How many have found practical uses for large language text models in your real life?
My calendar application changed the method to add events and broke my workflow.
My workaround: I typed basic schedule info into Perplexity and have it convert the data to a CSV file and import it.
I tried using the one that spontaneously generates text adventures, but they kept turning porny.
That sounds like fun - do you recall the name?
AI Dungeon, specifically the GPC3 "Dragon" narrator
Yes, I use it to generate code that I run at home. Also generate a ton of icons that look amazing, and content for a game I am working on.
What do you use for the icons? Are they vector images?
Right, svg. Not doing it anymore but we had a use case for dynamically generating information bearing icons based on changing conditions.
These days I see there are dedicated services that do this really well.
Another thing I got some benefit at was using it to generate thousands of configuration json files covering all permutations of a wide set of values (within rules and ranges).
I use it as a way to rewrite emails to sound more professional, especially when I'm too lazy to word good, or I'm mad at the person I'm emailing. I can say what I really want and have AI tone it down, or smarten it up.
I use it a lot for helping structure written reports. I can't use exactly what it spits out, of course, but it helps me get an idea of what reads well and what doesn't. It's made me a better writer, I'm still bad, but getting better.
Helps with trouble shooting common or obscure IT issues in a pinch, not always right, but tends to point me the right direction. It's great at reminding me about steps I skipped over. Also helps explain underlying technical causes.
I use it to explain certain industries, sciences, and technical reasons behind specific technologies I don't have enough experience with, medical, biotech, IT, industrial, chemistry etc. Why would they use this vs that, what are reasons why they would or world not.
Jargon translator! If you want to learn a subject but you're eyes glaze over at all of the jargon, ask an AI. I feel like there isn't a topic I can't learn now. If I don't understand something I just copy and paste it in and say Explain. Anything I still don't get I can ask for more details or a comprehensive breakdown. There isn't a level of abstraction it can't get to. Works just as well in reverse like when I want to quickly turn my explanation into something someone else will understand. Cross department communication is much easier, same for explaining something up the chain. My favorite is to have it explain things to the execs in layman terms.
Using AI feels like having my own smart...thing .alien .gnome by my side, or an extra lobe in my brain. I feel like I'm making new connections and learning faster than I would have without it. I think a big part of that has been my initial double checking ALL of it's answers. That gave me a good feel for it's weaknesses and strengths, when to doubt it, double check, or know when it's just saying what sounds correct. I started using it day one every day, it's definitely improved but still has a ways to go.
Finally, simple scripts. Anything more than that and I end up wasting too much time debugging.
Bedtime stories. Kids list anything they want in the story, as silly as they can come up with, and in seconds we have a short story to giggle to.
There are probably more, but those are the ones that came to mind. I want to list the things it's absolutely terrible at too but another time.
Edit: typos
The only thing I use an LLM for is answering random questions I have throughout the day. "Why does my car windscreen freeze over when the air temperature doesn't get that low?" "How does a circumflex change pronunciation?" "What's a simple recipe for honey-mustard dressing?"
I think your last question answers itself.
Small one time use scripts, which aren't worth the time investment to do properly.
I mainly use LLMs as a programming assistant, as a brainstorming buddy and creative assistant while working on adventures for the tabletop roleplaying campaigns I run, and as a search result summarizer/interpreter over at Bing for the sorts of generic queries I used to use Google for.
Aside from small scripts and stupid joke images, I used ChatGPT on a vacation for extra details about the area I was in. It was when it first came out and Wikipedia and Google were obviously better for facts or finding restaurants but being able to ask follow-up questions was sometimes useful. You can be more vague and slowly zero in on what you’re actually asking about.
So, I don’t rely on it for anything important but it’s handy for a number of edge cases (sometimes, if only to figure out what you need to do a regular search for).
The main practical use I've found for me is writing a personalized cover letters when applying for jobs. Just plug in the job description, edit as needed, and off it goes. The way I see it, if HR can use it for their purposes, so can I.
HR doesn't do shit. Go ahead and disrespect them
i used it successfully for social formal mail, like wishing business contacts i hardly knew farewell. you just tell it to write a farewell mail for john paste it in your email and change stuff you don't like
...
For science texts I sometimes need to rephrase sections of text. A perfect task for our in house chatgpt. I'll give it a quick glance afterwards to see if the content is still the same but honestly it never let me down when I used it correctly.
Then also sometimes I try to look up things and I just ask chatgpt a question and to give me citations for their answer. That stuff only works in 1/3 cases which is a shame.
I use them to worsen my personal impact on climate change, since those things are consuming vast quantities of energy. It makes me feel more manly to know that I'm contributing to the slow destruction of our habitat every time I use them. It's like having a second SUV to go to the closest corner store multiple times a day. These changed my life for the better, made it easier, and I can't live without them now.
I have.
nice
😂
I use it a lot for code improvements. "I'm doing XYZ in my code. This is not efficient. Is there an algorithm to improve this?" Often times, there is.