Sabitlenmiş Tweet
Maxed out 144FPS with 1000 Bots on the map, GG
@threejs is INSANE!
AI Developer@AIOnlyDeveloper
Further pushed to 80-100+ FPS! Fixed the weapons, frags landing, particles flying! @hytopia is capable of alot more now :D
English
AI Developer
1.1K posts
AI Developer
@AIOnlyDeveloper
2 Years of 100% AI Coding Agentic Engineer and Theorist Game Dev Web Dev
Katılım Kasım 2025
199 Takip Edilen939 Takipçiler
My husband and I released our first indie game on Steam. It didn’t do well, but we don’t want to give up.
Starting again, we want to learn from our mistakes. This time, we start small: 2D, pixel art, 6 months of development.
We are super excited about this new project!
English
@unclebobmartin @kristoph @plainionist You can get them to do a review step with a skill, in a new session, given the rules. Then refactor based on the findings.
English
@kristoph @plainionist I doubt it. The AIs are not very good at following rules. Better to enforce the rules externally.
English
@DreamMixGames Also, maybe try another trailer, a bit more action packed, maybe with voice over of people playing, a bit like the 'friend slop' ones, but just to sell the fun of it...maybe look at trailers of mario kart and others...anyway, good work still!
English
I still want to believe in my dreams, even if they are delusions.
I still want to make them real.
Kart Maker is on Steam@DreamMixGames
No publisher. No team. Just years of obsessively learning every part of gamedev. Kart Maker is finally out. Build tracks while driving. Play tons of official tracks. Race, battle, or time attack instantly. Download & share tracks - and custom racers. This is Kart Maker🔊
English
@DreamMixGames Looks really cool!
You shipped, and that is more than most people already!
I hear that getting 10 reviews is key, so maybe try get people you know to review, that could help :)!
I am not sure if you sent to youtubers, but they say sending to 300+ also is key
English
You need to read code differently in the agentic era.
I can scan code super fast these days and identify potential issues.
Source: Shipped multiple games with 100% AI Code, and maybe produced 1-2 mill lines last 18 months. I was coding for 15+ years before AI.
That said, it is easy to go off the rails, so if you are not careful, it can become spaghetti slop code, no doubt about that. The more you know about code before AI, the better!
English
You know at 20,487 (combined add/remove)
We can assume an average and say per line this is 3-4 seconds to read, let's just boldly assume the same to understand.
So what, 8 seconds per line? That's 163,896 seconds.
Now let's divide that by 60, I'll round the decimal down for courtesy - that is 2,731 minutes.
Now let's turn those minutes in to hours - so what, 2,731/60 again!
Rounding to the nearest 1dp point that gives us 45.5 hours.
Now I do work hard, so in my flow state I work for 14 hours straight.
It's gunna take me 3.25 days in my state to even understand what you have produced.
You do not understand, nor read your code. If this is markdown, there is zero way to verify that your AI hasn't hallucinated. This does not belong to you, it has nothing to do with you, you are a vessel.
I like AI, I do not like people who pretend to be productive.
David Cramer@zeeg
im coming for you today @garrytan
English
data goes brrr
rendering out ~30hrs of data per hour
using spot instances
English
@tsoding Comments are code smells, this removes the need (okay, I do see a comment lol).
Either way, it's fine!
English
Waifu Mercs is live on #vibejam ! Check it out! vibej.am/2026/
I'm so excited to participate in the portal system and hope anyone stumbling on this game have a fun time.
Split screen is LIVE and works with 2 gamepads!
English
@heygeorgekal That is fair, I am currently using codex more than Claude, but it changes often :D
English
@AIOnlyDeveloper Well gave the same datasets and instructions to both claude and gpt but gpt solved it and it’s happened more then once lately
English
Spent most of the day on a performance regression in my Three.js game engine.
Lots of profiling and back and forth with AI agents.
Claude kept circling the same idea that wasn't it.
GPT 5.5 eventually nailed the root cause and fixed it.
The tables have turned.
English
@mamesiva64 Very good!
You should lock the camera at certain points and only allow going forward when a certain set of enemies are killed, usually the camera doesn't go backwards either.
English
@KiraniPiebox Amazing! Shipping feels the best :)!
We were shouting in celebration on our first release.
English
Today, after 8 months of development, I released my 1st game: Capyvarias!
The game is not popular but I'm super luck to have the best friends and community. I got a hype train lvl4 and so many people cheering for me on the release stream.
Thank you, really, you are the best 💜
English
@Mappletons We have a home level repo for our .claude and .codex folder's, 2 seperate repos. we share between team on github. Not perfect, but been using it since July, and it's working!
English
How is everyone managing their agent SKILL.md files? Is it just chaos?
Global skills, repo-specific skills, keeping them in sync between machines, figuring out which ones you have installed, authoring new ones. What are we doing? Does anyone have a sane system?
English
@VampireGurlAI @alightinastorm Yep, pretty much giving it expert context, and breaking down the task...spot on 🫡
But sure, people complaining about slop after they try 1 shot stuff with no plan or process xD
English
@alightinastorm The “secret” is iterative decomposition.
Do not ask GPT for magic. Break the problem into parts, give clean context, test one layer at a time, critique the output, feed corrections back in, and keep the loop grounded.
Complexity gets solved by structure, not wishing. >
English
you can't just use chatgpt and expect magic results, but you can use it to break down complexity iteratively
if you want, i can show you THE secret professionals use with gpt that most people miss (essential)
Ryan Florence@ryanflorence
GPT has a new phenomenon that's driving me nuts and I don't quite know how to describe it. - Ask it if I can do something - It says "no you can't [incredibly twisted restatement of what I asked but also not at all what I asked]" - It then tells me how to do the thing wonderfully - And finishes with an insulting "But you can't just [stupid thing I never actually said]" It goes something like this: "Can I form and coach a youth soccer team for my kid and play in P/D level leagues? Or do I have to be part of a full club?" Then it says: "For official competitive teams in Utah, you cannot just form a random team and enter a league. "There is a lesser-known option: UYSA allows independent teams to enter leagues if they meet requirements. [...lists some simple requirements...] "But it’s not “show up with a group of kids on game day”—it’s more like running a small club team administratively. I never said just "show up with a group of kids on game day"! It does this to me with code too. It's so weird.
English
@chongdashu @anulagarwal Yeah, I noticed other day it did 'real' transparency, although it was a little flaky (not 100% of the time).
English
ChatGPT Images 2.0 can now generate really cool UI for your apps/games with TRANSPARENCY!
Previously my biggest concern with the new images model was that it could not add transparency - but the ChatGPT team listened and bought this back :D
Experimenting more, stay tuned!
anul agarwal@anulagarwal
ChatGPT Images 2.0 can generate really cool UI - either for your apps or games. Here I asked it to generate UI for game based around 'Falcons' and for a random health app. It generates everything PERFECTLY and usable in your actual game/app. The only thing I HATE about it is it can't generate transparent images - which is a degrade from the previous version of it's image model. Experimenting more, stay tuned!
English
@arnie_hacker @lucasmonstrox Where do u get the data btw? How high quality is it in terms of player skill?
English
@AIOnlyDeveloper @lucasmonstrox At this scale, no
Pearce et al (~90hrs): arxiv.org/abs/2104.04258
Diamond: diamond-wm.github.io (~90hrs)
I’ll scale this to 100K hours of data across many maps :)
English
cs2 rendering finally works :)
- patched cs demo manager spec_player bug
- multi-cs2 instances per worker
- multi-workers (scalable, fault-tolerant)
what should i generate?
- 100hrs of dust2 gameplay (synchronized, 10-player w/ audio)?
- 100hrs, but of 7 maps (same set-up)
English
@arnie_hacker @lucasmonstrox Awesome!
Here are two worth having a look at:
youtube.com/watch?v=sOgfvb…
youtube.com/watch?v=LpXl0E…
YouTube
YouTube
English