progchan

So, i'm a firm believer in ai and llms. i legit think they can automate npc behavior and make the human condition less miserable with stupid pointless timesinks like "jobs" and "work"

But i swear talking to other ai engs has made me hate that even in a niche highly specialized field we get absolute fucking morons who don't understand the first fucking thing about any of these principles or architypes or algos and just think "WOW I CAN TELEGRAM MY AI HOW AMAZING"

When i go to a tech party, i know i'm going to get some students bootlicking me for advice and some indians looking to get me to whitewash their shit, but to hear over and over people talking about the most retarded concepts it's just torture.

Hmm even before ai education was in shambles. There's really no solution. There's definitely an upper limit to llms and I think once we hit it people will be forced to perform jobs based on that new plateau. Like how "calculator" was actually just a job before it was replaced with the tool. I think the general populous is just too misinformed by hype. You look at a field like accounting. All computers do is math and numbers and yet every fucking company has a guy that sits at a desk writing and checking the numbers on his screen with what he has been taught over 2-6 years of education that the numbers should actually be. That's all any modern tech job is. Once we hit the double entry accounting levels of understanding of ai by the general public I think people will be less bananas about the ai hype train. Maybe ai never plateaus, but there'll still have to be some guy at a desk cross checking his ones and zeros with sort of informed educated idea of what they should be.

Whats it like up there?

up where

i dont think AI will ever go away, per se. its obviously a disruptive technology and represents a paradigm shift.

the ultimate goal would be to use AI to advance the human race and to make life better, and automate the dangerous or stupid or mindnumbing shit away.

would be nice if the disruption caused by AI would help the cause of UBI....

unfortnately for the meantime, it will be used to try to replace humans so that corpos can save money on payroll.

but what AI will look like in 5-10 years will be very different from this silicon valley investor slop shit AI bubble.

i think AI will have its applications and killer apps in the future, but once the AI bubble pops and investor generated hype dies down, it wont be stuffed into everthing like they are tryna do now.

AI being so compute and therefore energy heavy will be its downfall, and i hope that dumbfuck trumpo's Iran war will be the first domino to fall to burst the AI bubble.

i currently wish the capitalists, investors, big tech morons, hardwre scapers and AI glazers a very die.

i forgot to include this,

does anyone regard AI engineers seriously? like do they think this market will last 5+ years?

thats like being a metaverse engineer, literally a meme-tier job based on a silicon valley fad.

is anyone going to take an AI engineer when a ToS or service update could compleately make your job impossible to do, irrelevant or obsolete?

sounds like bullshit

I think you're mostly correct. People are just switching from Google Search to AI Asking. The hilarious thing though is even though these models take tons of power to create the distilled models are pretty free and easy to run even locally. So ya if ChatGPT fails to overtake Google in market share its over for them.

Programming in the current year is insane. It's both way easier and way harder. You look at software like the huge fucking mess claude code (BY THE PEOPLE THAT ARE INVENTING VIBE CODING) and you have to wonder: what the fuck are programmers going to be like in a few years? Cars, airplanes, building engineering software, ext. If stuff that actually matters ends up being slop generated by the same type of people that did claude code then were all in trouble.

the new gemma models are fucking insane ngl. great leap forward.