The AI Discussion Thread

[Clive]

maybe the data centers can be made near nuclear plants and connected to nuclear

That’s… that’s not how that works…

 

[Andysol]

maybe the data centers can be made near nuclear plants and connected to nuclear

You’ve solved the power issue!

…

While worsening the water issue.

 

[fryoj]

Just need to get AI to solve the desalinization and solar problem and boom. The AI problem is solved.

 

[Mad Dog]

Maybe AI will do us a favor and impoverish all of those tech elitist multi billionaires.

 

[Parkscope Joe]

Where in that piece does it say that AI fueled the development of Epic Universe?

Hell lets talk about timeframes: Epic Universe seems to have been developed between 2015 to 2022. The first legacy ChatGPT model just for chat purposes, not image generation, was in November 2022. Realistic image generation is still not there and it's still a recent phenomenon.

As you say later the term is being knowlingly abused - prior features that were machine leanring are now AI, LMS are AI, automatic red eye removal they've had in iPhoto since like 2005 is now AI. Hell I saw an article once say games have had "AI" for decades because they conflated enemy movement programing ("AI") with the new surge in LLMs.

 

[Netionster]

As I look at the threads, I'm noticing this topic constantly flies into our faces a bit more frequently. I figured I would take the initiative to create a thread for this as AI has been a topic I've been really keeping track of lately and I think it would generate some interesting discussions here. I created this topic in hopes to discuss what's currently going on in the development of AI and where we think it's going.

My personal opinion: It is going to be here to stay for better or for worse. As an aspiring web developer (I know, worst possible time to look for a job in that area right now LOL), there seems to be a massive scare that the software development industry is going to be completely replaced by AI. Although a lot of developers use AI as a handy tool to assist with their code, I do think we are many, many years away from that being a reality. In fact, I personally think AI is a bit overhyped.

I've seen several videos of people doing some research on this topic. I personally have come to the conclusion that while AI will be here to stay as a tool like ChatGBT, I don't think its advancement will be sustainable. AI is only as good as to what information you feed it and once there is so much content out there that is AI generated, the AI will start to "learn" from the content it put out, thus creating a snowball effect of degradation. Not to mention AI is also being trained using information that is proprietary, which I think is a big reason why an AI collapse is immenant. There are going to be (And already are!) many lawsuits fighting AI using their information. Adobe has been getting a ton of flack for claiming they have proprietary use of your work in their TOS. Also Scarlet Johansen is suing an AI voice software for using her voice despite her not giving them permission.

One guy points out below that Amazon got called out that their biometrics purchasing system at Amazon grocery stores was not AI computers utilizing it, but actually 1000 people in India running it. This was recent as April of this year! He also points out how AI has been stealing information from websites to train the AI.






What I do find genuinely useful are tools that focus on automating repetitive workflows rather than trying to replace developers outright. I've seen people experiment with things like https://latenode.com/products/agent-builder. to connect AI agents with APIs and existing apps, which seems a lot more practical than expecting AI to write entire production systems on its own.

So I'm curious to hear your thoughts on this matter. Again, I don't support advancing AI further as I just don't see how this would benefit humanity, but I don't think it's sustainable. Y'all remember the .com crash of the early 00's? It's starting to sound familiar. As Google and Apple are investing billions into AI and its sustainability being uncertain, not to mention that moral ramifications of it, I just don't see AI being the legs to keep moving forward the way it is. That's why despite the hype and fear-mongering, I'm still going to do my best to crack into the software development industry. I think things will turn around sooner or later. What do y'all think?

I think the hype cycle is real, but I'm less convinced we're heading toward an "AI collapse." We've seen this pattern before with new technologies: expectations shoot way past reality, investment pours in, some companies fail, and then the technology settles into practical use cases.

As for software development, I wouldn't let the current headlines discourage you. AI is genuinely useful for things like explaining unfamiliar code, generating boilerplate, writing tests, or helping debug. But understanding requirements, making architectural decisions, reviewing trade-offs, and maintaining large systems still require people. Even when AI generates code, someone has to know whether it's actually the right solution.

The training data issue is definitely one of the biggest questions going forward. Copyright, licensing, and consent aren't going away, and I expect we'll see more legal clarity over the next few years. That could change how future models are trained without necessarily stopping AI development altogether.