Honest question:
When was the last time you worked through a hard problem without opening an AI tool?
I think deep critical thinking is quietly becoming a rare skill.
With AI, we juggle too many contexts at once, with no real focus on any one problem. Yes, I can solve many problems at once, but the satisfaction just doesn't hit as hard.
I remember 9th grade, when it was just me and one math problem at a time. It was tedious but rewarding.
Are we trading brain growth for productivity and speed?
Anyone can build anything now.
A cool, differentiated product isn't the moat anymore.
The name of the game is distribution.
The best teams I know have their own in-house content pipeline and UGC system.
If you don't have one, you're falling behind.
We won 1st place at the @ycombinator ร @bitrig hackathon ๐
We built CRATE, a foldable AI sampler for iPhone Duo @Apple
Type "french jazz piano sample" โ a beat in 0.3s.
Finger drum on bottom screen, sequencer on top.
Build and drop with the hinge
We won 1st place at the @ycombinator ร @bitrig hackathon ๐
We built CRATE, a foldable AI sampler for iPhone Duo @Apple
Type "french jazz piano sample" โ a beat in 0.3s.
Finger drum on bottom screen, sequencer on top.
Build and drop with the hinge
Oasis at Speedrun Demo Day is building TikTok for software. Starting with hyper casual mobile games. Already partnered with Atari to recreate iconic IP. App is live on the App Store.
They started with prompt-to-game but never took it out of beta. The founder said "it's cool but I'm not leaving Roblox for this." So they went where AI quality actually works.
Any vibe coder can make games and launch them. People pay real money for virtual pool sticks. This market is barely touched by AI.
host - @brycent at @speedrun
Just built a full Divineko clone in about 4 hours using Cursor, and it actually plays well.
The core mechanic of drawing symbols to banish enemies was the hardest part. I used the $1 Unistroke algorithm (removed the rotation-invariance).
Claude delivered the structure. I spent half the time playtesting, tweaking the confidence threshold, adjusting enemy speeds, haptic feedback, and combo windows until nothing felt off.
Obviously, the game still has lots of room for improvement, but I'm happy with the progress so far.
Before, just getting to this stage would have taken weeks or months. Thereโs clearly so much more we can do in this new era.
You can check it out @playoasiz
#gamedev #AI
Told a friend from back home "and all these games are vibe coded"
I expected him to be impressed at how good they look.
Instead he stared at me blankly and said "...what's vibe coded mean?"
Exactly.
Game devs don't tell their players "this was built on Unity" or "we used Godot for this." AI is a tool, same as any other and it is the output that matters.
We put our games in front of people every day. They're brutally honest. They'll roast a bad game and love a great one. Whether it was vibe coded has never come up once (outside of the SF tech bubble ofc)
Here is one of the hyper casual ones they liked, live on @playoasiz
@redbeansgames Love the layered trick for 3d depth. I used to do light shading by adding highlight to one side and shadows the other to create the illusion of depth as the object moves around.
@sudoingX Itโs only a matter of time before weโre mass-producing AI games in-house. Maybe not today, but very soon. The real challenge isnโt creation, itโs distribution.
The best generator products will be the ones that minimize "viber's block."
I've seen so many AI creator apps where users churn fast because the cognitive load is just too high. They hit that writer's-block-like paralysis where they don't know what to prompt or iterate next, and they bounce.
Treat creative paralysis as a predictable phase to engineer around, not as user failure.