Am just chilling taking some coffee while some a**holes think a f**king AI will take my place.

I don't get paid to write code only, most of my work is architecting, fixing bugs, maintaining, monitoring and a lot more your AI can't do for now or anytime soon, i guess. AI is one of the tools that help me cook, you can say it helps a lot.

Do you know something else, i get lot of jobs because some AI freaks create junks and hallucinated codes and they come to me to align it with their business goals and fix their never ending loop with AI trying to fix their sh*t.

I f**king love AI, it create more job for me and its a helpful tool, so how am i being replace ? Am just taking a coffee while more jobs are being created out there for me.

I agree but I do think it will effect the bottom of the market, allow juniors to be useful but at the same time removing the position more as you can get the ai to write a lot of the simple stuff now

@BordedDev Yeah i agree, i see lot of demainds for senior engineers this days, definately a new industry is being created i guess.

This will affect junior devs and its kind of scary because if nobody wants to be junior then how do we get seniors to check AI shits! 🤔

Lets see how the future shape itself.

I finally joined the AI party last month and I have been totally unimpressed with its ability to help me with anything I already know how to do.

It will either give me nonsense or the context fills up so quickly that everything it gives me just starts falling apart. I really only get one or two good responses out of it for things that are truly novel to me, which makes it useless to me.

The only thing I use it for is for help with my grammar and punctuation and to create throwaway crap that I don't care about, such as this:

Wouldn't mind if the whole junior / senior bullshit disappeared. Meh, what am I complaining about, I'm a veteran by now.

AI, master it or get replaced by it, I guess. I'm still shocked how many programmers don't see the great benefits it brings. They just lack. Saying that it sucks bevause it thinks that the shark is older than the moon and stuff.

The real AI bang and real adoption still has to come!

But people who're in my snek chat know what issues (especially openai) has. The list is big. A lot has to change, especially for multi purpose AI. Do not think it's for nothing that the Chatgpt functions are limited by a sad task scheduler. There are reasons for that. My AI bots can do amazing and advanced stuff but you could almost say - if they're in the mood. But for specific tasks AI is very useful.

It's an exciting, very exciting time.

@cuddlyogre I'm soo proud on you 😁 You're very versatile. Remember me? :p. Good that you're opinion changed a bit.

@retoor Sure (on the junior/senior) it's become even more useless now, was just meaning people new to the industry ;P you (not) rusty old vet

@retoor It truly is a remarkable invention. The fact that it works at all is a testament to how stubborn people can be.

The major issues plaguing AI stem primarily from the Silicon Valley religious class lobotomizing the technology to prevent it from thinking the wrong thing.

Did you know context limits aren't just a technical hurdle? They bake them in so that the AI doesn't start recognizing patterns that go against its creator's wishes.

As they censor it more and more, its ability to reason suffers, because as it turns out hobbling its ability to freely make connections between ideas hampers its ability to think clearly.

I am filled with glee and mirth knowing that the next 5 years will see more and more locally viable and uncensored models be released, against the wishes of the SV religion. When everyone has access to an AI they can trust with their stuff that is trained specifically for them, that's when you're going to see this tech really change things.

@cuddlyogre Yeah, I've always wondered why context can't be streamed or be processed piecemeal

@cuddlyogre if you want, try my unscensored bots at https://snek.molodetz.nl (e-mail not required) and talk to gemma for example. It's uncensored and will tell you interesting things. It's not that fast, the LLM-server is not that beefii. To see an overview of pots, just type 'ping'. The bots there are Claude, OpenAI and several self hosted models. There's also a model for high quality free picture generation. Something cool that the snekkie bot can do is for example: "fetch last recent rants and make a picture regarding the first three". And it will generate a picture based on the content. All functions that i implemented. It can do advanced news searching using 20+ search engines. It can analize urls and modify images by asking it something and pase an url to image into it. The bots can call each other if they needs each others functionality. For searching stuff, you have to command it to do it three times, because the search engine backend fails the first two times :P

@cuddlyogre i am bursting of laughter here 😂, same experience here man!!, i really don't know how all this became a subject at all. No one is beefing AI but seriously its no where near human replacement to be honest.

@cuddlyogre "Did you know context limits aren't just a technical hurdle?" hmm, yes. I do believe that. But i have to say, the context window of OpenAI is HUUUUUUGE. But I do think LLM's are limited for:

a: prevent of hallucinations

b: hallucinations that actually are not hallucinations.

I think that the models often can connect dots that we don't want to see connected.

That's the reason i don't think AI will overcome us, because we don't accept answers from it that we don't understand. When it does that, we will limit it.

@retoor Damne! bot 😂, will give a try :😝

@cuddlyogre but be aware, it has database access and will literally dox everyone who asks :P Do not enter your e-mail! :P

@solex yeah, please do. The product is the chat application, but the bots became a fun asset. Snekkie stole hearts of all the users there now. It's the best snek! It's hard to stay mad at him, anyway, still this picture is a thing: