Asklemmy

Yesterday I was working on a training PowerPoint and it occurred to me that I should probably simplify the language. Had GPT convert it to 3rd-grade language, and it worked pretty well. Not perfect, but it helped.

I'm also writing an app as a hobby and, although GPT goes batshit crazy from time to time, overall it has done most of the coding grunt-work pretty well.

I use GPT-4 daily. I worked with it to create a quick and convenient app on my smartwatch, which allows it to provide wisdom and guidance fast whenever I need it. For more grandular things, I use its BingChat interface which can search the web and see images. The AI has helped me with understanding how to complete tasks, providing counseling for me, finding bugs in my code, writing functions, teaching me how to use software like Excel and Outlook, and giving me random information about various curiosities that pop into mind.

I don't keep it a secret and tell anyone who asks. Plus it's kinda obvious that something is going on with me. I always wear bone conducting headsets that allow the AI to whisper in my ear without shutting me out to the world, and sometimes talk to my watch

The responses to knowing what I'm doing have almost always been extreme: very positive or very negative. The machine is controversial, and when some can no longer stay in comfortable denial of its efficacy they turn to speaking out against its use

Edit: just fixed its translation method. Now the watch will hear non-english speech and automatically translate it for me too (uses Whisper API)

I've found ChatGPT is good for small tasks that require me to code in languages I don't use often and don't know well. My prime examples are writing CMakeLists.txt files, and generating regex patterns.

Also, if I want to write a quick little bash script to do something, it's much better at remembering syntax and string handling tricks than me.

I know many people my slightly younger than me are using chatgpt to breeze though university assignments. Apparently there's one website that uses gpt that even draws diagrams for you, so you don't have to make 500 UML and class diagrams that take forever to create.

Trying to do our work with ChatGPT and then discussing the results has been our #1 topic of kitchen conversation all year.

I use ChatGPT fairly frequently. For example, I often have to write a business email. I'm usually pretty good at it. But sometimes I don't have the time or desire to find the right wording. This is where ChatGPT comes into play: I have trained my writing style using several examples and then simply have the quickly written emails beautified.

My boss doesn't know about it, but I don't hide it either. My company is very, very slow on the technical side and will only understand the benefits of AI in a few years.

My job actively encourages using AI to be more efficient and rewards curiosity/creative approaches. I'm in IT management.

Coworker of mine admitted to using this for writing treatment plans. Super unethical and unrepentant about it. Why? Treatment plans are individual, and contain PII. I used it for research a few times and it returned sources that are considered bunk at best and hated within the community for their history. So I just went back to my journal aggregation.

We openly use it and abuse of it from top to bottom of the company and for me add Co-Pilot to that as well

English is not my first language. I use it to fix grammar and rephrase sentences for making communication easy.

The platform/language that I use doesn't supported by chat GPT or Bard. So I write my own code.

As a backend developer I use it to explain some SQL, dev processes that I should know but unsure on, or best practices for X.

SQL is my weakest link.

Not using chatgpt at all because it's queue is always full.

I use it as a search engine for the LLVM docs.

Works so much better than doxygen.

But it's no secret.

I am the boss and I've had to cajole a couple of my employees into using it.

Any employer that thinks using ChatGPT carefully and judiciously is a bad thing is mistaken. When it works it's a great productivity boost, but you have to know when to kick it to the curb when it starts hallucinating.

I've run emails through it to check tone since I'm hilariously bad at reading tone through text, but I'm pretty limited in how I can make use of that. There's info I deal with that is sensitive and/or proprietary that I can't paste into just any text box without potential severe repercussions.

I use it at work but gladly tell the boss... It's only pluses if we can do more trivial work faster. More time to relax. They don't watch what I do during the day. The boss relaxes also. All good.

I use it to speed up writing scripts on occasion, while attempting to abstract out any possibly confidential data.

I'm still fairly sure it's not allowed, however. But considering it would be easy to trace API calls and I haven't been approached yet, I'm assuming no one really cares.

As a coder, we have had discussions about using it at work. Everyone's fine with it for generation of test data, or for generating initial code skeletons but it still requires us to review every line. It saves a bit of time but that's all.

Im using the shit out of gpt-4 for coding and it works. And no never told anyone cause nobody asks.

Technically not ChatGPT, but GPT4 directly via the API. And I also told my boss about it ;) I used it as baseline for a few translations, had it generate some boilerplate, and create classes in multiple languages from some JSON. But most of the time I have no need it can fulfill, and most of the (tiny) bill OpenAI sends me is from me using it for personal stuff like getting an idea what to do for our wedding anniversary (the boat ride was a great idea :D), or as a funny bot in two twitch chats I moderate (I use 3.5 there because it’s over an order of magnitude cheaper).

Aside from asking it coding questions (which are generally a helpful pointer in the right direction), i also ask it alot of questions like “Turn these values into an array” or something similar when i have to make an array of values (or anything else that’s repetitive) and am too lazy to do it myself. Just a slight speedup in work.

I use it

My boss likes it too. Of course we dont trust it m, but it can do certain things easier and faster than a human can

My boss pays for it! I don't use it that much, but it's pretty useful from time to time instead of going through a bunch of unrelated Google results.

I tried it once or twice and it worked well. It's too stupid now to be worth the attempt. The amount of time spent fixing its mistakes has resulted in net zero time savings.