How do you currently make use of generative AI?

[Puffy]

I work in digital so I see generative AI (like chatGPT, DeepSeek, etc) getting overhyped and talked about a lot.

My sense is that even within digital it's in experimental stages where there aren't really any universal standards or regulations around how it should be used - UK government is all over the place with regards to this - more individual pioneers exploring and experimenting with it to see what works and doesn't.

How are you currently making use of these tools in your personal and work life? What have you found to be the most compelling use cases for it for you so far?

 

[Puffy]

I'm studying thai massage at the moment so one thing I've been doing is ahead of doing a case study with someone is run any health conditions they have through the tool as a means of learning more about the condition and potential contraindications or things to keep in mind.

I'm not quite at a point of feeling I can rely on this as if it hallucinates and gets something wrong that has a risk of causing harm. So I tend to use it as a summary and then fact-check against a textbook for more detail.

 

[fluffy]

I am still in a design phase.

Cognitive processes I mostly understand.

(Executive functioning)

But I need more research on emotions.

All in all I don't really know what I could use a.i. for - it's been a long time since I thought about it?

People out there have much better designs than I do so it's like what's the purpose?

It just is that I need to buy a new computer first. The old one doesn't work in the way I would need. Something like a PS5

 

[Hadoblado]

I use AI for summarising readings and as a study-assist. It saves enormous time and deepens learning, but you need to know enough to be able to tell whether it's bullshitting you. I have it cite all references it uses and check that the sources are real, for instance (It's about 75/25 depending on which AI I use). When it's not hallucinating, it's the single most valuable learning tool I've ever used. I've been harassing Kuu for tips on which AI is good at what.

I also use it for formatting reference lists, for restructuring paragraphs, for brainstorming, for grading by rubric, and for orienting myself/sanity checking.

 

[dr froyd]

i've used it a bit for coding. It's good for getting a quick start in things you don't know well. For example i don't really know javascript, but managed to cobble together a UI in js using mostly just LLM prompts. You can learn interactively by asking it to clarify and explain things, etc.

for anything more complicated it's a bit of a mess tbh and rarely produces anything usable. For anything that requires mathematical reasoning it's practically useless.

none of this is surprising, of course. An LLM is good at that which has been done many times before and asked many times on e.g. Stack Overflow. It does the job you can expect from a statistically trained language model - nothing more, nothing less.

 

[Cognisant]

Cover letters, knowledge base documentation, writing polite emails when I'm not in a polite mood.

 

[birdsnestfern]

Maybe thru viewing art

 

[threeStepfourStep]

I do it for research purposes mainly, to find the right vocabulary on things so I can use those keywords on more established databases.

Most of the information openai bases itself on are really bad articles I find. Honestly I don't think it's possible to create an language ai model that can replace higher learning. It's mostly a more advanced Google for me.

Also the substance of what I research, like Asian politics or history/economics is really limited. If the source is not widely published it's not going to show up on the model. Not to mention that other languages need the input of those language users to make the model more proficient- so researching in other languages, when those languages are the ones I'm trying to find information on, are scant.

 

[birdsnestfern]

The Uncanny Valley feeling. I avoid it.

Uncanny valley - Wikipedia

en.wikipedia.org

Uncanny Valley: Why Realistic CGI and Robots Creep Us Out

The uncanny valley is a phenomenon that occurs when the more realistic a robot becomes, the more likely it is to seem eerie or even disgusting. Learn why it happens.

www.verywellmind.com

Ok, but generative AI is like microsoft co pilot? I tried to disable it in microsoft and ended up it just keeps getting automatically reinstalled. I do not like it or use it. I don't even speak outloud in case its listening. It makes me paranoid honestly, I don't want to become dependent on anything, I want my own mind to decide.

 

[birdsnestfern]

Instagram pays people for videos that get a lot of likes or views.
AI generates a lot of fake videos that make money, sometimes 5-10K!

More here: https://www.404media.co/inside-the-...ters-and-misery/?ref=daily-stories-newsletter