AI and entering the market
Hey guys , I wanna raise a discussion about AI , lately some CEOS annouce to their employes that they muse use AI to levrage their work , im learning by my self to code in order to enter the market , ok so boiler plate code is done automated , but what about for example there is a new version or react or next js , the ai is not updating live , so it does help or not ? do i have any chance to enter the market and write my own code , test ,debug etc.. ?
31 Replies
AI is a tool that's finding its place in the industry, so it's hard to tell where it's going to end up fitting in. I very strongly doubt it's going to be the end of developers like some C-level idiots think, simply because 70% of a developer's job has nothing to do with writing code. And even that remaining 30%, an AI can maybe reliably do 5% of. and AI is plateauing in its capabilities
So should you learn to work with AI as a new developer? Probably, yeah. At least you need to know what it can and can't do.
Will AI change the industry? Sure, a little.
Will AI replace developers in the next decade+? No, definitely not.
as for the latest version of stuff: Most places won't use the latest version anyway
Thanks a lot , now as a someone who wants to get into the market , when ai can code simply projects , what should i code as junior for my portfolio ?
just make projects yourself
even time simplest ones ?
Build ones that fit your skill level
if I'm hiring a senior dev, I don't care about their first todo project unless it's very special. If I'm hiring a starting developer, a todo can be interesting to see
basically just build the hardest thing you can build well
yes but in a pov of the hiring manager , you could summon me to an interview and could ask , hey i can make todo or idk simon says in a click by generating from the ai , why did you make or... how can i know that you didnt do that by asking the ai for ?
you know what i mean now?
How can they know you didn't just copy it off someone else's portfolio? Or had a friend make it for you?
And, to be brutally honest... If all you have are apps you could vibe code, getting hired on your portfolio alone is going to be tough in this market
so what should i code and put on my portfolio these days ?
This
As a new one , should i use ai in terms of guidance ?
I wouldn't put any generated code in a portfolio project, but you could use Ai if you wanted to
yes yes im with you , only for guidance , not to generate code for me
no. never. ai is good at generating plausible-looking bullshit
if ai generates something that's complete garbage, your limited skills won't be able to know that, and you will spend too much time polishing a turd
and in the end, you will have to throw that away
i was talking only about guidance not code
okay, that doesnt change the fact that ai lies and is very good at it
unless you trully think that using white glue to prevent cheese from falling off of your pizza is a good idea...
haha
A bit late to the convo, but I like giving my opinion, so there lol
lately some CEOS annouce to their employes that they muse use AI to levrage their workThey are either stupid, gullible, or they're the CEO of a company that sells LLMs. Ignore them
ok so boiler plate code is done automatedThat was true even before AI. That's what boilerplate code is: common code that's copy/pasted (or
npm install
ed or the equivalent).
but what about for example there is a new version or react or next jsMost of the time you'll be working on a legacy codebase (one that existed before you started working there) so it'll be using an older version
do i have any chance to enter the market and write my own code , test ,debug etc.. ?Probably, but only time will tell. Sure, a lot of "influential" folks are saying it's the end of the junior dev. But how the hell do these CEOs et al think they're gonna get new senior-level devs? Do they grow on trees and sprout at senior level? No, they bloom from carefully cultivated junior dev seeds. Organically grown with the richest of soil and clearest of water…ok, the analogy is getting away from me. If these companies want more senior level devs they need younger junior level devs in order for them grow into senior level. Otherwise senior devs will become extinct. It's like what's happening to the US government (and the financial sector in general). All the super fuckin' old code written in COBOL isn't understood by any new devs because they prefer the shiny new shit. They think that
new === good
. But it's not always. COBOL has been running the entire world's financial ecosystem since the beginning. And if not COBOL then C or C++ (depending on the industry). These super important programming languages are being forgotten and it's going to cause issues in a few years when something breaks and there's no one around who knows how to fix it.the cobol point is really good
Anything, everything, nothing.
As Jochem said, put the hardest thing you've done in your portfolio. But note that not everything you make needs to be in your portfolio. You can make "silly" things to learn a new concept, programming pattern, API, or whatever. But you don't need to include it in your portfolio. Only include the things that took you a lot of time, effort, and work to do.
For example, don't include a codepen you did of a random fact generator using a free API to learn how
fetch
works. But do include the to do app that uses that fetch knowledge to dynamically update the status of your tasks across multiple devices.
As a new one , should i use ai in terms of guidance ?Honestly I'd say no, not yet. Not until you can tell when it's lying to you. Sure, you're not asking for code—which automatically makes you better than the YouTube "vibe coders"—but until you know enough to know what could be wrong you might learn some bad habits. Asking it for advice on what steps to take could work, especially if you read the sources it drew from so you can get a better understanding of the process. So much of being a programmer has nothing to do with code, but lines of code is one of the metrics that companies use to determine the "worth" of a programmer as it is measurable. But there's the planning, the brainstorming, the logic puzzling, and other non-tangibles that go into coding that is, sadly, under-appreciated. If you can, learn how to organize and plan on your own. Ask LLMs for help if you think you can tell when it's feeding you 💩 . But never fully rely on it, use it as a starting point and inject your own brainpower into it. The important bit is this: Always think for yourself. Don't let LLMs "think" for you, or you'll never have an original thought again
Do you guys know or familiar with a game called DLE ?
-# …is this question related to the initial question? If not, should be asked elsewhere (either in a new post if a Question or in #gaming-general if it's just to discuss the game)
yes yes it is
Ok, then best to ask the actual question. Don't ask to ask 😉
there is a game called dle for some things , like i know for some animes , like guess the characther , guess the finishing move and its daily , so im a huge wrestling fan and i wanna create something like that but i feel like its really look overwhelming AF
This? https://www.gamedle.wtf/
Narutodle
Daily guess Naruto and Naruto Shippûden characters!
for example
Ok, so it's a guessing game. Not sure what that has to do with AI in the workplace
no , we were also talking about projects , that's it 🙂
It looks overwhelming, sure. But break it down into bite-sized pieces. What do you need for a guessing game?
* Questions and answers (so there's a database)
* a UI (that's almost always a given these days, but a terminal app is a popular first project for some languages)
* A way for the UI and DB to communicate (webapps use HTTP almost exclusively, but WS or SSE are other options)
Now that you know the three main parts, break them into smaller bits. For the DB, either you make the questions/answers yourself or you find a list on the internet. For the UI, make it simple. A place to display the question along with places for the potential answers to be displayed. Probably a button to click on. As for the communication, that's a simple HTTP call (using
fetch
if you're using JS).
It's all about taking a big task and breaking it down into smaller segments. Continue until you have simple, actionable steps
Then do it🙂