But the technology is fundamentally immoral because of the way the models were trained, and fundamentally destructive to human culture and thinking.

All big-media models, yes, but it's not something inherent to the process, unless you're taking into account the energy expenditure to calibrate them, in which case there is no workaround.

There is that too. The devastation and unloading of the burden onto communities that can't fight back.

But the technology is fundamentally immoral because of the way the models were trained, and fundamentally destructive to human culture and thinking.

All big-media models, yes, but it's not something inherent to the process, unless you're taking into account the energy expenditure to calibrate them, in which case there is no workaround.

It's worth noting that the survey you linked is from 2025. The tools have vastly improved since then, especially Claude Code. That's part of the reason I'm "optimistic" about their abilities, I've seen them get so much better in real time over the past 9 months. I doubt the tools in 2025 could have vibe coded Halo-in-a-browser, to tie it back to the original post.

The paper was from August 2026 regarding productivity when code review is taken into account.

I’m sure one day it will get even better. But the technology is fundamentally immoral because of the way the models were trained, and fundamentally destructive to human culture and thinking.

If we let it. It's a legitimate criticism regarding some of the training. It does raise the question, though, how else would you train it? If you want a computer to know things, how does it learn? From what sources?

I think your concern is about art, and I share it, but bad art existed before AI. A lot of bad art makes money. The unique talents that individuals bring to art is what makes it great, though, and by definition AI art, without artistic direction, is going to be mid. Did painters go away after the invention of photography?

It's worth noting that the survey you linked is from 2025. The tools have vastly improved since then, especially Claude Code. That's part of the reason I'm "optimistic" about their abilities, I've seen them get so much better in real time over the past 9 months. I doubt the tools in 2025 could have vibe coded Halo-in-a-browser, to tie it back to the original post.

The paper was from August 2026 regarding productivity when code review is taken into account.

I’m sure one day it will get even better. But the technology is fundamentally immoral because of the way the models were trained, and fundamentally destructive to human culture and thinking.

Dude, you are being stubbornly obtuse about this. Multiple people on this forum work with code for a living and have said that it’s useful. Do you think they’re lying?

No, but study after study on professional coders has shown that the ones using AI coding agents estimate they are more productive, when in reality they are less productive overall when you account for everything. It appears to be most effective in doing repetitive tasks rather than solving novel problems.

Are they the exception to the data? Likely not.

In a survey of 49,000 developers, the majority do not trust the accuracy of AI coding agents.

https://stackoverflow.blog/2025/12/29/developers-remain-willing-but-reluctant-to-use-ai-the-2025-developer-survey-res...

The bottleneck appears to be human review of the code, with times and workload increasing when AI coders are used

https://fion.ac/jellyfish.pdf

I'll go by data and not anecdote .

You're mostly right, Cody. AI is not as useful as the hype would make us believe. It's something of a magic trick. It looks like it's doing something it's not, but that doesn't mean it's not doing anything useful.

I won't give you an 80 page paper to read like you gave us, but I will give you Cal Newport. I think he is one of the saner voices on this topic. This article supports some but not all of what you're saying.

AI is going to be good enough for a lot of things. You're a connoisseur of quality. I don't know how discriminating you are about fashion, but it's obvious that a tailored suit is going look better and be superior to something off the rack from Mens Warehouse. But technology made good-enough suits accessible to many people when they need them. Is the world better because of that? I think so, but a high-end fashion designer might not think so. The fashion designer still has a job, though, and people still buy tailored suits.

AI is really good at one thing: pattern recognition. And it's better than anything has ever been at that. That aspect is going to change the world. I believe that many needles in many haystacks are about to be found. In this one aspect, the hype is deserved. And there are many applications for this ability in all kinds of work, including game development.

Newport, at the end of that article, refers to AI as normal technology. That's actually not an original statement (see AI is Normal Technology). That's the space I'm trying to inhabit, but it's tough, with everyone losing their minds in every direction and viewing AI as God or the Devil. I do think we'll be better off if people take a less religious view of it.

Personally, I'm glad to have access to AI. There are a few tasks in my job that I do not enjoy (especially on the administrative side) and AI has made those things less of a burden. Soon it's going to analyze test results that I have to review and give me summaries in seconds. That's the kind of thing that took an hour or two of me combing through details and making notes. I do worry about the misapplication and over reliance on AI, but I don't mind asking it to take care of some of the tedium.

Dude, you are being stubbornly obtuse about this. Multiple people on this forum work with code for a living and have said that it’s useful. Do you think they’re lying?

No, but study after study on professional coders has shown that the ones using AI coding agents estimate they are more productive, when in reality they are less productive overall when you account for everything. It appears to be most effective in doing repetitive tasks rather than solving novel problems.

Are they the exception to the data? Likely not.

In a survey of 49,000 developers, the majority do not trust the accuracy of AI coding agents.

https://stackoverflow.blog/2025/12/29/developers-remain-willing-but-reluctant-to-use-ai-the-2025-developer-survey-res...

The bottleneck appears to be human review of the code, with times and workload increasing when AI coders are used

https://fion.ac/jellyfish.pdf

I'll go by data and not anecdote .

https://www.bloomberg.com/news/articles/2026-10-12/epic-games-delays-disney-project-asks-developers-to-use-more-ai

Of course, it also wouldn't surprise me to find out you use some 15 year old version of your editing software because updates only ruin things (not an unreasonable position given the previous 15 years).

More like 15 years ago was the last time you could actually just fucking buy a piece of software and own it forever instead of having to pay a monthly fee.

Of course, it also wouldn't surprise me to find out you use some 15 year old version of your editing software because updates only ruin things (not an unreasonable position given the previous 15 years).

This is actually true. You would not believe the number of pre 2018 Avid suites in Hollywood that are still in use. And Thelma Schoonmaker still cuts on Lightworks for some strange reason.

I can't even remember the last job I worked that was on the latest version of any NLE. Heck, on my current job we downgraded versions.

I definitely put this on my list. I enjoyed Ace Combat 7 (the first and only Ace Combat I've ever played), but I didn't love it the way lots of people seem to.

YOU my friend need to first finish FFVII Rebirth, then play Metal Gear Solid 1-5 followed by Rising Revengence.

Yeah, that's the list I added Ace Combat 8 to. It's the "Hopefully, maybe someday, but probably not actually gonna happen" list.

Speaking of AC6, you can now order your squad around again. Just don't leave them on "at will SP usage" for too long. They chew through those things...

Btw, I was just flying my F-22 and got knocked on my ass by a sheep.

You're not supposed to hit the baa-rricade

I definitely put this on my list. I enjoyed Ace Combat 7 (the first and only Ace Combat I've ever played), but I didn't love it the way lots of people seem to.

YOU my friend need to first finish FFVII Rebirth, then play Metal Gear Solid 1-5 followed by Rising Revengence.

Of course, it also wouldn't surprise me to find out you use some 15 year old version of your editing software because updates only ruin things (not an unreasonable position given the previous 15 years).

Btw, I was just flying my F-22 and got knocked on my ass by a sheep.

I can only provide my own experience (and see the results of others) that what once took a month, can now be done in a day. Now, I'm no computer science major and nor am I doing anything computationally complex (mostly just data processing and calculation tools), so maybe for video game development the gains aren't so substantial, but I don't see how a developer could justify forgoing their use.

"I've sold my soul to the Devil, and as a result I can work 30x faster! Such productivity! How can anyone justify NOT selling their soul?!"

I can only provide my own experience (and see the results of others) that what once took a month, can now be done in a day. Now, I'm no computer science major and nor am I doing anything computationally complex (mostly just data processing and calculation tools), so maybe for video game development the gains aren't so substantial, but I don't see how a developer could justify forgoing their use.

If you're a big enough software developer to have yearly reviews for your employees, AI usage is going to be metric, and you'll be asked how you improved something with it--if only a process. A positive frame is that it can reduce tedious tasks, which frees up energy and time for more creative tasks. A negative frame is that coders don't know the code or they don't know it as well as they would have had they written it.

I definitely put this on my list. I enjoyed Ace Combat 7 (the first and only Ace Combat I've ever played), but I didn't love it the way lots of people seem to.

If that was your first experience in the series, I would highly recommend you try out Ace Combat 6. IMO, the dynamic objectives are more fun, your squad mates are more involved and the story is less cringy. I think it might still be an xbox exclusive but give the AI decompilers some time and I'm sure it'll be PC available if not already.

I'll get 8 when it goes on sale.