Jokes aside, NGL...stuff like this could have a benefit on game development insofar as shortening dev cycles. If we want AA to return at more significant levels, if we want AAA games to get back to 3-4 year dev cycles and $50-$100 million budgets instead of 6/7 years & $200 - $300+ million costs, it's going to involve stuff like Muse at least in some capacity. I'm sure other 3P, and platform holders like even SIE and Nintendo, are exploring similar things when it comes to how AI can speed up & lower costs of certain parts of the pipeline.
But like usual, these companies better be very careful they aren't ripping off work of hard-working people without their knowledge, without compensation. That can't be tolerated. They can do this stuff but it has to be trained 100% on their own data, IMHO. They better also get out in front of potential regulatory hassles, and ensure they don't cut off too many employees by replacing some ridiculous percent of staff with AI. Like some self-imposed mandate that x amount of actual people need to be staffed among such and such position. Yes, this would be a quota. Sometimes quotas are a good thing.
Ultimately this AI stuff means nothing if you don't have flesh-and-bone human beings involved in every step of the process in some decent capacity. Hopefully if/when the tech is more widespread, those let go from larger teams can just form teams of their own and implement a reasonable amount of AI tech to supplement a portion of staff & costs. I think things like game types and amount of content will just solve themselves over time. I.e not everyone's going to start making GTA clones or try making infinite-hour GAAS because player time & money and finite no matter what.
And ultimately, if only a very small handful make games with the majority of revenue in dollars & playtime, it'll destabilize the market, collapsing the supply chain for tech, leading to massive price increases for the remaining, negating all of the benefits.