Pedro
Admiral
I *really* wish we'd had the tools we have available now when starting the WC4 Remake. AllTinker has liborigin about to release which opens up most of the gains there. The flip side of that to community is I think it's arrived at just the right time to be disheartening; what was a "solved" issue is now being re-solved in a fraction of the time. I certainly look at some of this stuff and it takes the wind out of my sails.
For the rest... you get these claims of "one shotting" a game. But the truth is you could always do that, just hit one of the sample buttons in a popular 3D engine - and you'd start with a base that you're able to maintain. Between originator and Unity you could have something up and running in a day which looked the part. If you look at a 4 year title, especially now - by the end of year 1 you have something largely recognisable as the final product. Creating a polished product means actual added lines of code drop off a cliff at some point and it becomes about refinement and tweaking which is a lot harder in a codebase you didn't create.
I think the barrier to entry is a lot lower now so we will see an absolute flood of attempts. I'm more sceptical as to how many will turn into polished products you actually want to play. But as they say, where there is a will there is a way; this many people experimenting is going to result in something.
For the rest... you get these claims of "one shotting" a game. But the truth is you could always do that, just hit one of the sample buttons in a popular 3D engine - and you'd start with a base that you're able to maintain. Between originator and Unity you could have something up and running in a day which looked the part. If you look at a 4 year title, especially now - by the end of year 1 you have something largely recognisable as the final product. Creating a polished product means actual added lines of code drop off a cliff at some point and it becomes about refinement and tweaking which is a lot harder in a codebase you didn't create.
I think the barrier to entry is a lot lower now so we will see an absolute flood of attempts. I'm more sceptical as to how many will turn into polished products you actually want to play. But as they say, where there is a will there is a way; this many people experimenting is going to result in something.
Vidmaster
Rear Admiral
I am not very interested in the output of a environment-destroying, technically inefficient and massive-theft-derived machine sold by techno-faschists.
I am very much interested in experiences made by humans who cared enough to make them in the first place and to invest their time into learning the required skills and performing the necessary tasks.
While some middle ground does exist, especially in the tool-space, where AI can speed up some repetitive and easily verifiable tasks, its usage is still not very ethical.
None of us are perfectly moral and ethical all the time, but I certainly draw the line at paying money for such a thing or relying on it for anything actually important.
Finally, I also want to point out that the majority of AI-driven remaster mods are garrish and miss the mark. An HD texture in a sparse level does not improve a game's look.
I am very much interested in experiences made by humans who cared enough to make them in the first place and to invest their time into learning the required skills and performing the necessary tasks.
While some middle ground does exist, especially in the tool-space, where AI can speed up some repetitive and easily verifiable tasks, its usage is still not very ethical.
None of us are perfectly moral and ethical all the time, but I certainly draw the line at paying money for such a thing or relying on it for anything actually important.
Finally, I also want to point out that the majority of AI-driven remaster mods are garrish and miss the mark. An HD texture in a sparse level does not improve a game's look.
Last edited:
AllTinker
Commodore
Well for my part I got pretty close to just dropping everything over this stuff. I've been through denial and bargaining, and am finally pretty much at acceptance. In the same way I "accept" other things I have no control over and make me depressed. I don't reject the tech outright; there's some stuff it does well (and more or less ethically) - but its limitations are completely obvious and inherent, and there are of course much deeper and more serious problems lurking.
As someone who revels in the process and craft of software engineering, visual art, audio engineering, composing music, writing and so on - I'm not going to be changing my approach much. I'm perfectly fine with working slower to create things which are actually mine, and that I know the quality of - inside and out. It doesn't matter how "good" genAI gets, it's really besides the point.
I particularly reject any suggestion that programming is just a means to an end - that attitude is extremely prevalent though, and in my experience is the root cause of a lot of the issues we see in so much software. In a cut-throat soulless corporate environment it makes perfect sense... But I thought this was supposed to be the opposite of that - certainly is for me - i.e. people actually giving a crap about making something, not just producing something.
What genAI has changed is my interest in open source. I'll continue to contribute to human-developed open source projects, but that's where my interest stops. I have negative interest in "feeding the machine", and thus I'm afraid libOrigin is unlikely to ever be released publicly.
To summarise; a lot of this stuff makes me quite upset, and after some soul-searching I'm pouring that irritation into making more stuff. I'll continue to take longer to do things, and do my best to make sure they're worth the wait.
As someone who revels in the process and craft of software engineering, visual art, audio engineering, composing music, writing and so on - I'm not going to be changing my approach much. I'm perfectly fine with working slower to create things which are actually mine, and that I know the quality of - inside and out. It doesn't matter how "good" genAI gets, it's really besides the point.
I particularly reject any suggestion that programming is just a means to an end - that attitude is extremely prevalent though, and in my experience is the root cause of a lot of the issues we see in so much software. In a cut-throat soulless corporate environment it makes perfect sense... But I thought this was supposed to be the opposite of that - certainly is for me - i.e. people actually giving a crap about making something, not just producing something.
What genAI has changed is my interest in open source. I'll continue to contribute to human-developed open source projects, but that's where my interest stops. I have negative interest in "feeding the machine", and thus I'm afraid libOrigin is unlikely to ever be released publicly.
To summarise; a lot of this stuff makes me quite upset, and after some soul-searching I'm pouring that irritation into making more stuff. I'll continue to take longer to do things, and do my best to make sure they're worth the wait.
Vidmaster
Rear Admiral
Then do not. Share it with any human who asks nicely, so that true knowledge and cooperation my flow freely but in channels outside of the AI shit.
The data-gobble is real. Here in Germany, there are growing reports of antiquarian bookstores getting 200+ book orders at a time from Canada. Investigations are pending, but it is likely that American AI companies are using Canadian offshoots/fronts to obtain more books to scan and then destroy, circumventing copyright laws (because it is not fair use to do so over here).
I am the lead software architect in the company I work for. Our company has banked on Anthropics' Claude and we are evaluating this stuff.
My preconceived opinion has been that this stuff does not actually make everyday work more efficient, plus all the obvious dependency and software sustainability issues. Honestly, that opinion has not been fully confirmed.
The AI stuff can indeed be useful in practice, for everything of low importance and low life expectancy which can be easily tested/evaluated. Things like GUI-prototypes or single-use-developer scripts, which you throw away after using. It is also surprisingly good at managing CSS. But real backend code, code that needs to be maintained long term, code of real value, of a business value... For that you need a human.
And all the companies who fail to see that will run into giant problems within the next 5 years.
Still, one must (professionally) admit that AI usage can indeed be beneficial in some sections of software development process.
If that benefit is actually economical however, there is still a big question mark over that. Costs will rise, see enshitification. Then there are the hidden costs, you loose knowledge and the ability to train juniors to be good at their jobs. All good for the AI company of course, as these companies want you to "lock in", to be depended, because this is how monopolies can succeed within a platform economy. Not very sustainable if your business utterly depends on external services.
Nowhere is this as apparent as in the so called autonomous business approach that is being promoted here, which truly is the most bonkers PR-lie that I have seen in decades, yes decades! This is such a hard opposite that it can only be described as Orwell-like, think war is peace, freedom is slavery.
For those in favor of AI-first and unwilling to see the economic or politic truth of how problematic this is, let's try a technical argument instead:
The current LLVM-based AI approach will not improve much further and hallucinations are indeed an unsolvable problem (see here), once of the reasons that some predict that symbolic/logic AI could make a comeback within the next decade, potentially in some form of combined system.
Right now, AI systems are degenerating instead because they are getting trained on AI-generated data. But since that does not fit the story, the very important story to keep the unbelievable amount of investor money following, we keep to the success stories. Instead, we read about models so smart that they cause security incidents or must be prevented from being used by foreign citizens.
If you are a software expert/developer/engineer/something, the last three years have been rough when it comes to job opportunities and general media perception. But that phase will end, once it all comes crashing down. Suddenly, people able to analyze problems and actually solve problems will be in very high demand, although it will be very frustrating to deal with the AI-generated legacy code stack then. There is also the danger of big and popular frameworks being ruined in the meantime, of commonly used tech like some DBMS being essentially contaminated and rendered unusable without strong alternatives being ready (I currently predict that Microsoft's software stack is the first to collapse, that Win11 OS is barely working reliably by now).
The AI push is one of the biggest economic gambles our capitalistic societies have ever attempted. The gold rush is driving up prices of consumer goods while telling executives to eliminate the purchasing power of these consumers by giving their jobs (and the executive's value-production-chain) to AI. No one really benefits from this, not even the AI companies (remember this webpage!), except Nvidia who builds the hardware.
Now, they are all hoping that they will benefit in the future, once enough people are locked in and can be squeezed because they have no longer a choice. This might come to pass. Or it will not, who can say.
I do not think it will, I think there is a giant crash coming in a few years. But if I am wrong and the AI evangelists are right, how this a cyberpunkian future, potentially techno-faschist future, any better that our current time? A future in which a few giant mega-corps control most of the means of production? Who really thinks that "we will have more free time" as Bill Gates did put it? The vast majority of people will probably have worse lives than before, scraping to get by, while a minuscule elite will control an unthinkable part of our everyday existence. Real, dystopian shit.
Sorry, this is turning into a rant. Personally, I say fuck off AI. Professionally, I think the AI stuff will collapse hard and those people with real skill at something, be it technical, creative or otherwise, with profit from that again in the long term. The meantime is ugly, however.
Last week, the famous Ring play at the yearly Bayreuth Festival in Germany was changed into an AI-driven performance to do something different for the 150th birthday. And everyone, absolutely everyone, from press to performers to audience, hated it.
Also last week, I overheard a meeting at a restaurant by some marketing person with his customer. PR was among the earliest fields impacted by AI, the possibilities seemed endless. No more photo ops, reduced costs, sounded great for those in charge. According to this meeting, they are already in a hard shift to focus on actual human contact, on human-hosted in-person events, on hiring humans to have a personal business relationship with the customer. Because this is how you stand out from a sea of slop.
AI affects a lot. All of us, in the connected western societies. But there really is little benefit, except for Nvidia.
I'll continue to take longer to do things, and do my best to make sure they're worth the wait.
Your work, and that of many others around here like Pedro from above, is valued. Very valued.
Last edited:
Bandit LOAF
Long Live the Confederation!
I've heard a lot about how AI is going to let people that can't think their way out of a paper back make incredible video games and create the remasters of their dream and so on and so on. I've been told repeatedly it'll be so easy to do things like play Super Wing Commander or Arena thanks to all this. But all we actually get is scammy companies whipping up morons to proselytize using slop videos. This too shall pass, I think the best we can do is use it as a teaching opportunity to show people the actual amount of human thought that has to go into anything worthwhile.