I hate the art part of gamedev. I know, it's weird. I care about how it works and whats the gameplay loop. I care about that. I had Fable have a mental breakdown today by the lack of art direction lol. I whish I knew an artist in my life.
#threejs#gamedev
I keep coming back to this thought while doing game dev:
When done properly, game development is often an order of magnitude more complex than building a “normal” app.
And I don’t really care whether that app is banking, insurance, SaaS, collaborative software, whatever. Those systems can absolutely be complicated.
But games add an entirely different category of problems: real time.
In a normal app, if something takes 1 second instead of 300ms, the user might notice.
If it takes 3 seconds instead of 1, you probably have a performance problem.
In a multiplayer game, if something happens 100ms too late, you may already have a gameplay problem.
Then you open the whole can of worms:
Networking. Packet loss. Latency. Prediction. Reconciliation. Interpolation. Authoritative servers. Physics. Collision. Animation. State synchronization. Tick rates. Bandwidth. Cheating.
And all of these systems have to agree on what happened, where everybody is, and what should happen next continuously, in real time.
Oh, and while doing all of that, you should probably render the thing at 60+ FPS too.
That real-time requirement creates an entire class of problems you can spend a career building conventional applications without ever really encountering.
A CRUD app can be architecturally ugly and still work.
A multiplayer game can be architecturally beautiful and still feel like absolute garbage because a packet arrived 80ms too late.
I’m really trying to give T3 Code a fair shot on Ubuntu 24
Desktop app: Clerk login dance + random OAuth cancellations CLI trying to open a .extension file: “No Apps Available”
(npx t3@latest): node-pty native module goes boom , rm -rf cache, retry.
Android linking: also not vibing . (qr gives "not right credentials"?)
The onboarding experience is currently… character-building.
Still rooting for it though. Linux users are suffering for the greater good 🫡@theo
My GitHub issue for n8ao got closed with “Fixed in 2.0.1”
(https://t.co/FK4yn5cUUR)
and I’m choosing to believe they made a whole release just for me 😭
Meanwhile I slightly improved the placement ghost
and made the shadows stop looking like smeared potato blobs.
It’s not much, but it’s honest work.
#threejs #n8ao #indiedev
@NeoAIForecast Because it's a sign how gullible one can be. If one is willing to throw 4k for "intelligence at home" , chances are, one maybe willing to buy a bridge in another state.
@rxssula@theo I might it a shot. Cause I'm going insane managing multiple codex and multiple Claude on multiple machine. Initially "thought" Bs influencer crap. But that may not be the case indeed.
Were you the guy confused between gigabyte and gigabit ? no wonder you spend money on useless stuff like that. by the time you spend 14K on a frontier model ,which 2 orders of magnitude better then anything you can run "locally", will be 2040, and probably run on a phone. Stop deluding your "fans". There is no "local intelligence". Unfortunatly .it's a fairy tail.
Function? Load-bearing.
Abstraction? Load-bearing.
Config flag? Load-bearing.
Random comment from 2022? CRITICAL load-bearing seam. DO NOT TOUCH.
A) Load — present
B) Bearing — present
C) — not present
D) — not present
I picked C + D.
A + B already exceeded the project’s load-bearing budget.
Just stress-tested the death animation + loot system in my Rust PVE prototype by hatcheting this NPC to death
He dropped wood & stone like a gentleman, I looted the body while standing over him, and everything still somehow works .
Also added a debug room because I’m a professional It’s still pure garbage… but the hatchet feels good #threejs #indiedev #gamedev
@kunchenguid Couldn't have said this better. It's good for "here blender mcp, make me some assets for the prototype". thats it. Oh and is also good if you ask codex to review a plan via opus skill.. genuine find good bugs and gaps. thats it. 20 $ plan and you extrapolate 100% for this.