729
Junior Prompt Engineering
(lemmy.ml)
Tech guy invents the concept of giving instructions
With clear requirements and outcome expected
Why did no one think of this before
Who does that? What if they do everything right and it doesn't work and then it turns out it's my fault?
It would be nice if it was possible to describe perfectly what a program is supposed to do.
Someone should invent some kind of database of syntax, like a... code
But it would need to be reliable with a syntax, like some kind of grammar.
That's great, but then how do we know that the grammar matches what we want to do - with some sort of test?
How to we know what to test? Maybe with some kind of specification?
People could give things a name and write down what type of thing it is.
A codegrammar?
We don't want anything amateur. It has to be a professional codegrammar.
What, like some kind of design requirements?
Heresy!
Design requirements are too ambiguous.
Design requirements are what it should do, not how it does it.
That's why you must negotiate or clarify what is being asked. Once it has been accepted, it is not ambiguous anymore as long as you respect it.
I'm a systems analyst, or in agile terminology "a designer" as I'm responsible for "design artifacts"
Our designs are usually unambiguous
What did you said?
I think our man meant in terms of real-world situations
And NOT yet another front page written in ReactJS.
Oh, well, that's good, because I have a ton of people who work with Angular and not React.
This still isn't specific enough to specify exactly what the computer will do. There are an infinite number of python programs that could print Hello World in the terminal.
I knew it, i should've asked for assembly
Yeah but that's a lot of writing. Much less effort to get the plagiarism machine to write it instead.
Ha
None of us would have jobs
I think the joke is that that is literally what coding, is.
OP just chatting with themselves so they can screenshot it?
That is some telegram group and both messages shows from left with profile icons(which got cropped). The screenshot person sent the last message which shows double ticks
In the desktop client the positions of bubbles also depend on the width of the window.
Great attention to detail!
That's just a fake conversation in general, look at the timestamps between the messages from the interlocutor. Several minutes to type a complete sentence?
Hey, i can take a few hours to reply sometimes :c
Could be a group chat but we all know they're a twat
I wrote a shell script like this (it admin , notna dev) for private use.
The prompt took me like 5 hours of rewriting the instructions.
Don't even know yet if it works (lol)
Calling GPT a neural network is pretty generous. It's more like a markov chain
it legitimately is a neutral network, I'm not sure what you're trying to say here. https://en.wikipedia.org/wiki/Generative_pre-trained_transformer
You're right, my bad.
The core language model isn't a nueral network? I agree that the full application is more Markov chainy but I had no idea the LLM wasn't.
Now I'm wondering if there are any models that are actual neutral networks
I'm not an expert. I'd just expect a neural network to follow the core principle of self-improvement. GPT is fundamentally unable to do this. The way it "learns" is closer to the same tech behind predictive text in your phone.
It's the reason why it can't understand why telling you to put glue on pizza is a bad idea.
the main thing is that the system end-users interact with is static. it's a snapshot of all the weights of the "neurons" at a particular point in the training process. you can keep training from that snapshot for every conversation, but nobody does that live because the result wouldn't be useful. it needs to be cleaned up first. so it learns nothing from you, but it could.
"Improvement" is an open ended term. Would having longer or shorter toes be beneficial? Depends on the evolutionary environment.
ChatGPT does have a feedback loop. Every prompt you give it affects its internal state. That's why it won't give you the same response next time you give the same prompt. Will it be better or worse? Depends on what you want.
I've played with markov chains. They don't create serious results, ever. ChatGPT is right just often enough for people to think it's right all the time.
Neural network: for when saying LLM doesn't sound smart enough
It's just what it was called in the nineties.
this post was submitted on 01 Apr 2025
729 points (100.0% liked)
Programmer Humor
23345 readers
352 users here now
Welcome to Programmer Humor!
This is a place where you can post jokes, memes, humor, etc. related to programming!
For sharing awful code theres also Programming Horror.
Rules
- Keep content in english
- No advertisements
- Posts must be related to programming or programmer topics
founded 2 years ago
MODERATORS