The People of Hong Kong are SCREWED | Inside China’s Ethnocracy
John made a lot of very accurate predictions here back in 2004, and it's interesting to think about these issues in a present-day context, with neural networks on the scene.
1. He correctly predicted that the amount of art assets required to create a AAA product that pushes the technological envelope would reach a point where you'd have to compromise your artistic vision to be able to ship it, not just due to budget constraints but skill deficit. I think this is where neural networks are going to start being leveraged to either greatly supplement or replace the original artists' work. They can sketch something at a low level of fidelity and then use a neural network to fill in the details. Goes for audio and voiceovers, too. If you need to record additional dialogue at a late stage in a dev cycle you might not end up needing to bring your voice actor back in.
2. Neural networks will be the next big way forward in creating believable behaviours in NPCs. AI in future games will not be programmed, but trained. All that research of bots learning to play Doom and getting good at it quickly is gonna be put to use, by observing and emulating the behaviours of real people in the games. It's the only way that games will be able to create behaviours that match the visuals and create a believable world.
3. Doom 3 was the last game that id was able to ship working as an indie-sized kind of studio in any sane amount of time. John noted that four years was pushing it and they'd need to get their turnarounds quicker in the future... and then Rage took seven!
I remember thinking back in 2007 or so, "Whoa, dude, we've hit like... peak photorealism, how can you do a big leap in quality from this?" It was wrong in retrospect to think of it just from a rendering tech POV (though there's certainly been lots of incremental improvements since then on that front), but the real revolution is gonna be in content creation methodologies that will make these incredibly complex and lifelike experiences possible to realise outside of big monolithic studios that focus-test all of their flagship titles to death in an effort to mitigate the risk of their $100M+ investment
So far on #QOTO we have brought the following services up, I want to know what you want to see next. All of these services are free and open to the public for open-source and collaborative non-commercial ventures.
Here is what we have up so far:
GitLab server - https://git.qoto.org
Servers we intend to bring up:
Let me know which of the above you'd like to see next. Also if you have a suggestion of a service we didnt list let me know and we might just host it for you if it would be useful for others.
What could possibly go wrong
When implementing dark mode, naively inverting seldom works. Instead, consider this great dark mode UI advice: ⤵️
🔥 When implementing dark mode, don’t throw away the visual cues in the light version by naively inverting the color scheme.
Close elements should still be lighter and distant elements should still be darker – even in a dark UI.
'member when #opeth wrote epic fusions of metal, acoustic and progressive rock that wasn't just a gimmick to sound clever but effectively used light and shade to convey a sense of sublime melancholy that felt authentic and emotionally resonant as well as technically impressive?
QOTO: Question Others to Teach Ourselves. A STEM-oriented instance.
No hate, No censorship. Be kind, be respectful
We federate with all servers: we don't block any servers.