this post was submitted on 04 Aug 2025
31 points (61.3% liked)

Ask Lemmy

33863 readers
834 users here now

A Fediverse community for open-ended, thought provoking questions


Rules: (interactive)


1) Be nice and; have funDoxxing, trolling, sealioning, racism, and toxicity are not welcomed in AskLemmy. Remember what your mother said: if you can't say something nice, don't say anything at all. In addition, the site-wide Lemmy.world terms of service also apply here. Please familiarize yourself with them


2) All posts must end with a '?'This is sort of like Jeopardy. Please phrase all post titles in the form of a proper question ending with ?


3) No spamPlease do not flood the community with nonsense. Actual suspected spammers will be banned on site. No astroturfing.


4) NSFW is okay, within reasonJust remember to tag posts with either a content warning or a [NSFW] tag. Overtly sexual posts are not allowed, please direct them to either !asklemmyafterdark@lemmy.world or !asklemmynsfw@lemmynsfw.com. NSFW comments should be restricted to posts tagged [NSFW].


5) This is not a support community.
It is not a place for 'how do I?', type questions. If you have any questions regarding the site itself or would like to report a community, please direct them to Lemmy.world Support or email info@lemmy.world. For other questions check our partnered communities list, or use the search function.


6) No US Politics.
Please don't post about current US Politics. If you need to do this, try !politicaldiscussion@lemmy.world or !askusa@discuss.online


Reminder: The terms of service apply here too.

Partnered Communities:

Tech Support

No Stupid Questions

You Should Know

Reddit

Jokes

Ask Ouija


Logo design credit goes to: tubbadu


founded 2 years ago
MODERATORS
 

I know the reputation that AI has on Lemmy, however I've found that some users (like myself) have found that LLMs can be useful tools.

What are fellow AI users using these tools for? Furthermore, what models are you using that find the most useful?

(page 2) 48 comments
sorted by: hot top controversial new old
[–] Libb@piefed.social 3 points 4 days ago* (last edited 4 days ago)

I know the reputation that AI has on Lemmy, however I've found that some users (like myself) have found that LLMs can be useful tools

I know the reputation that AI has on Lemmy, however I've found that some users (like myself) have found that LLMs can be useful tools.

Their utility is not questioned. It's their true cost and how they're developed that's the issue.

No doubt a machine able to do some quick and dirty jobs that would take us a lot more time is a fine tool (like mentioned already, denoise, quick text summaries and stuff like that) edit: even complex and highly skilled stuff. The tool is already impressive today, and I don't doubt it will get much better quickly.

The issue is how it learned to do what it can do and how it is monoetized. I mean, learning from humanity common knowledge (no AI at all without it being allowed to learn from us all) and making it... subscription-based for us to use? WTF? The issue is also how it is destroying many things in the exclusive profit of a handful of very rich people and their shareholders. The issue is how we, mankind, have zero control over a tool that is threatening to make a lot of us go bankrupt...

Feel free to downvote, obviously.

And to answer your question:

What AI tools have you found useful?

I would say, the off button... of which there is none I can find.

[–] gigachad@piefed.social 3 points 4 days ago

I use predictive AI for certain classification tasks daily at work, however I call that Deep Learning and not AI. I don't want to be too specific, but you can imagine we are classifying certain objects - is this a traffic light, is this a tree etc. It is a task that cannot be solved geometrically very good, so Deep Learning is the perfect use case there.

[–] Nusm@piefed.zip 3 points 4 days ago (2 children)

I know people dislike and complain about it, but I absolutely love Suno. LOVE IT. I’ve created what I think are some really cool songs. Will they ever be hits on the radio? Nope. Will anyone else listen to them besides me? Probably not. But boy, after tweaking, I’d rather listen to some of the songs I’ve created than the garbage on the radio!

[–] FaceDeer@fedia.io 2 points 4 days ago (1 children)

There's also Udio.com and Producer.ai out there, and possibly some others - music generation is becoming fairly widespread. I didn't mention any of this in my list of recommendations though because OP specifically asked for LLMs. :)

[–] Jordan117@lemmy.world 2 points 4 days ago (1 children)

I don't know how Suno has become so much more popular than Udio. Every Suno track I've heard has sounded like the same generic pop, and the vocals always have this noticeable "synthy" quality.

[–] jcg@halubilo.social 1 points 4 days ago (1 children)

Have you heard the stuff from the new v4 model? The vocals are so much clearer and the instrumentation gets pretty varied (ymmv depending on how specific you get with the styles though)

[–] Jordan117@lemmy.world 1 points 4 days ago

I have -- I had high hopes for it, but the vocals are still pretty bad. They all have this metallic, tinny quality, liked they're subtly vocoded or being sung into those little toy spring microphones for kids. It's a constant reminder that it's artificial and just completely takes me out if it.

[–] Varyk@sh.itjust.works 1 points 4 days ago

Holy crow, that freaked me out. That's really impressive. Pretty uncanny valley, but I can definitely see the appeal.

[–] Keyboard@lemmy.world 3 points 4 days ago (1 children)
[–] tpihkal@lemmy.world 2 points 4 days ago (3 children)

I'm going to need you to elaborate on this one...

load more comments (3 replies)
[–] piefood@feddit.online 2 points 4 days ago

I've been self-hosting my own AI stuff for a bit using Ollama. I use it to create images, design a tattoo, run a chatbot, write emails, write code and commit-messages, run a D&D game, explain concepts that I'm not familiar with, translate languages, etc.

I've been toying with different models, and I'm not sure that I have one that I would say is a goto. I am liking Ollama to be able to easily pull in and test new LLMs, as well as Stable-Diffusion for image generation/modification.

[–] HubertManne@piefed.social 1 points 4 days ago (1 children)

If we are talking chatbots I see them as another level of abstraction to search and is useful but I have concerns on the energy use. Other uses I have encountered is just sorta a convenience thing. Where it can do a bunch of things that individual software can do but at a one stop shop. I have not directly been involved in other aspects but im aware how they are baked into things like facial recognition and tracking and such.

[–] PeriodicallyPedantic@lemmy.ca 1 points 3 days ago (1 children)

Dont use AI chat as a replacement for search except on popular subjects with broad consensus. Which unfortunately is when you generally don't need it.

[–] HubertManne@piefed.social 1 points 3 days ago (1 children)

its fine as long as it gives references to check out. I mean its not fine because of the energy usage but if that is solved I would use it for search. again as long as it tells me sources.

[–] PeriodicallyPedantic@lemmy.ca 1 points 3 days ago (1 children)

If what you want is sources, a regular Google search will do a better job if it's a popular subject.

Chatgpt and it's kin will be inexplicably creative in its choice of sources, and in its summary thereof. And in the sources themselves sometimes.

If it's something you care to get right, just skip AI.
If it's meaningless, then it's harmless in its potential inaccuracy

[–] HubertManne@piefed.social 1 points 3 days ago (1 children)

See just like a normal search its up to you to evaluate it. ai search wise is as I said another abstraction. Not using it is like turning off the little snipets search engines do nowadays and going back to just clicking an each and every link. The problem is people just taking the response as gospel with no critical thought.

[–] PeriodicallyPedantic@lemmy.ca 1 points 3 days ago (5 children)

Like a normal search, except it only provides like 4 links, it's choice of links is even worse than Google SEO, and it provides inaccurate summaries of them rather than relevant text snippets.

So yes, people just taking the response as gospel is bad. But also it's just worse than search if you're using it as a search.

load more comments (5 replies)
[–] cheese_greater@lemmy.world 2 points 4 days ago (3 children)

I like it for coming up with quick, modular code that produces whatever direct result I want without having to reivent the wheel provided I more or less understand how it works and how to tweak it to refine what I want or how it does it

load more comments (3 replies)
[–] burrito@sh.itjust.works 2 points 4 days ago (2 children)

I'm running ollama and open-webui and some unsloth modified models for some general purpose stuff.

The https://huggingface.co/unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF model has been pretty good. Beware it's a Chinese model so you can get some funny results if you ask about Tiananmen Square or if certain people resemble Winnie the Pooh. For making Linux configurations it works great.

Some gemma3 models are okay but it doesn't seem as good. Same for Phi4 models.

[–] occultist8128@infosec.pub 1 points 4 days ago (1 children)

What are the minimum requirements?

[–] burrito@sh.itjust.works 2 points 4 days ago

It can run with a variety of systems. You just need to have enough VRAM on your video card to fit the model and then it can run pretty fast. There are models down to a couple hundred MB in size, but they're quite limited. There are other models that are 245GB in size, though the bigger ones use a "mixture of experts" where only portions of the model are loaded as needed, and the rest stays unused for the particular task at hand. If you don't have enough VRAM to fit the model, it will fall back to running on the CPU and using the system ram. Most of the operations are limited by the speed of the memory that's running the model. Video card memory is much faster than system memory so that's what helps it run a lot faster. It can still get the job done but you will have to wait quite a while for the output. There are ways of making the models smaller by using quantization. Quantization reduces the precision of the models parameters (the number with the b next to it in models i.e. 4b, 8b, 14b, 30b, etc.) by taking it from 32-bit data down to 8-bit or smaller. This allows more data to be packed in a smaller space, but it reduces accuracy a bit.

[–] FaceDeer@fedia.io 1 points 4 days ago

I've currently got a fancy Python script using that model to scan through 15 years worth of transcripts of audio recordings figuring out what the context and subject matter of the recordings are, so that I can sort and search them. Previously they were just a giant pile of audio files.

I used WhisperX to transcribe them. The hardest part was getting pip to install WhisperX correctly, everything else has been just routine Python coding.

[–] Appoxo@lemmy.dbzer0.com 1 points 4 days ago

For some situations I used Copilot to script an auto-translator for XML-EPG in bash.

For that it worked okay enough.

[–] Mr_Fish@lemmy.world 1 points 4 days ago

The only time I've seen real use from an ai tool is at work, we are using it to get data from an invoice/quote/etc from the pdf that we get emailed to data that we can put in the database. It's not a perfect solution, but there isn't really anything else we can find other than getting people to do it, which is slower and more expensive.

[–] chaosCruiser@futurology.today 1 points 4 days ago* (last edited 4 days ago) (1 children)

I think I’ve found the one area where LLMs really excel: business books / self help literature. The real life examples in that genre are pretty awful and dragged out as it is, so you can’t really make it much worse, now can you? The information density is kept low to fluff up the page count, and oh boy, are LLMs good at that. So, if you want to become a self help guru, but can’t be bothered to write your own book about magical hotels, marriage advice, productivity tips and communication, LLMs can take care of that for you. Copilot has turned out to work well for projects like that.

If you raise the bar, you’re going to have to read and edit the text manually. You also need to keep track of what has already been mentioned elsewhere and avoid repeating them again, depending on the genre. In business books though, that’s not a problem at all.

BTW, if you wonder about the downvotes, it’s because Asklemmy@lemmy.world isn’t a safe space for AI related discussions. Consider posting somewhere else.

[–] goldenbug@fedia.io 3 points 4 days ago (1 children)

A safe space? Sorry but disagreeing with you is different from actual hatred campaings

[–] chaosCruiser@futurology.today 1 points 4 days ago* (last edited 4 days ago) (1 children)

The main point of the post is to ask a question. Apparently that is something people disagree with. Maybe they don’t like what the question implies.

[–] goldenbug@fedia.io 2 points 4 days ago (1 children)

You say in your post 'BTW, if you wonder about the downvotes, it’s because Asklemmy@lemmy.world isn’t a safe space for AI related discussions. '

A safe space is where we can debate, disagree or downvote each other. If you receive death threats, harrassment, stalking and so on, this is definitely not a safe space.

People disliking Generative stuff have their reasons for it.

load more comments (1 replies)
[–] PixelatedSaturn@lemmy.world 1 points 4 days ago

I was testing a lot of tools for work (UX), but they all sucked. Basically I just use gpt, cursor for tab coding,l and image generation for various purposes (Leonardo).

Most apps work poorly and using a model directly works better.

I worked on a few apps that use ai without a chat bot, but we found users expected one, maybe that will change in the future.

[–] d00phy@lemmy.world 1 points 4 days ago

I’ve enjoyed messing with Perplexity and Duck AI.

[–] lukaro@lemmy.zip -1 points 4 days ago

I chat with llm's to help me remember shit I already know, or to change the tone of stuff I write but thats about it.

[–] Tracaine@lemmy.world -1 points 4 days ago (6 children)

I mostly use it to generate smut for gooning. It's useful for that. Chat gpt is absolutely filthy if you get it going the right way.

[–] FeelzGoodMan420@eviltoast.org 2 points 4 days ago (1 children)

Do you enjoy providing data companies with your personal porn preferences?

[–] tpihkal@lemmy.world 1 points 3 days ago

To browse Pornhub with cookies on or with cookies off? That is the question 💀.

load more comments (5 replies)
load more comments
view more: ‹ prev next ›