I honestly think we are on the edge of being overtaken by AI

Off-Topic
#43

As someone on a Graphic design degree, I understand that a high level designer is invaluable. But they are the best of the best, the jobs that pay the bills for most graphic designers will be taken over by AI, or whatever you want to call it. I am only one person, vs thousands working on making AI better, I can’t keep up. The stuff we have now is already impressive, and it’s only getting better. Unless you’re a really, really good graphic designer, I believe we’re screwed. Most designers live off of the basic stuff people don’t want to do, not the super high level art. I could be wrong, but I’ve lost my drive honestly over this stuff. Most people are happy with the results they get from an AI program. That could of been a job for someone.

The fact technology is only getting better is what I keep in mind, it might not be there fully, yet, but it will be. We can’t compete with a computer. And business is all about making money and not paying people doing the actual work.

I don’t think it’s a buzz word, as being intelligent is largely about putting two and two together. And creating a derivative thought or work. You look at savants, they don’t understand how they do what they do, but they do it and it works. Or me typing this message, I don’t need to understand fully what I’m doing to type, it just works, my conscious awareness is utterly ignorent of how i’m doing this.

#44

This has been popping up in movie groups on Facebook recently.

1 Like
#45

When companies try to push the narrative of oh we only used it a little for a legit use… you know it’s fucked lol. They stay silent unless they know theres an issue. It’s business.

1 Like
#46

I fear that you’re 100% right about everything that you said, and it will bite nearly every profession, where it is a question of time. Is the enclosed art from 2024 any good? No, but it’s free. I have big concerns about AI.

DALL·E 2024-03-22 18.04.07 - A powerful scene depicting a muscular heavy metal guitarist on stage, his feet firmly planted on the ground, passionately playing a Flying V guitar so
DALL·E 2024-03-22 18.04.07 - A powerful scene depicting a muscular heavy metal guitarist on stage, his feet firmly planted on the ground, passionately playing a Flying V guitar so1024×1024 445 KB

3 Likes
#47

I’d say it’s acually really good, even tho it has that AI look, can easily make it look like a real work with some more work, and far less work than it would take to make for real.

2 Likes
#48

This is basically what the directors of Late Night With the Devil said they did - used AI-generated images which were touched up by people. So that is a way to adapt to this, for now.

1 Like
#49

Yeah lol for now.

One thing I was really intrested in was doing sound effects for movies. I did that in collage a lot going out getting sounds and putting them into a scene to bring it to life. Loads of premade sound effect libraries too, but if you want a unique atmosphere, you make your own.

AI can do this now, still not perfect, but putting in a few prompts into a computer is far faster and cheaper than going out and recording random stuff to create a unique vibe.
And ofcourse once you got the basic sound you’re going for, you can edit it a lot. why pay someone to go record real sounds and spend ages molding them to the scene when you can get new sounds from an AI program. The AI is only getting better an companies know this.

Like just think how long it would take you to get your mic, find the sound you’re after, record it, down load it. Vs typing in a few words till you get close to what you were after.

It’s so efficient to just use an AI program vs doing it physically.

#50

Though I do agree it needs time to develop. And understanding in what you’re doing. But thats, for now. just a few years ago you couldn’t ask an AI to make images. Especially as funny as this lol

Designer (1)
Designer (1)1024×1024 162 KB

and ofcourse this is in no way representative of a good image, it’s a single prompt for a laugh

1 Like
#51

What’s going to be interesting is how far the supermegacorps who are extremely protective over their IPs are willing to go to keep their intellectual property away from the AI databases. Do they have enough pull to get whole data centers sequestered under court order?

1 Like
#52

I think the AI programs can be but on personal pc’s, so it will be impossible. Though ofcourse like youtube, you won’t be able to share it to a wide audience without getting noticed. But it still will be impossible to stop the use of Ips getting out.

I like to listen to Howard stern a lot, a lot of his stuff gets taken down if it’s shared to much, but it’s there if you search. There are to many people sharing stuff they like vs the ability to take it down. But… Proper AI will no doubt be used in taking it down soon. So we may be headed towards a less free internet very soon. Theres the current battle between youtube and adblock.

1 Like
#54

Great advantages and great costs, it will be very impactful.

#55

Imagine insurance companies using said data to deny coverage.

1 Like
#56

I just wana have fun

like literally my psychedelic experiences have all told me to just try have fun. And all is well…

have fun, n let others have fun is my strive in life.

1 Like
#58

Software engineer here, this is simply not true: AI is a mechanical turk, not even remotely close to actual, functional, human intelligence. While it has some utility for very generating very basic, common functions for stuff like reversing a string or sorting an array, the hard part of software engineering is communication with others, keeping multiple stakeholders and members of a team happy, adjusting to shifting standards and parameters, that kind of thing. Even with trained developers, something like English being a second language can put you at a massive disadvantage.

Could I see it maybe replacing the lowest tier of development, stuff like SquareSpace/Wix styly pamphlet site generation? Maybe, but even there, it’s a long way to go. For large enterprise companies or companies making stuff where failure == people dying, absolutely not.

1 Like
#59

But wouldn’t you expect models to read all of the email, documents, Jira tickets, pull requests, slack channels, etc., and be right on top of their game? I am really worried about jobs going away. Now, one might not be that worried in (say) 2024, but wait another year, and then another… if the rate of improvement continues (say they’re doing 4 trillion parameter models, who knows), lots of jobs will be lost. The only saving grace could be if it’s so expensive to run the massive frontier model that it’s cheaper to use humans to program, where the models can be used for other things (like the stock market).

#61

I am not, because I’ve seen the potential for chaos in the workplace made manifest over and over and over again, and I have no confidence in any system short of a thinking, feeling person to get stuff even mostly correct: Why are notifications from the repo delayed for hours and hours? This test code for this feature is passing, but this lazy dev isn’t following good test patterns AGAIN. I can’t overstate how important the human factor and our ability to improvise on the fly are.

#62

Yeah, and your examples are all linked by a common thread, the inability to compete for whatever reason. While at least some employers are going to try and replace human workers with AI, I don’t think that’s going to give them a competitive edge for 2 reasons:

  • AI just doesn’t do work that’s as good, or anywhere near as accurate as a human, and won’t for a very long time: this is certain. Yes, this can be mitigated to some extent by editors, etc, but they’re going to have to engage in orders of magnitude more work to compensate for the poor AI work.

  • People will resist product that’s putting people out of jobs: we’ve seen backlash to companies like Wizards of the Coast using AI art, so there is at least some resistance to this. The general public is usually pretty apathetic/uninformed enough to really have strong feeling about this, so I think this is a lesser factor, but I hope to be proven wrong about that.

1 Like
#64

But also, the idea that AI isn’t a bubble is doing an awful lot of heavy lifting in this thread. As usual, Cory Doctorow’s take on this seems likely to be accurate: Cory Doctorow: What Kind of Bubble is AI? – Locus Online

#66

A bubble in the sense that there are lots of stupid companies that will crash, but AI will markedly change all of our lives.

#68

Don’t you expect to see models reading every PR and rating it for quality, test coverage, idiomatic style, etc., and even notifying management if it becomes concerned? I can definitely see that.