields that it’d be unfair.
Translation. LLMs are pretty much perfect for this.
Triaging issues for support. They’re useless for coming to solutions but as good as humans without the need to wait at sending people to the correct department to deal with their issues.
Finding and fixing issues with grammar. Spelling is something that can be caught by spell-checkers, but grammar is more context-aware, another thing that LLMs are pretty much designed for, and useful for people writing in a second language.
Finding starting points to research deeper. LLMs have a lot of data about a lot of things, so can be very useful for getting surface level information eg. about areas in a city you’re visiting, explaining concepts in simple terms etc.
Recipes. LLMs are great at saying what sounds right, so for cooking (not so much baking, but it may work) they’re great at spitting out recipes, including substitutions if needed, that go together without needing to read through how someone’s grandmother used to do xyz unrelated nonsense.
There’s a bunch more, but these were the first five that sprung to mind.
Translation. Only works for unified technical texts. The older non-LLM translation is still better for any general text and human translation for any fiction is a must. Case in point: try to translate Severance TV show transcript to another language. The show makes a heavy use of “Innie/Outie” language that does not exist in modern English. LLM fail to translate that - human translator would be able to find a proper pair of words in the target language.
Triaging issues for support. This one is a double-edged sword. Sure you can triage issues faster with LLM, but other people can also write issues faster with their LLMs. And they are winning more. Overall, LLM is a net negative on your triage cost as a business because while you can process each one faster than before, you are also getting way higher volume of those.
Grammar. It fails in that. I asked LLM about “fascia treatment” but of course I misspelled “fascia”. The “PhD-level” LLM failed to recognize the typo and gave me a long answer about different kinds of “facial treatment” even though for any human the mistake would’ve been obvious. Meaning, it only corrects grammar properly when the words it is working on are simple and trivial.
Starting points for deeper research. So was the web search. No improvement there. Exactly on-par with the tech from two decades ago.
Recipes. Oh, you stumbled upon one of my pet peeves! Recipes are generally in the gutter on the textual Internet now. Somehow a wrong recipe got into LLM training for a few things and now those mistakes are multiplied all over the Internet! You would not know the mistakes if you did not not cook/bake the thing previously. The recipe database was one of the early use cases for the personal computers back in 1990s and it is one of the first ones to fall prey to “innovation”. The recipes online are so bad, that you need an LLM to distill it back to manageable instructions. So, LLM in your example are great at solving the problem they created in the first place! You would not need LLM to get cooking instructions out of 1990s database. But early text generation AIs polluted this section of the Internet so much, that you need the next generation AI to unfuck it. Tech being great at solving the problem it created in the first place is not so great if you think about it.
You’re bringing up edge cases for #1, and it should be replacing google translate and basic human translation, eg allowing people to understand posts online or communicate textually with people with whom they don’t share a common language. Using it for anything high stakes or legal documents is asking for trouble though.
For 2, it’s not for AIs finding issues, it’s for people wanting to book a flight, or seek compensation for a delayed flight, or find out what meals will be served on their flight. Some people prefer to use text or voice communication over a UI, and this makes it easier to provide.
For 3, grammar and spelling are different. I said it wasn’t useful for spellcheck, but even then if you give it the right context it may or may not catch it. I was referring more to word order and punctuation positioning.
For 4, yeah for me it’s on par in terms of results, but much much faster, especially when asking followup questions or specifying constraints. A lot of people aren’t search engine powerusers though, so will find it significantly easier, faster and better than conventional search than having to manage tabs or keep track of what you’ve seen without just scrolling back up in the conversation.
For 5, recipes have been in the gutter for a decade or more now, SEO came before LLMs, but yeah, you’ve actually caught on to an obvious #6 I missed here of text summarisation…
What I’m getting overall though is that you’re not considering how tech-savvy the average person is, which absolutely makes them seem less useful as the more tech savvy you are, both the more you’re aware of their weaknesses and the less you benefit from the speedup by simplification they bring.
you’re not considering how tech-savvy the average person is
I am actually having hard time understanding where all of that hype is coming from. The first time I’ve seen AI solve a problem better than a human was back in 1996. I have used various generations of AI tools ever since. LLMs are fun, but it is not like they are that much different from the other AI tools before them. Every time a new AI technology comes around I am finding a use case for it in my own flow. LLMs have their uses as well. But I am not trying to solve ALL the problems with the new tech.
I do not understand “the average person”. And I guess I never will.
Nice, here’s a gold star for finding one case of it doing something wrong. I’ll call the CEO of AI and tell them to call it off, it’s a good thing humans have never said anything like that!
1rre@discuss.tchncs.de 3 weeks ago
I’m going to limit to LLMs as that’s the generally accepted term and there’s so many uses for AI in other fields that it’d be unfair.
Translation. LLMs are pretty much perfect for this.
Triaging issues for support. They’re useless for coming to solutions but as good as humans without the need to wait at sending people to the correct department to deal with their issues.
Finding and fixing issues with grammar. Spelling is something that can be caught by spell-checkers, but grammar is more context-aware, another thing that LLMs are pretty much designed for, and useful for people writing in a second language.
Finding starting points to research deeper. LLMs have a lot of data about a lot of things, so can be very useful for getting surface level information eg. about areas in a city you’re visiting, explaining concepts in simple terms etc.
Recipes. LLMs are great at saying what sounds right, so for cooking (not so much baking, but it may work) they’re great at spitting out recipes, including substitutions if needed, that go together without needing to read through how someone’s grandmother used to do xyz unrelated nonsense.
There’s a bunch more, but these were the first five that sprung to mind.
voronaam@lemmy.world 3 weeks ago
Translation. Only works for unified technical texts. The older non-LLM translation is still better for any general text and human translation for any fiction is a must. Case in point: try to translate Severance TV show transcript to another language. The show makes a heavy use of “Innie/Outie” language that does not exist in modern English. LLM fail to translate that - human translator would be able to find a proper pair of words in the target language.
Triaging issues for support. This one is a double-edged sword. Sure you can triage issues faster with LLM, but other people can also write issues faster with their LLMs. And they are winning more. Overall, LLM is a net negative on your triage cost as a business because while you can process each one faster than before, you are also getting way higher volume of those.
Grammar. It fails in that. I asked LLM about “fascia treatment” but of course I misspelled “fascia”. The “PhD-level” LLM failed to recognize the typo and gave me a long answer about different kinds of “facial treatment” even though for any human the mistake would’ve been obvious. Meaning, it only corrects grammar properly when the words it is working on are simple and trivial.
Starting points for deeper research. So was the web search. No improvement there. Exactly on-par with the tech from two decades ago.
Recipes. Oh, you stumbled upon one of my pet peeves! Recipes are generally in the gutter on the textual Internet now. Somehow a wrong recipe got into LLM training for a few things and now those mistakes are multiplied all over the Internet! You would not know the mistakes if you did not not cook/bake the thing previously. The recipe database was one of the early use cases for the personal computers back in 1990s and it is one of the first ones to fall prey to “innovation”. The recipes online are so bad, that you need an LLM to distill it back to manageable instructions. So, LLM in your example are great at solving the problem they created in the first place! You would not need LLM to get cooking instructions out of 1990s database. But early text generation AIs polluted this section of the Internet so much, that you need the next generation AI to unfuck it. Tech being great at solving the problem it created in the first place is not so great if you think about it.
1rre@discuss.tchncs.de 3 weeks ago
You’re bringing up edge cases for #1, and it should be replacing google translate and basic human translation, eg allowing people to understand posts online or communicate textually with people with whom they don’t share a common language. Using it for anything high stakes or legal documents is asking for trouble though.
For 2, it’s not for AIs finding issues, it’s for people wanting to book a flight, or seek compensation for a delayed flight, or find out what meals will be served on their flight. Some people prefer to use text or voice communication over a UI, and this makes it easier to provide.
For 3, grammar and spelling are different. I said it wasn’t useful for spellcheck, but even then if you give it the right context it may or may not catch it. I was referring more to word order and punctuation positioning.
For 4, yeah for me it’s on par in terms of results, but much much faster, especially when asking followup questions or specifying constraints. A lot of people aren’t search engine powerusers though, so will find it significantly easier, faster and better than conventional search than having to manage tabs or keep track of what you’ve seen without just scrolling back up in the conversation.
For 5, recipes have been in the gutter for a decade or more now, SEO came before LLMs, but yeah, you’ve actually caught on to an obvious #6 I missed here of text summarisation…
What I’m getting overall though is that you’re not considering how tech-savvy the average person is, which absolutely makes them seem less useful as the more tech savvy you are, both the more you’re aware of their weaknesses and the less you benefit from the speedup by simplification they bring.
voronaam@lemmy.world 3 weeks ago
I think you are correct at the main point:
I am actually having hard time understanding where all of that hype is coming from. The first time I’ve seen AI solve a problem better than a human was back in 1996. I have used various generations of AI tools ever since. LLMs are fun, but it is not like they are that much different from the other AI tools before them. Every time a new AI technology comes around I am finding a use case for it in my own flow. LLMs have their uses as well. But I am not trying to solve ALL the problems with the new tech.
I do not understand “the average person”. And I guess I never will.
Jomega@lemmy.world 3 weeks ago
Right, except they suck at all of those things. Especially the last one. Unless you think glue is an acceptable pizza topping.
1rre@discuss.tchncs.de 3 weeks ago
Nice, here’s a gold star for finding one case of it doing something wrong. I’ll call the CEO of AI and tell them to call it off, it’s a good thing humans have never said anything like that!
Jomega@lemmy.world 3 weeks ago
Bruh, you were the one that picked the examples. If you had a better argument you should have used that one instead.