Is Blizzard going to switch to AI in any particularly notable way? No. They're the premium brand and the whole point is that it's the higher-quality, higher-priced option (or, with how software scale works, they're higher-quality which they then aim to translate into more scale).
So yes I assume artists at prominent, quality-focused studios are fine.
That's like what, < 1% of total employed digital artists? For the other 99%, this is an armageddon. I'm sure nobody's dream job is to make the 934652352345th hot anime girl for generic gacha mobile game #235624563426, but it pays the bills and can be a stepping stone to jobs they might like better.
Software is already somewhat hardened against this because while there are (were?) plenty of jobs making generic hot anime girls, in software we'd just have 1 person made hotanimegirl4j and then we can all copy that forever. E.g. nobody is out there employed re-writing GNU coreutils over and over (is it a fair comparison to say that hot anime girls is to art as the ls program is to software?)
So yeah. I feel for these people who will be economically impacted by this. You can already see this in the patreon numbers of artists who specialize in certain types of content. People are getting their fix by DIYing it, and competition exploded since there's so many more people offering that content now.
Sure graphic designers and artists will do the main stuff but about all the background stuff or user interface stuff?
Why would a major studio have an entire team for something an AI can do very well?
What about the tens and tens of thousands of lines of code that exist only to make other code work correctly, securely, redundantly, efficiently, etc.
What abouta chatGPT powered NPC system?
The "AI Armageddon" is almost more significant to an actual studio than a freelance or small time designer. EA can adopt all the cool stuff Microsoft throws at them from OpenAI - they can afford it.
I'd say that a very large portion of games, that come to market, are almost 100% copies of other games. The whole industry is copying ideas from everyone else. So, yes, an ai is perfect for the job. You need to launch faster than ever, you need to iterate faster than ever, and a machine, that can produce games 24/7 is the wet dream of every game studio.
Of course the games will be shit, but the most successful ones already are, just look at the charts in the (mobile) app stores.
I think AI will do best with already established content - Star Wars for example. An AI can make endless expansions to a Star Wars game, if the game was made so that it could do so.
That's more the future than an AI being made/licensed by EA that creates Madden 2030 from nothing. Changing Madden 2029 into 2030 and updating the jnterface/visuals and making it faster/more streamlined and identifying places to add further content, creating that content, managing that content - all of that an AI system can do.
Yeah, lots of generic copy mobile type games are coming but this is absolutely going to be more than that.
Once they have to pay the actual costs of AI after all the VC money dries up, it will be clear that humans are much cheaper than AI. A $20 AI tool is just an introductory price, and that's mostly because companies don't want to pass on the real cost because they want people hooked.
I disagree. If it really sees widespread use, the AI providers can still make tons of money while being cheaper. This is especially true for higher paying jobs, jobs with benefits, and the reduction in potential legal risk (after the initial firings anyways).
AI image gen in particular is really cheap. You can run the best state of the art open source model, which also gives you much more control than something like DALL-E, on a consumer GPU.
I didn't expect image gen models to be so widely used. Would be nice to have some concrete numbers instead of these anecdotes the article provides. Bullish case for ai though
11 comments
[ 2.4 ms ] story [ 50.7 ms ] threadIs Blizzard going to switch to AI in any particularly notable way? No. They're the premium brand and the whole point is that it's the higher-quality, higher-priced option (or, with how software scale works, they're higher-quality which they then aim to translate into more scale).
So yes I assume artists at prominent, quality-focused studios are fine.
That's like what, < 1% of total employed digital artists? For the other 99%, this is an armageddon. I'm sure nobody's dream job is to make the 934652352345th hot anime girl for generic gacha mobile game #235624563426, but it pays the bills and can be a stepping stone to jobs they might like better.
Software is already somewhat hardened against this because while there are (were?) plenty of jobs making generic hot anime girls, in software we'd just have 1 person made hotanimegirl4j and then we can all copy that forever. E.g. nobody is out there employed re-writing GNU coreutils over and over (is it a fair comparison to say that hot anime girls is to art as the ls program is to software?)
So yeah. I feel for these people who will be economically impacted by this. You can already see this in the patreon numbers of artists who specialize in certain types of content. People are getting their fix by DIYing it, and competition exploded since there's so many more people offering that content now.
Why would a major studio have an entire team for something an AI can do very well?
What about the tens and tens of thousands of lines of code that exist only to make other code work correctly, securely, redundantly, efficiently, etc.
What abouta chatGPT powered NPC system?
The "AI Armageddon" is almost more significant to an actual studio than a freelance or small time designer. EA can adopt all the cool stuff Microsoft throws at them from OpenAI - they can afford it.
Smaller studios maybe not.
That's more the future than an AI being made/licensed by EA that creates Madden 2030 from nothing. Changing Madden 2029 into 2030 and updating the jnterface/visuals and making it faster/more streamlined and identifying places to add further content, creating that content, managing that content - all of that an AI system can do.
Yeah, lots of generic copy mobile type games are coming but this is absolutely going to be more than that.