Back in the 80's, some of the defining games of the decade were made by a single person, and dev teams hardly consisted of more than two or three people. With the progress in computer hardware, with better resolutions and 3D cards, games developed too, into a space where the best games on the market required teams of 200+ people to design, develop, produce, market, and sell them.
Then, in the 2000's, the indie scene started to gain traction in the public eye. Not because they could compete with AAA studios on their game courts, but because they found clever and novel ways to make games that piqued gamer's interest who were perhaps a bit bored with the increasingly formulaic output they saw from major game developers.
But indie games also had to retract to such alternative game ideas because producing top of the line games had just gone way too far out of reach for a small team with a small budget - let alone a one-man studio.
I think we're at the beginning of new dawn in independent game development. The tools available, even if you just look at open source tools, has never been so rich as they today. Now, with AI entering the scene, the dependence on very large teams will likely go down more and more, at least when combined with clever game design.
Will one person be able to create a solo project that challenges the AAA games of the top-selling studios any time soon? Of course not. But I think we're on the verge of another indie game revolution.
AI threatens entire professions in a way that was previously thought impossible. Even a couple of years ago, hardly anyone would have imagined that jobs that typically require years of practice, training and skill could become equally under pressure as the traditional labour jobs that have been replaced by robots. We cannot even begin to imagine the the long-term effect on the economy at large.
On the upside, for gaming, the consumer will likely benefit from this development as new avenues and opportunies are opened to indie developers. Creating a character sheet automatically is just the beginning. This trend will continue, and it is absolutely foreseeable that further parts of the design process will be automated. Music and game sound generation will also be perfected by AI in the very near future. Actually, that is probably true for all game assets, it's just a matter of time.
Even storytelling - if future games even need predesigned narratives (why not let an enticing story develop on the fly, thanks to AI, that really takes every user action into account and puts it into a coherent context?).
Fortunately/unfortunately (depending on your POV), there's a constant variable that limits all output, and that is peoples time. Instead what we'll see is the collapse of the pillars of culture (much like the collapse of the music industry, when the supply 10xed and delivery time fell to 0). Everyone will be king of a small anthill.
For a more pertinent current example, all you need to do is compare the mobile games industry to the computer games industry. The cost to make a mobile game is very low, and tons of games are released every day. The cost to make a successful mobile game is still very high, because all the money that is saved on development costs are spent in marketing instead.
Also, the video market collapsed in the 80s precisely because of low quality/oversupply games.
You are right, and to go a step further - Generating full games with AI is possible now. Computer games that weren't even considered possible 2 years ago.
I've been playing text-based adventure games that I have made with ChatGPT and they are fascinating. Try the below command in ChatGPT (don't read the text below if you actually want to play it though!).
> Hi, I would like to play a text-based adventure game with you. I will say commands and you will respond either describing the situation or saying what the characters have said. Responses should be kept fairly short. The game will not end until I say the word 'endthegame'. In a combat situation, you will tell me how many hitpoints the character I am in combat with has, and I will describe the attack - you will evaluate that attack and decide how many hitpoints to deduct from the enemy. The game is set on a spaceship, however the spaceship is under attack. There is a portal gun. Part of the solution should be to teleport to the enemy ship and kill the enemies, and this should be telegraphed to the player after a few turns. The goal will be to stop the enemy ship from destroying our ship. I will say some things that will not be possible in context, or that are too vague, and if this happens you should respond saying why I cannot do the action, or that it was too vague. In general you should try and accomodate the players actions though, even if they have negative consequences - but not allow them if they are too vague or are multi-step.
Can the above now be considered 'source code' to a game? A text-based game which is fun to play (at least I find it fun!) and with free-flowing narrative, where your actions have consequences.
It's pretty wild to go into this game, where you can go up to characters and start 'proper' relationships with them, talk in natural language, and be able to make genuine choices about how to solve a problem that the developer didn't even dream of.
You can also generate an infinite amount of quasi-educational text adventures by prompting it with a historical or fictional setting. Give it a piece of classic literature and it will try to come up with something on-theme.
That would be awesome! Huge opportunity in education.
I'm still experimenting with the right prompts to make gameplay fun.
Without prompts GPT-3 will tend to 'play the game for you' or allow you to do non-sensical actions (e.g. if a house is on fire, and your command is 'blow the fire out with your gigantic lungs', it will respond with 'you blow the fire out!' rather than 'you blow but this does not have any effect'), but if you tell it to ignore actions that would not work it tends to go 'on rails' and expect a specific sequence of actions from the player, so just trying to balance those two things!
Imagine a system of tags and account linking which pair your preferred games and genres (on steam and others) to your goodreads list, your spotify, etc.. and generate assets and narrative based on that, while using existing game engines (and their increasingly complex automations) to put those things together into a multimedia experience. Maybe for a while this could create a genre like "RPGMaker" games did, but it has the potential to be more.
For a while at least coherence and consistency are not within reach of those systems, but that doesn't seem impossible. Perhaps for a little while longer I can see jobs emerging for curator of prompts, able to fine tune the experience for what is being generated and knowledgeable about the rapidly changing state of the art.
>Even a couple of years ago, hardly anyone would have imagined that jobs that typically require years of practice, training and skill could become equally under pressure as the traditional labour jobs that have been replaced by robots
This has always happened right through history. Although it is correct that technology developed at such a slow pace until the industrial era (ie cart with 4 wheels major transport until the car, only like 60 years from first flight to landing on the moon).
Previously the transition was over the ages, ie iron, bronze, steel etc. But recent progression means we're going to start seeing stuff like that happen faster; I think atm a huge automation stopper is the cost. We manufacture stuff in China because it's cheaper, they don't automate as much as they could because: * Human workers can still be cheaper, because they pay them so little * Job creation is forced by gov because it would otherwise become a HUGE problem (in China/similar).
Then, in the 2000's, the indie scene started to gain traction in the public eye. Not because they could compete with AAA studios on their game courts, but because they found clever and novel ways to make games that piqued gamer's interest who were perhaps a bit bored with the increasingly formulaic output they saw from major game developers.
But indie games also had to retract to such alternative game ideas because producing top of the line games had just gone way too far out of reach for a small team with a small budget - let alone a one-man studio.
I think we're at the beginning of new dawn in independent game development. The tools available, even if you just look at open source tools, has never been so rich as they today. Now, with AI entering the scene, the dependence on very large teams will likely go down more and more, at least when combined with clever game design.
Will one person be able to create a solo project that challenges the AAA games of the top-selling studios any time soon? Of course not. But I think we're on the verge of another indie game revolution.
AI threatens entire professions in a way that was previously thought impossible. Even a couple of years ago, hardly anyone would have imagined that jobs that typically require years of practice, training and skill could become equally under pressure as the traditional labour jobs that have been replaced by robots. We cannot even begin to imagine the the long-term effect on the economy at large.
On the upside, for gaming, the consumer will likely benefit from this development as new avenues and opportunies are opened to indie developers. Creating a character sheet automatically is just the beginning. This trend will continue, and it is absolutely foreseeable that further parts of the design process will be automated. Music and game sound generation will also be perfected by AI in the very near future. Actually, that is probably true for all game assets, it's just a matter of time.
Even storytelling - if future games even need predesigned narratives (why not let an enticing story develop on the fly, thanks to AI, that really takes every user action into account and puts it into a coherent context?).
We're living in the future.