Post

This is why we need to go back to PS2/Gamecube limits for game graphics We could be making this type of stuff with extremely short dev cycles because the art tools are so fucking good nowadays, but all the higher ups are interested in is making sure we can see Naked Snake's nose hairs

Burnout 3 is underrated as one of the very, very best games of all time because you can show Burnout 3 footage to people two decades later and their reaction will still be "hell yeah, this rules. I wanna play this right now."

It's absurdly easy to make a game that looks gorgeous β€” and will stand the test of time based on looks alone β€” as long as you're willing to accept limits to things like polygon count and texture sizes The indie space already knows this, but god help you if you suggest lower fidelity at a AAA studio

3 reposts59 likes

Consider these characters. I made 12 over roughly three months, and the vast majority of that time was spent just on R&D. Once I had the process down, I was knocking out two per week. That's literally impossible with current AAA specs. bsky.app/profile/xavi...

I had the pleasure of working with [REDACTED] on these 12 portraits over the last few months of 2025, and it was a fantastic technical and artistic challenge. I have permission to talk about some of the technical process, so I'll post a bit about that below!

We're doing too much work with no payoff. Nobody cares if your characters have hair that's strands or hair that's alpha cards. Nobody gives a shit about muscle sims and ML deformers. Not a single goddamn person will tell you that an 8k facial pore occlusion map changed how they look at art.

I dunno, I think muscle sims and deformers are pretty cool on more muscular characters and monsters and make a pretty big impact. And in third person games, hair rendering does have a huge effect considering the head takes up like 10% of your screen at all times

I feel like those are just waiting for simplified "good enough" optimizations. The hair tech in Indiana Jones didn't make the game play better, and looked visibly worse than standard card methods at anything below Ultra settings. What's the point of adding filesize/vram bulk at that point?

When I started at my new studio I straight up asked "so do we have a rock library that we re-use between games" and they sorta laughed at me. Someone said the normal maps from the previous project wouldn't be good enough. I was trying to think of why on earth any players would care.

I cannot even count the number of times someone has come to me worried about something in game, that 3 players are ever going to notice, and not one will give a shit about. β€œYes, I see the LOD on that model changing 100m away. This is a boss fight with 8000 lasers firing at once.”

So why not like…make the rocks with obscenely large textures you can just scale down for different platforms and needs? Or at least just have a mesh and adjustable smart material ready to go. There are options, I don’t understand why anyone would laugh at this

The tools to make incredible things are here now and we're wasting hours every day making perfect, shiny corporate art We could be making lots of different beautiful things that please us, make our lives richer and more interesting, and also run at 60fps on a Compaq from 2004