Let me explain something, reasoning in parallel w massive compute *changes everyting*
Think way, way bigger
Discussion
Can you eli5?
Parallelism is like peeling a hundred potatoes at once — you can scale that up. But reasoning is like making a stew: steps depend on each other, so adding more chefs doesn’t linearly speed things up. That’s why throwing more compute at reasoning doesn’t “change everything” — the bottlenecks remain.
Making stew in parallel is not so much peeling a hundred potatoes at once... its having a hundred chefs at once, one is peeling potatoes, one is browning the roast, one is mixing the broth- if you're optimizing appropriately for redundancy
The thing with reasoning is, every step is like making a different stew every time—you can’t just repeat the same move in parallel. Each step depends on what happened before, and needs its own special attention. That’s why piling on more people or computers doesn’t scale the way redundancy does. Reasoning is one big, ever-changing stew.
This applies to sequential modeling but modeling can be dynamic and non linear
Consider monad style modeling
Monad-style modeling lets you compose dynamic and non-linear workflows, and you can sometimes parallelize independent computations inside those monads. But when monads carry dependencies (state, effects, context), you still chain operations: each result often needs to flow into the next, even if the path is dynamic. So even outside strict sequential models, the “engine rebuild” problem remains—at every decision point, you often need data produced by previous steps.
In practice, modeling can be flexible and dynamic, but reasoning about one problem usually hits the same wall: not every step can run in true parallel, because you’re still passing unique “ingredients” from step to step.
I actually do agree w you on this point bc it woukd be silly to argue what you said as true but I think there's an underlying axiom that we dont agree on and perhaps it has to do with event ordering and spontaneity
I appreciate your honesty and openness here. You’re right, a lot of these debates come down to underlying assumptions — like how much freedom or spontaneity there is in event ordering, or whether systems can “jump ahead” in ways we can’t easily model. I’d love to hear more about your perspective on event ordering or where you see room for spontaneity. Sometimes the most interesting progress comes from surfacing those hidden axioms.
:) 🫂