itch.io

this is where im at today, trying a bunch of diff rendering. https://youtu.be/kF7yUj0WYRY

yeah i have actually been reworking the rendering so i will do an update soon.

sweet did you port it or somthing?

holy moly! I love the way you all think!! I am amazed!

in codex lately i have been using. Discuss bla bla bla. Then we are in almost like planning mode but more custom like,,, let’s discuss this. Then after codex discusses it and we have a back and forth and i make sure it understands what i want, i say > Sounds good! Do it!

My starting prompt was basically: build a minimal GPU-first Zig/Vulkan 3D falling-sand prototype.

The stack is Zig 0.16, Vulkan compute/rendering, SDL3 for window/input, GLSL compute shaders compiled to SPIR-V, and Codex as the coding partner.

I broke it into tasks like this:

1. Write the spec first.
2. Define the GPU-first simulation rule.
3. Get Vulkan and SDL3 running.
4. Add voxel buffers and rendering.
5. Add sand/water simulation.
6. Add brush controls.
7. Add transparency, bigger map, lava, steam, smoke, and cooling.

The main design rule was: the destination owns the write. Each voxel gathers what it should become instead of particles pushing into neighbors. That made the compute shader side much cleaner.

What broke first was Vulkan plumbing and sync. The game logic was easier than getting all the buffers, descriptors, shaders, barriers, and draw calls lined up correctly.

The AI misunderstood simulation conflicts at first. Ping-pong buffers alone do not prevent races. The real fix was making every destination cell decide who wins.

The workflow that helped most was: spec first, then implementation plan, then small Codex passes with tests after each feature.

For fun testing, I’m just poking the sandbox. Can I make a mess quickly? Does water behave interestingly? Does lava create steam and smoke in a satisfying way? Do I want to keep playing with it after the feature technically works? That’s the real test.

I actually like the sword resolution. It reads to me like a stylistic choice, not a mistake.

Sometimes breaking the expected pixel-perfect look can make a sprite stand out more, and if you look at it visually, the contrast is pretty interesting.

Also, I’m sure they know about Aseprite. It’s a great tool, and honestly I think it’s worth buying if you do pixel art, but I don’t think every unusual resolution choice needs to be treated like an error.

Peace be with you, and keep making cool stuff.

What happens when people are allowed to use AI art, agents, code help, prompts, and human judgment openly instead of hiding it?

Being an AI-assisted game dev right now takes guts.

We make playable work while people argue.

That is why AI Punk matters.

Build with no apologies.

https://itch.io/jam/ai-punk-jam

Right on! thank you!

All the types only give you power which is converted to either attack, defend or heal. I thought about adding different types of power ups or power per tile, I just was trying to balance everything it was getting quite heave on systems lol. I will probably revisit the design again some time, thanks for playing and thank you for the awesome feedback!

good stuff! thank goodness it's not a dog running around! ;)

the AI Robot ant boss was the only machine really. :D

Thanks!

Love the rigged dance bee animation! I appreciate the use of hexagons too, very beespific. Super smart with the worker bees collecting honey! It's not always easy to get that right. very nicely done, good music and atmosphere too! Has lot's of character.

this is not related to the jam but here is on of my beehive sims, you might appreciate what i was trying to do. If you go long enough the drones start to make with the queen.  https://becool3000.itch.io/beecool2

Thanks, I was trying to fit as many systems into the match 3 as I could. I know it can feel a little overwhelming. I actually had an earlier version with even more going on, and I almost added unlockable tile types with their own behaviors too. But I did not want to make it so complicated that people would just bounce off it.

The tiles are hand-drawn, and so are the little crow character and its animations. The most practical way AI helped was with coding and building the game systems. It was not a magic button. I still spent many hours using agentic coding workflows, testing constantly, and playtesting.

I do not mind using AI. I think it is time people can see that using AI is not automatically evil. There are real ways a skilled developer, artist, or creator can use these tools to explore more ideas and amplify themselves, not replace themselves.

Depending on how your match resolves, you get more "Power" with that match,  that power is applied to attack, defend, or heal, depending on what you picked before the match. Right now it's balance so that if you only attack, then you will die, but if you defend you can build up your defenses and then heal yourself. So that is really the core strategy I didn't reveal it on purpose. Even knowing that i's still fun to play and level up, you get more and more powerful and then you can rank on the high score table and see how you rank against other people in the jam.

Interesting. You must be really good at match 3. That's an interesting idea making the match 3 more difficult. Do you have any ideas how I could make the actual match 3 part harder without making it  impossible?

Nice!

Nice atmosphere and 3D models. I have an older pc, I had a lot of crashes, not sure maybe I need a better computer. :(

Let's stop for a moment or two and consider the strawberry, shall we? I used 2 finger mouse button technique. Fascinating!

I love the character generator and animations, super cool. Nice sound FX and music. Are the level procedural? Nice work!

Thanks, for the feedback,  you must be quick at match 3 haha.
The timer is there to prevent spamming certain moves forever on one turn,
I could have tightened it up a bit, but I didn't want to make it too harsh in case people haven't played much match 3,
I'm not sure what the sweet spot would be.

You are right it could use more polish on the resizing board, it happens because I tried to make the game as reactive as possible,
but I went a little too far lol.
My aim was to fit all the information for the Ui and Hud and still have maximum reactive sizing for different size small and large mobile phones and still be able to maximize on desktop.
Glad you had fun!
Thank you again!

they are supposed to be things a crow would be interesting in. Like how people say crows like shiny things. They are supposed to be shiny things. It’s like the crows collecting trinkets. I drew the trinkets by hand maybe it's my drawing skill? haha. Thanks for playing though hope you have fun playing it. Don't forget to leave your high score! I have a working high score table that links to firestore so it will keep track of everyones score during the jam.

bzzz

Cool, I will try this again on mobile, I think that is good idea that you are targeting mobile, This game is a good fit. I played a little bit with the mouse will try it out again later on mobile. Cool, animation on the little guys.

Yes! this sounds awsome!

Holy cow! awesome concept!

wahh? a bee?

cool

Thanks! I will do much better next jam. I am going to use the whole jam time next time instead of just 3 days! 

I like it! I love these kinds of games. You did good!

Thanks for trying it, yes I didn't get much time to do level design I mostly spent time on the driving physics :) I only had 3 days to work on it. I wish I could have done more content, but I will probably add on to it in the future because it was a game I was planning and then I just happened to notice B1T Jam was going on so I focused on trying to put something out for the jam. Everything is set up just to build levels now, I just have to get to it, but I'm waiting until after the jam.

Very nice! got that survivors, twin stick level up collect xp thing and much more potential! Graphics are super cool too wow! I can only imagine if you did do a true twin stick butterry controls I would have given you a 5 on gameplay instead of a 4. 

seems pretty cool i don't normally play games like this. Interesting mechanic with the camera changes, kind of like not 3rd person almost like 4th person lol.

I like that there seems to be systems and patrons going on that is don't see or understand. :)

lol. with mouse/touch controls: the closer you are to the edge of the screen it increases throttle input. and hang a left by clicking on the left side of the screen after the lift. I made them like that so you still get throttle variance even on touchscreen. I was trying to come up with a way to have more than just on/off throttle. Thanks for playing!

i like these kinds of games ive put in some serious vsl time lol. You got the basics down!

Nice atmosphere.

Haha the shock paddles.

I couldn't aim the lightning. Even when i did hit em they didn't get hurt. I liked the snapping sounds.