[Discussion] From DOOM to "Quantum Compositing": What’s Next for our Batch?

Hey everyone,

You’ve probably seen the headlines lately: “DOOM running on a quantum computer” or “Biological neurons playing DOOM.” It’s cool and it definitely makes a buzz, but what does it actually mean for us Flame Artists? Are we looking at the next big shake-up for our workstations?

To kick things off, here’s a quick breakdown of what’s really going on behind the hype:

1. The “Quandoom” Case (Quantum Computing)

Don’t be fooled by the headlines. This isn’t a real quantum machine running the game in real-time. The Quandoom project is essentially a clever software workaround that rewrites the game’s code into quantum logic gates.

  • The Reality: It needs 70,000 qubits to run, which is way more power than we have today. It’s basically just a clever simulation running on a regular PC.

  • For Flame: Think of it as an extreme stress test. If we ever manage to get this kind of computing power (or even hybrid processors) into our workflow, we’re talking about a massive leap in how we handle complex fluid sims or massive datasets.

  • Source: Hackster.io - Luke Mortimer Proves Doom Really Will Run on Anything

2. The “Biocomputing” Case (Synthetic Biology)

Cortical Labs recently made waves with their “CL1” interface. They basically grew 200,000 human neurons on a chip to “teach” them how to play DOOM using electrical stimuli.

Toward a New Era for Flame

The “Quantum DOOM” stories are fun, but the real question is how this stuff actually ends up in our favorite “Hub.”

Flame is already the heart of the finishing process, especially with all the ML nodes we’ve got now. In the future, it might not just calculate things. It could interpret them:

  • Real-time Physics: Moving from “good enough” simulations to perfect, instant physics right inside Action.

  • Semantic Tracking & Roto: Having the scene analyzed instantly so that rotoscoping becomes a “one-click” job, no matter how messy the footage is.

  • 8K+ Pipeline: Using compression based on information theory to handle RAW masters as smoothly as we currently handle proxies.

References (for the curious)

If you want to dive deeper into the tech:

The Big Question

If Autodesk announced a partnership tomorrow to bring a “quantum” or “neuromorphic” accelerator into Flame, what task, currently a major chore or a bottleneck, would you want to see become a single button in your Batch?

  1. “Real-time” physical simulations in Action?

  2. Perfect, dynamic rotoscoping with zero keyframes?

  3. The ability to relight any 2D element as if it were native 3D?

I’m curious to hear your take. Is this just cool sci-fi, or are we closer to the future of compositing than we think?

Just reminded me…whenever a vendor or demo referred to some new tool that ‘does it perfectly realistic’, I always think about the number of times a client has walked into my sessions and looked at the screen claiming “yeah, that’s better but it still doesn’t look real. The water looks better but still fake, the clouds are more transparent and the boat isn’t bouncing correctly in the water”. Then I would explain that the screen was showing the same real life reference footage as yesterday.

Clients

Can’t keelhaul them

Well…

Can’t throw ‘em down a well either…