4 ms·
Well, in my case, I don't believe our universe is a simulation, but I'm open to discussing the idea for fun and it does seem like a possibility. Whereas, most p
by dntrkv 5y ago
Well, in my case, I don't believe our universe is a simulation, but I'm open to discussing the idea for fun and it does seem like a possibility. Whereas, most people that believe in creationism, believe it 100% to be the case and if you don't believe the same you are going to hell. I grew up in an evangelical Christian community and you can't really compare the two groups. Evangelicals are ready to die for this belief.
- UncleOxidant 5y agoThis is mostly the YECs (Young Earth Creationist - "the earth is 6000 years old" camp). There are other flavors like ID (Intelligent Design) that tend to hold things a good bit looser - and there are many different flavors of ID as well. But yeah, the YEC folks are completely "it's our way or the hellway!" and the Evangelicals have pretty much doubled down on YEC - that wasn't always the case, there used to be a lot of Evangelicals that were theistic evolutionists and had no problem with a 4.5B year old earth. EDIT: maybe we need another word in this context besides 'creationist' since it has a lot of baggage in the culture at this point. What else to call someone who hypothesizes that there is some kind of intelligence behind the universe? The simulationists seem to fit into that category as do the various flavors of 'creationist', 'intelligent design', 'theistic evolutionist' and probably even Hindus, etc.
- ericb 5y agoThe term "deist" fits some of those items, although. Interestingly, I think some of the distinction as to why this idea is more palatable is that it doesn't require "supernatural" or "magic" deities. The "creator" could be just like us. We already have evidence that creating virtual worlds is possible--we do it ourselves with games, so I think it takes a lot less faith, as we have a limited proof of principle already. Also, most magical creationism is totally untestable. You can make some predictions about a simulation, though. If simulations are subject to constraints, which is likely, you should be able to ascertain, in the design of the universe, that items with the biggest O might be subject to performance optimizations. If you find lazy loading, caching, or other performance optimizations at the smallest scale (biggest O), which is what this might predict, you at least have some hints.