00:00It seems that we have more insights from CDPR's Mikhail Novakovsky, and this time he's saying
00:06that he is skeptical of fully AI-generated games. Now, this is a very peculiar topic,
00:11so bear with me, because I'm going to go a bit more in-depth than I usually do.
00:16First and foremost, I think Novakovsky is acknowledging something many people in the
00:20industry privately believe. AI is inevitable as a tool, but that doesn't automatically mean
00:25it should become the creator. And the example he shared, 40 prototypes in a week, five games
00:31selected and one shipped within three weeks, that really sounds impressive on paper. But then it
00:37raises a question, are we optimizing for creativity or just for output? You need to understand something,
00:43games aren't just collections of assets and code. They are the results of thousands of creative
00:49decisions, iterations, failures and happy accidents. Games that people remember for decades like The
00:55Witcher 3, Red Dead Redemption 2, Baldur's Gate 3 or Elden Ring, they weren't created because
01:01somebody found the fastest path to release. They were created over years. Also, when you think about
01:08these things even further, the iconic characters in the games were also the product of this particular
01:13process. Take Kreia, for example, from Knights of Dwarf Republic 2, Sith Lords. I don't really think
01:20that AI can create that type of a character. Because if you think about it, nobody even came close to
01:26create someone as iconic as Kreia. But basically, if everyone can produce 40 prototypes a week,
01:33then prototypes themselves become cheap. And then the bottleneck isn't generating ideas. It's pretty much
01:39recognizing which ideas are worth pursuing and refining them into something special. Also,
01:45if AI makes game production nearly instant, platforms could become overwhelmed by thousands
01:50of derivative titles chasing trends. And then what happens is, discoverability becomes harder,
01:55and you may struggle to separate genuine creativity experiences from algorithmically
02:01assembled products. And to be honest, I don't think anyone wants this long term, at least. So ironically,
02:07fully AI generated games might make handcrafted games more valuable, not less. And eventually,
02:14human creativity could become a selling point. I mean, think about it, if you just keep dishing out
02:19slop. Eventually, because it will take some time, you will figure out what's good and what's bad. And
02:25then, long term, human creativity will become the most important thing ever. So if you think about it,
02:32we just need to weather the storm with all this AI bullshit, and things will eventually become balanced.
02:38Now, it seems that I am a huge opponent of AI. But I actually like the technology. And I believe
02:45where
02:46AI has enormous potential is in eliminating tedious work. For example, generating placeholder assets,
02:53assisting with localization, helping with bug fixing, accelerating testing and QA. All these roles
03:00could be enhanced by AI, which can be a powerful tool that amplifies human creativity rather than replacing it.
03:07So essentially, what I'm saying, and what everyone should look at AI, it's basically a tool. Use it
03:13to accelerate stuff, don't use it to replace humans. Because we've seen what happens when you replace
03:19valuable people, everything goes to shit. Because, as always, human resources are the most important
03:26thing for any company. And if you, for whatever reason, forget that, I don't think that you're going
03:32to have a pretty decent future to begin with. One interesting point that I would like to make is
03:38that games are more than efficiency. This needs to be hammered really hard. Efficiency isn't always
03:44the highest goal in art. Nobody praises a novel because it was written in three days. I mean,
03:50you can do that, but again, what outcome are we going to have? Similarly, nobody says their favorite
03:55album is great because it took two weeks to produce. In this sense, the emotional impact matters
04:00more than the production speed. Which brings me to my next point. Some of gaming's greatest
04:06moments came from people obsessing over details. For example, Arthur Morgan's journey in Red Dead
04:12Redemption 2, the bloody Baron questline in The Witcher 3, and all those amazing relationships
04:18between characters in Baldur's Gate 3. These experiences feel meaningful because human beings poured
04:25years of thought and emotion into them. But then you can say, AI can do this if you feed enough
04:30data
04:31into it. And yeah, you can say, I can imitate patterns, but it doesn't have experiences, memories,
04:37frustrations, passions, or personal perspectives to express. Will that be a thing in the future?
04:43Well, I don't know, maybe, but as of right now, it's a no-go as far as I'm concerned. Okay,
04:48so I talked
04:49about big games that we all know, but what about indie developers? Well, indie developers and small
04:55teams who simply don't have the resources of a CDPR or Rockstar are definitely going to rely on AI
05:00heavily. In that sense, AI could democratize game development by allowing one or two people to
05:06accomplish things that previously required dozens. Not to mention, it can also significantly reduce
05:12development time as well. And that aspect is exciting, because some amazing games could emerge
05:18from creators who otherwise never would have had a chance. So basically, the problem isn't AI itself,
05:24it's the pursuit of replacing people entirely. Because companies have always sought to lower costs
05:30and increase output. And that's exactly what AI offers. If you know a bit about economics,
05:36the biggest expense for any company are the salaries. And if you reduce the salaries, you immediately
05:42increase the profits. And investors love that kind of efficiency. But history shows that
05:48cheaper production doesn't necessarily produce better art. Take Hollywood, for example. They can
05:53generate endless sequels and franchises, but audiences still celebrate exceptional movies.
05:58Music production software, for example, became accessible to everyone, but truly memorable artists
06:04remain rare. I mean, I, a common pleb, I can start doing this music production bullshit, but I will never
06:11have that thing that pro artists have. Because simply put, I'm completely retarded when it comes to music.
06:18However, my point here is that the ease of creation doesn't eliminate creativity. It made originality
06:24more important. So pretty much when you listen to my rambling, my stance when it comes to the AI is
06:30pretty
06:30much in the middle. I am not for AI 100%, but also I'm not against it. But I know one
06:37thing. A future where
06:38studios prioritize volume or revision is never a good one. Imagine hundreds of games designed by
06:45algorithms analyzing engagement metrics. And then they optimize to maximize retention and spending,
06:51constantly chasing trends instead of taking risks. To me, that sounds less like an artistic medium and
06:57more like a content factory. You gotta understand, the greatest games often come from weird ideas and
07:03passionate creators willing to fail. And those things are hard to quantify and impossible to optimize
07:09perfectly. For example, if we made every single game based on, I don't know, metrics and whatnot,
07:16then Mugenics would never be a thing. And people just adore that freaking game. It's insane. So you might
07:22ask yourself, can AI ever make something like that? And I think the answer is pretty much obvious. In any
07:29case,
07:29I don't think fully AI-generated games will replace human-made games any more than synthesizers
07:35replace musicians or cameras replace painters. AI-generated games will exist that's 100% accurate,
07:42and some will probably be fun. Maybe some of them will eventually become hits. But when people talk about
07:49masterpieces they'll remember 20 years from now, I suspect they will still be talking about games
07:55shaped by human vision. Like I said before, AI may help build those games, but I doubt it will
08:01replace the thing that gives them meaning. And at the end of the day, I think that's ultimately what
08:07Nowakowski is getting at. Or at least I hope that's what he is getting at. Just because something can be
08:13automated doesn't mean automation is the destination. Sometimes the value lies not in how quickly something
08:20was made, but in why it was made in the first place. And that's basically the news. Thanks very much
08:26for
08:27watching, and as always, like, share, subscribe, comment, all that good stuff. And that's it from me, until the next
08:33one.
Comments