Absolutely Astounding: Billion Dollar Industry
When I first came to this network in 2017 I had grand plans; I was going to build cool stuff and get paid for my work. I do have a computer science education, and the corporate world was not what I was looking for. Crypto seemed like the perfect anarchist fit.
However building cool stuff without a dedicated team is quite difficult. I assumed by now that I'd be a multi-millionaire and could fund these things myself. Alas: it is what it is. I am seemingly trapped in the 6-figure club, for now.
One man army?
However now that AI assistance keeps getting exponentially better over short bursts of innovation I find myself wondering if I could just start cranking stuff out super fast on my own. I'm starting to get highly motivated; now more than ever after I've seen what Grok3 can do.
Behold the Insanity:
~~~ embed:1892950882989273506 twitter metadata:RGFubnlMaW1hbnNldGF8fGh0dHBzOi8vdHdpdHRlci5jb20vRGFubnlMaW1hbnNldGEvc3RhdHVzLzE4OTI5NTA4ODI5ODkyNzM1MDZ8 ~~~
~~~ embed:1892728822878875878 twitter metadata:ampmcmVldGhpbmtlcnx8aHR0cHM6Ly90d2l0dGVyLmNvbS9qamZyZWV0aGlua2VyL3N0YXR1cy8xODkyNzI4ODIyODc4ODc1ODc4fA== ~~~
~~~ embed:1892539000344207536 twitter metadata:cnRlY2hyb21hbmNlcnx8aHR0cHM6Ly90d2l0dGVyLmNvbS9ydGVjaHJvbWFuY2VyL3N0YXR1cy8xODkyNTM5MDAwMzQ0MjA3NTM2fA== ~~~
~~~ embed:1894033762062913866 twitter metadata:cGFuZGV5cGFydWx8fGh0dHBzOi8vdHdpdHRlci5jb20vcGFuZGV5cGFydWwvc3RhdHVzLzE4OTQwMzM3NjIwNjI5MTM4NjZ8 ~~~
~~~ embed:1892506837800431882 twitter metadata:VFNMQXNoYXJlaG9sZGVyfHxodHRwczovL3R3aXR0ZXIuY29tL1RTTEFzaGFyZWhvbGRlci9zdGF0dXMvMTg5MjUwNjgzNzgwMDQzMTg4Mnw= ~~~
~~~ embed:1892690353599234090 twitter metadata:bWlja2V5eGZyaWVkbWFufHxodHRwczovL3R3aXR0ZXIuY29tL21pY2tleXhmcmllZG1hbi9zdGF0dXMvMTg5MjY5MDM1MzU5OTIzNDA5MHw= ~~~
~~~ embed:1892338112640266549 twitter metadata:WHlvbnpoaXN0b3J5fHxodHRwczovL3R3aXR0ZXIuY29tL1h5b256aGlzdG9yeS9zdGF0dXMvMTg5MjMzODExMjY0MDI2NjU0OXw= ~~~
~~~ embed:1892619614858760626 twitter metadata:S2V0dGxlYmVsbERhbnx8aHR0cHM6Ly90d2l0dGVyLmNvbS9LZXR0bGViZWxsRGFuL3N0YXR1cy8xODkyNjE5NjE0ODU4NzYwNjI2fA== ~~~
~~~ embed:1893087549234798620 twitter metadata:TmlsZXNTYW5rZXl8fGh0dHBzOi8vdHdpdHRlci5jb20vTmlsZXNTYW5rZXkvc3RhdHVzLzE4OTMwODc1NDkyMzQ3OTg2MjB8 ~~~
~~~ embed:1892712755750203887 twitter metadata:Z2FyeXNoZW5nfHxodHRwczovL3R3aXR0ZXIuY29tL2dhcnlzaGVuZy9zdGF0dXMvMTg5MjcxMjc1NTc1MDIwMzg4N3w= ~~~
~~~ embed:1893778151295053959 twitter metadata:cG9ueV9wb240NzF8fGh0dHBzOi8vdHdpdHRlci5jb20vcG9ueV9wb240NzEvc3RhdHVzLzE4OTM3NzgxNTEyOTUwNTM5NTl8 ~~~
~~~ embed:1893131535751995903 twitter metadata:WGFuYXJlc198fGh0dHBzOi8vdHdpdHRlci5jb20vWGFuYXJlc18vc3RhdHVzLzE4OTMxMzE1MzU3NTE5OTU5MDN8 ~~~
~~~ embed:1892836942678806585 twitter metadata:S2ltS2F0aWVVU0F8fGh0dHBzOi8vdHdpdHRlci5jb20vS2ltS2F0aWVVU0Evc3RhdHVzLzE4OTI4MzY5NDI2Nzg4MDY1ODV8 ~~~
What the hell!?!
25 games... all created in less than a day's worth of work.
These are the kinds of projects they'd give you in college and might give you a month to complete (or even the entire semester). AI is pumping them out within hours. I just had to test it myself, but before we get into that let me pick out a couple examples from the list above and explain why I find it so impressive.
So this is the guy who created the Pacman clone and claimed that other people who said they created their game in a single prompt were full of it. He seems to know what he's talking about here, and says that ChatGPT is still better for coders, which is remarkable if true. I've really been sleeping on this tech, so it seems.
link: "roguelike LoS"
This one stood out to me because of the line-of-sight mechanic.
Objects in the character's view are obfuscated by structures.
It also shows you how you can start with ascii characters and move to graphics later.
link: "3D city-building simulation"
This one caught my attention because it seems a lot more advanced. This guy clearly knows exactly what he's doing. I also had a idea for a joke game called Simp City and I realize after seeing something like this that it no longer takes months and months of full-time work to get something like that completed.
link: "freestyle"
This one is noteworthy because the AI isn't given a prompt at all... and is rather instructed to create a game based on "what users on X are talking about this week". Which is insane to me that it actually just came up with something on the fly.
link: "turn based strategy"
I've also considered building one of these on Hive. Was thinking about modeling it based on an old game called Mother of all Battles (shareware) pictured above. Apparently something like this can get done in less than a week now. Insanity.
link: "gravity simulation"
It looks like AI is able to handle rudimentary physics with ease.
Impressive.
link: "dominoes"
Not a game but again the physics are impressive for something created instantly.
I could go on but you get the idea.
Testing it for myself
I asked Grok3 how difficult it would be to create a clone of Castle of the Winds, a very very old shareware game from the Windows 3.1 era. Not only did it know everything about the game but it gave very detailed plans about how to get it done.
Notice how AI is not a replacement for coders.
While it is true that AI is going to displace millions of jobs it's also pretty obvious that understanding how to use AI as a tool while still having underlying knowledge of what's going on is going to be an invaluable skillset. Notice how Grok tells me it could take months to build this thing if I have no idea what I'm doing and days if I'm an expert. I guarantee the original developer of this game took at least 6 months to a year building it. So AI is essentially turning complete novices into what would be considered competent coders just a couple of years ago.
I asked why I wouldn't be able to get this game done in a single prompt and Grok reminded me that CotW is a lot more complex than anything these people are posting online today. So I toned down the scope (suggested by Grok) and get a tiny snippet of code just to get something up and running immediately.
# Basic Castle of the Winds starter: Player '@' moves on a 5x5 grid with walls
player_x, player_y = 2, 2 # Start near center
grid = [
['#', '#', '#', '#', '#'], # '#' = wall, '.' = floor
['#', '.', '.', '.', '#'],
['#', '.', '.', '.', '#'],
['#', '.', '.', '.', '#'],
['#', '#', '#', '#', '#']
]
def print_grid():
for y in range(5):
row = ''
for x in range(5):
if x == player_x and y == player_y:
row += '@' # Player
else:
row += grid[y][x] # Wall or floor
print(row)
def move(direction):
global player_x, player_y
moves = {'w': (0, -1), 's': (0, 1), 'a': (-1, 0), 'd': (1, 0)} # Up, down, left, right
dx, dy = moves.get(direction, (0, 0)) # Default to no move if invalid input
new_x, new_y = player_x + dx, player_y + dy
# Check bounds and if the new spot is a floor (not a wall)
if 0 <= new_x < 5 and 0 <= new_y < 5 and grid[new_y][new_x] != '#':
player_x, player_y = new_x, new_y
# Initial display
print("Use WASD to move '@'. Press 'q' to quit.")
print_grid()
# Game loop
while True:
action = input("Move: ").lower()
if action == 'q':
print("Goodbye, adventurer!")
break
move(action)
print_grid()
First: the Python version
Python is a language I haven't used in over a decade, but I miss it.
Python is pretty awesome.
We can see here that all this code does is create a 3x3 grid to move around. The '#' are the walls and the character is not allowed to proceed past them. Even though this is super basic and not nearly as impressive as these other games I was still surprised it really is as easy as copy pasting the code and it works without any kind of error. In fact even the comments within the code are pretty decent...
# Check bounds and if the new spot is a floor (not a wall)
if 0 <= new_x < 5 and 0 <= new_y < 5 and grid[new_y][new_x] != '#':
player_x, player_y = new_x, new_y
Anyone who understands what the AI is doing here can see that some of this code is redundant. The first place I checked was this snippet where the program checks to see if the player is out of bounds or not. It's making sure that we are not only within the 0-4 range of the array, but also that we aren't standing in a wall.
I immediately realized I can delete a chunk of this code and the program will still work exactly the same.
if grid[new_y][new_x] != '#':
player_x, player_y = new_x, new_y
I did this just now.
I deleted the check to make sure we were within the 5x5 square, and it works exactly the same. This is due to the fact that we start within the square and we can't get out of it because of the walls. However it probably is best practice to make sure we are within the bounds of the array or we could experience a runtime error that crashes the entire program. The point is that there's still extreme value in knowing how it works. These tools are very powerful but people still need to know how things work.
<!DOCTYPE html>
<html>
<head>
<title>Castle of the Winds Mini</title>
<style>
#game { font-family: monospace; white-space: pre; }
</style>
</head>
<body>
<div>Use WASD to move '@'. Press 'q' to quit.</div>
<div id="game"></div>
<input id="input" placeholder="Move: " autofocus>
<script>
let playerX = 2, playerY = 2;
const grid = [
['#', '#', '#', '#', '#'],
['#', '.', '.', '.', '#'],
['#', '.', '.', '.', '#'],
['#', '.', '.', '.', '#'],
['#', '#', '#', '#', '#']
];
function printGrid() {
let display = '';
for (let y = 0; y < 5; y++) {
for (let x = 0; x < 5; x++) {
display += (x === playerX && y === playerY) ? '@' : grid[y][x];
}
display += '\n';
}
document.getElementById('game').textContent = display;
}
function move(direction) {
const moves = { 'w': [0, -1], 's': [0, 1], 'a': [-1, 0], 'd': [1, 0] };
const [dx, dy] = moves[direction] || [0, 0];
const newX = playerX + dx, newY = playerY + dy;
if (newX >= 0 && newX < 5 && newY >= 0 && newY < 5 && grid[newY][newX] !== '#') {
playerX = newX;
playerY = newY;
}
}
// Initial setup
printGrid();
const input = document.getElementById('input');
input.addEventListener('keypress', (e) => {
if (e.key === 'Enter') {
const action = input.value.toLowerCase();
input.value = '';
if (action === 'q') {
document.getElementById('game').textContent = 'Goodbye, adventurer!';
} else {
move(action);
printGrid();
}
}
});
</script>
</body>
</html>
Javascript version
The frontend on this one is even cleaner because it updates the webpage in real time and doesn't spam the console with a new map on every move. Again even something super basic like this would take hours to accomplish as a novice in college and the professor might give you a week to complete it. AI cranks it out in two seconds. And yeah that might make teaching more difficult (or just easier to cheat rather) but this is the direction we're moving in so there is no avoiding the new paradigm. On a very real level it makes teaching a lot easier for those willing to learn. And more importantly is supercharges the power of the user to be able to do things they would have never imagined possible at blindingly fast speeds.
Gaming: The Next Shitcoin Casino
So what does any of this have to do with crypto? Well it's very obvious to me that now that AI can make it trivially easy to create a basic game... we will almost certainly see a convergence with crypto. Attach a shitcoin to a game like this and all of a sudden people are gambling their asses off based on a product that could have been created by a 12 year old in Mom's basement. After seeing what I've seen today I have no doubt that the next narrative for crypto is going to be gaming, and that's honestly pretty exciting for me personally because I've been waiting for the better part of a decade for this to happen.
Conclusion
Truly incredible stuff coming out of AI these days.
I can no longer ignore the power of this technology.
Not only can AI help me from getting stuck for hours on end...
It can also prevent the kind of burnout that makes me quit altogether.
It's time to get to work.
AI has come a long way since I last looked at it. Time to dive in I'm thinking.
Some interesting results for not much input.
New crypto narrative sounds good. 👍🏼
I'd love to see some kind of blockchain based CIV clone. I don't know how everything would work, but I still think it would be pretty sweet. I love that game!
I just really want one of these models to be capable of building with Unreal Engine 5. I have been testing tools to create 3D models of images to be used with Unreal Engine 5.
Wow some crazy stuff, things are moving so quickly, won’t be long before we are getting AAA games in weeks rather than years
!BBH
Saving a comment here to read later