We've been slowly chipping away at the idea of absolute determinism and arrived at the solution that a probabilistic multiverse is entirely possible and if so -- the idea that it could be a simulation becomes increasingly likely. If that wasn't enough, our protagonists even started discussing the possibility of it being an optimization simulation.
Features of a universal optimization
BrainI'm not just pulling it out of my arse you know. There are clues. For example, it has a very clear objective function.✓✓
An objective function, you say? What is it optimizing for? You torturing me with late night philosophy?
It's minimizing free energy, obviously. The Laws of Thermodynamics.✓✓
or maximizing human stupidity? Or people being an arse. That's certainly an act of free will.
What is the act of being an arsehole called? Arseholery?
Arseholstery?
Nah, that implies it's about holstering your arse, which is what underwear is for
Assery?
Isn't this where donkeys are made?
The implication is that the timeline (and perhaps other timelines) has an objective -- to minimize free energy according to the laws of thermodynamics. If it's truly and optimization, be it created or self-defined, it must have unknowns or variables whose values it needs to fill. It won't work otherwise.
Why have alternative timelines?
BrainHere's the thing. You can't optimize something without an objective. So if it is acting like an optimization algorithm, then it must have an objective at any point in time. So far this tracks. You also need to have unknowns or it won't run. We know it runs because:
I think, therefore I am
What's the unknown?
✓✓I don't know dude, I just want parallel timelines to exist.
Right. There's a pretty decent chance they do.✓✓
Search spaces?
Search spaces.✓✓
If you're looking for an unknown via a simulation your best bet is to run multiple simulations. It speeds up the process through parallelization.✓✓
I don't see your point. If it's an optimization like you say then it's a colossal simulation, implying above godlike level of power. If someone or something has this much power why bother simulating anything?
Yeah, can't they just get to the answer? That's a great point. Why bother with all the steps?✓✓
Are you accepting it could be purposeless?
Of course it could be purposeless as an initiative but it's definitely operating with a goal as an algorithm. By extension -- it's searching for an optimum.✓✓
Alright, I'll bite. What's the unknown?
If it's probabilistic then there's stochasticity to it. It could just be trying to fill in the variable values which satisfy the laws of thermodynamics, moving towards the most amount of entropy. So it's a bit like a dynamic optimization, a differential equation if you will.✓✓
I've always felt like running optimizations as differential equations is like having a car and pushing it instead of driving it. You don't take advantage of the real power of the technology -- abstracting time. At the same time, your objective can only see one time step at a time.✓✓
I agree with that. Just run an Ordinary Differential Equation (ODE) at that point, even if you need extra constants. Don't even bother with optimizing. The thing is though -- ODEs are deterministic.
You could make them stochastic but the point is you'll be using the wrong tool for the job. The power of optimization is in abstracting.✓✓
So the alternative timelines could be a giant data collection exercise, trying to see what you may do when placed in different circumstances? That's the current working idea and you don't have to subscribe to it, just entertain it for now.
By the way, isn't this what Buddhism is all about? The only deviation from the standard understanding of it is that you may be undergoing all of your reincarnations simultaneously. I gotta say -- this sounds like a pretty efficient use of time. Perhaps time isn't the right word here.
What is time anyway?
What the hell is time?
BrainDoesn't that contradict what you're saying? Optimization abstracts time away. Yet we live in a very time-centered universe.
True but that's from your perspective. The actual simulation might have began its equivalent of a nanosecond ago. The time it simulates may not behave the same way as the time it experiences.✓✓
Hmmm, like a graph optimization? Time is expressed as the flow of causality, the order in which the flow hits up different nodes.
Yes! That's how I explain myself time in this worldview. It's really only from our perspective.✓✓
Time is nothing more than the order in which things happen. We just romanticize it because we don't know exactly what could've happened otherwise. The brain was right -- it's very much like nodes on a graph. Each node represents an event. The order in which these events get connected is the timeline. The simulation itself doesn't need to calculate time explicitly. The feeling of time spontaneously arises from the perspective of the nodes.
But what of the purpose?
What is the unknown in this simulation?
BrainIt still irks me that the simulation you're describing is so damn specific. It's a niche method. You said that if it's a simulation and alternative timelines exist that it must be searching for something that it can't calculate.
Right. If it's a spontaneous simulation it may not even bother with alternative timelines, which would be against our time travel constraint and the pasta model.✓✓
So if time travel exists then it must be interested in the result. It could also be that the multiverse is completely purposeless and just happens to be self-parallelized but that should be the last possible conclusion for us, as it's not actionable.✓✓
Kind of like how astronomers say that it's not aliens unless it's nothing else?
Right. This principle has worked surprisingly well but there will come a time when it will be aliens.
Do you think it's searching for what free will would do?
That would be extremely congruent with the rest of our logic.✓✓
So it's running randomly diversified versions of the timeline, collecting the data to see how free will tends to influence decisions?
That's what I think.✓✓
Are the timelines incumbent solutions then? To be filtered out so the one best one is picked?
I said it's an optimization but that doesn't mean it's optimizing by searching for the one true timeline. It's an optimization with respect to how the simulation of individual timelines is running, i.e. the minimization of free energy according to the laws of physics. There could be other constraints and laws too.✓✓
Most importantly, time travel would be impossible if the alternative timelines were incumbent solutions. Jumping between incumbent solutions would break causality. I think the multiverse resembles more a stochastic sampling operation. Each timeline is its own optimization. You're the data leak.✓✓
So a simulation with free will as the main variable and samples that can influence one another. Sounds like a flask with cells to me more than your fancy search algorithm.
Cell cultures do behave much like stochastically diversified samples. Each one is an independent optimization algorithm but they influence each other. Evolution is a really powerful optimization algorithm but it's actually not that perfect. You'd like that since it implies a relatively more pointless multiverse.✓✓
Time travel is possible. Free will is required. The timeline is an objective-minimization simulation. The purpose could be to see what you make out of your gift of free will. This makes sense to me personally. How would you know what is someone truly made of unless you put them in different situations for every second of their lives.
If that didn't give you an extra brain wrinkle I don't know what would. Perhaps what the topology of the multiverse must look like?
[ < Prev ] [ Next > ]
![map[email:you@example.com name:Alex Mihnev]](/images/avatar.png){kind=avatar}
