A.i. is a blatant case of darwinism.
There are those who adapt, those who will keep moaning about it and finally those who believe it can do everything.
First one will succeed, second one will be replaced, third one is going to get hurt.
I believe this article and the people it mentions are mostly from the second category. Yet no one with all his mind can deny that ai makes writing code faster, not necessarily better but faster, and games at the end are mostly codes.
Of course ai is going to get pushed hard by your ceo, he knows that if he doesn't, another competitor who use it will be able to produce more games, faster and less expensive.
>Yet no one with all his mind can deny that ai makes writing code faster, not necessarily better but faster, and games at the end are mostly codes.
It's actually quite easy, and not uncommon, to deny all of those things. Game code is complex and massively interwoven and relying on generative code that you didn't write and don't fully understand will certainly break down as game systems increase in complexity, and you will struggle to maintain it or make effective modifications -- so ignoring the fact that the quality is lower, there's an argument to be made that it will be "slower" to write in the long term.
I think it's also flat wrong to say games are "mostly codes" -- games are a visual medium and people remember the visual/audio experience they had playing a game. Textures, animations, models, effects, lighting, etc. all define what a game is just as much if not more than the actual gameplay systems. Art is the main differentiating factor between games when you consider that most gameplay systems are derivative of one another.
So on that basis you think the market is happy with shit things made very fast?
I can assure you it's not. And people are starting to realise that there is a lot of shit. And know that LLMs generate it.
The “market” for almost any definition of that you can choose is demonstrably happy with “shit things made fast” provided that the “shit” thing still mostly works. Within rounding error, no one is paying for any artisan hand crafted products. Today it’s AI, last decade it was Chinese products, before that it was Japanese products. If you can get an 80% solution for 50% of the price, most people are doing that. How often do you “sort by price” when shopping? When was the last time you bought the more expensive higher quality version of something you don’t have a passion for?
Or if you want to keep it in the realm of computers, “worse is better” clearly won out. The world uses Linux, not Unix much to the chagrin of the people who wrote the Linux Haters Handbook (regardless of how tongue in cheek that might have been).
And the take away from history should be that AI might be “shit” now, but it will improve. If you don’t remember the days when “Made in Japan” was a marker of “shit”, that’s because things that are “shit” can improve faster than things that are “not shit” can maintain their lead.
You assume it will improve. There is an asymptote from what I can see.
Surely you're not seriously that we've hit the peak (or even the top of an S curve) of what AI is capable of right? Even if you think LLMs have a limit to what they can do and current AI is overhyped (both of which I would agree with) I find it difficult to believe you really think what we have today is the best AI will ever be able to do.
The problem is there isn’t a viable revenue model. The moment you ask money for it the customers disappear. That means there isn’t a revenue stream to support the extreme cost of model training.
And this is all paid for by people who expect a return. In the middle of a very volatile market.
It’s dead even from a non technical perspective.
From a technical perspective every reducing gain requires more money than the last step. That isn’t something that ever works.
> another competitor who use it will be able to produce more games, faster and less expensive
And yet this is no guarantee they will succeed. In fact, the largest franchises and games tend to be the ones that take their time and build for quality. There are a thousand GTA knock-offs on Steam, but it's R* that rakes in the money.
Exactly, cheap buggy shovelware has always been available to gamers. They just choose not to buy it because they know the experience will be sht since no one spent any time polishing it from a POC to a product.
AI generates code that's harder for humans to understand so that polishing process is takes longer and is even more costly when you have AI shtting out code at breakneck speed.