I previously wrote about AI-powered image and movie generation, and speculated about its impact on various female-dominated industries. Progress has been quite rapid, so much so that there is a double or triple whammy coming, with AI gearing up to pummel journalism, HR, modelling, and movie making. In this post, I want to highlight a trailer I came across a few days ago:
Midjourney and Runway are AI tools for the visual arts. The former is used for generating images, the latter produces videos. As you can see from the trailer above, the status quo is not that bad. Of course, full videos are not yet within reach, and this trailer surely is not nearly at the level of industry leaders like Wētā FX or Industrial Light & Magic, but this is missing the point. For years, the movie industry has not served us based and red-pilled content, just like the triple-A gaming industry prefers putting ugly chicks in their games, just to spite you. They go the extra mile to create ugly female characters.
In gaming, however, we have seen a genuine revolution, with smaller studios or even one-man teams putting out highly successful games. There are too many examples to mention. Just skimming the current top 100 at Steam, I see games like Rust, Unturned, DayZ, Dead by Daylight, Stardew Valley, Terraria, BattleBit Remastered, or Factorio. Those are all titles that have had greater longevity than the vast majority of their tripe-A brethren, and cost a fraction to make. More importantly, they serve a market the big publishers do not care about. If you want deep games that do not nickel-and-dime you, then there are plenty of options. There is also no plethora of “spicy” games if this is what you are after.
Similarly, creating movies with AI that rival the production quality of the latest Avatar movie or the usual Marvel Cinematic Universe schlock will be out of reach for a long while. However, as an intermediate step, imagine getting movies in 1080p, written by some based and red-pilled creators, that cater to male tastes! I would not be surprised if we got spiritual sequels to 1980s Conan movies within the next five years, and I would be more than happy with that.
The gap in technology is ever-shrinking. Small game studios with a few dozen people can put out really polished products. Similarly, small teams will be able to produce movies of a pretty decent quality within a few years. On a related note, the porn industry is already as good as done as the technology behind Midjourney and Runway has been applied to this area as well, and it is obviously a lot easier to generate variations of the same few scenarios than it is to come up with a story, sketch out a movie with storyboards, and then get AI to produce all the scenes that may only exist in your mind. There is plenty of evidence that we will get there, sooner rather than later.
For gaming, can you foresee a scenario where AI handles over 90% of the testing? I imagine a creator oversees the development of a “level”, for example, with the aid of AI, then goes a step further by letting the AI run a bunch of tests to see if it can break the game. Afterwards, it provides specific instances of errors and even feedback on how to correct said errors, or just patches them automatically with the approval of the human creator. Something like that for TotK could have potentially saved them months or even years of development time.
I can very well imagine such a scenario. In fact, based on what I read of manual QA, part of the work consists of trying anything people can think of to break a game. Nintendo is extremely good at this. They even honor their testers and refer to them as the “Super Mario Club” in the credits of their games. In the West, testing is normally outsourced to third-world countries, or seen as a low-skilled profession, so apparently you get bottom-tier applicants with drug and discipline problems, which leads to high turnover.
With TotK, I am not so sure that they could have shaved off years of development time as large projects are really difficult to coordinate. Getting a few hundred people to collaborate on a creative project over years is a huge challenge. If anything, it is surprising that Nintendo was able to put out such a highly polished game in such a short amount of time. There is no Western equivalent.
“…There is no Western equivalent.”
I’m curious to see how well received Starfield is, as well as if it is broken to smithereens upon launch. I heard about how you can drop a fork on some random planet, for example, and it will always stay in the spot that you left it in no matter how far you travel away or how much time has passed. I don’t know why you would want that, though. I get that it can be cool for other things such as a base you constructed on a moon or something, but I kind of prefer the video game-y behavior of TotK to the ultra realistic style of open world games that the West seems to churn out.
Bethesda has not released a game that was embraced by the public in over a decade, so I would caution. Starfield has a lot of hype behind it, but we are getting into Metal Slug territory where most or all of the people who made the original games great have been scattered to the four winds. Speaking of Bethesda, it could very well be that Skyrim was more of an accident than a deliberate effort, and the fact that the game’s longevity is due to mods instead of the quality of the main game seem to support this position. Thus, Starfield could very well turn into a steaming pile of garbage. There is also the fact that this studio has been bought by Microsoft and it seems that most studios they buy lose enough of their key members of staff to lead to a situation where their post-acquisition output does not in the slightest measure up to the name recognition of the studio.
Allowing the modding community to takeover Skyrim was actually a pretty praise worthy decision at least, I think. Without it, I believe it’s safe to assume that Skyrim’s relevancy would have faded away long ago. A game like Titanfall 2 could have really benefitted from such an environment, whereas now the game is unplayable due to server attacks and the dev’s unwillingness to remedy to situation. However, I still have not forgotten the debacle that was Fallout 76. If I had a GamePass account (which I never will) I’d check Starfield out.
I re-read Hackers by Stephen Levy and Masters of Doom by David Kushner every few years. I think people like John Harris/Carmack/Romero were lucky to live at a time when one guy could be an auteur. That is, one guy could have both a vision and the skills and tools needed to make it a commercially viable, even lucrative, product. An afterword to Hackers, written in the early 2000s, has Ken Williams of Sierra comparing the John Harris days to the modern AAA industry where you need hundreds of developers, artists, animators, voice actors, etc.
I’ve often thought that as the complexity of making games has exploded, the tools haven’t kept up. Carmack and Romero, when making Doom, were inspired by the holodeck in the then-new Star Trek: The Next Generation in 1987. You just say what you want, and there it is. It seems like AI has some promise of fulfilling this.
I’ve been working on AI art for the last few month. So far I haven’t been impressed with the results. But the potential is there. I’ve heard the market for commissioned art has already tanked because of this. Aaron has written about how e-thots will likewise be displaced. (Maybe they’ll have to fall back on more honest work like meatspace escorting.)
Same with voice acting. You can just feed existing audio into an AI and it can say new things in the same voice. It’s far from perfect, but it’s close enough to let you suspend disbelief most of the time.
I agree that the results are hit-or-miss at this point, but the progress is undeniable. A few months ago, there was a story about an artist winning a competition with an AI-generated picture. This is just the beginning. There was a time when humans played against chess computers, expecting to win. Nowadays, chess engines are better than any grand master. Give it a few years, and the best artists may be AIs, but I do not refer to bogus abstract art. Instead, I can picture a future in which realism makes a comeback, with people generating genuinely beautiful art again, thus reclaiming or recreating what the societal elites destroyed about 100 years ago.
The latest Stable Diffusion model, SDXL, dropped a week or so ago. As I understand it, it’s SD 3.0.
With SD 1, those darn kids quickly found you could make NSFW with it. Of course the usual suspects then went REEEEEEEEEEEEEEEEEEE, so SD 2 was “nerfed.” That went over about as well as Nintendo’s no-blood port of the original Mortal Kombat. It’s almost as if listening to the screeching of neurotic Stalinists, instead of to people who buy or use your product, is a bad business decision. Anyway, they seem to have learned from that, because from what I hear SDXL is again NSFW-capable.