AI summary by TwinMind
The Simulation Hypothesis is Pseudoscience
Sabine Hossenfelder
10m 5s
Simulation Hypothesis Overview
- Everything we experience coded by intelligent being
- We are part of computer code living in computation
- Laws of nature are mathematical - universe computing those laws
- Controversial aspect: assumes another reality level where someone controls/interferes with natural laws
- Similar to monotheistic religions: omniscient hidden being that can interfere with nature
Nick Bostrom's Argument
- Philosopher claims likely we live in computer simulation
- Logic: If (a) many civilizations exist AND (b) they build computers running simulations of conscious beings, THEN (c) many more simulated conscious beings than real ones exist
- Therefore you're likely living in simulation
- Elon Musk supports this view ("most likely we're in simulation")
- Neil deGrasse Tyson gives >50% odds of being correct
Science vs Religion Distinction
- Religious stories don't explain HOW things work (e.g., Jesus healing blind/lame)
- Believers take on faith
- Science requires explanations for how something works
- Simulation hypothesis mixes science with religion
Problems with Bostrom's Premises
- Could be wrong: maybe no other civilizations OR they're not interested in simulations
- Doesn't invalidate argument structure, just means conclusion can't be drawn
- No good evidence either way
Consciousness Simulation Issue
- Critics assume human-like consciousness can't be simulated
- Actually requires explanation to assume it's NOT possible
- Consciousness likely property of systems processing large amounts of information
- Physical basis doesn't matter: neurons, transistors, or transistors believing they're neurons
- Simulating consciousness not the problematic part
Core Physics Problem
- Bostrom assumes possible to reproduce all observations using different underlying algorithm instead of natural laws
- Implicitly claims easy to reproduce physics foundations with something else
- Nobody knows how to reproduce general relativity + standard model of particle physics from computer algorithm
- Can approximate known laws with computer simulation (done regularly)
- But if nature actually worked this way, we could see the difference
Experimental Evidence Against
- Physicists searched for signs natural laws proceed like computer code
- Search came up empty-handed
- Possible to tell difference because algorithmic reproduction usually incompatible with symmetries of Einstein's special/general relativity theories
- "Not easy to outdo Einstein"
Quantum Computer Limitation
- Doesn't help to assume simulation runs on quantum computer
- Quantum computers are special purpose machines
- Nobody knows how to put general relativity on quantum computer
Information Compression Problem
- Civilization needs to simulate many conscious beings
- These beings will themselves try to simulate conscious beings (recursive)
- Must compress information universe contains
- Bostrom assumes possible to not care about details in unobserved areas
- Fill in details when someone looks
- Doesn't explain HOW this works
Technical Implementation Issues
- What algorithm can identify conscious subsystems and their intentions?
- How to quickly fill in required information without observable inconsistency?
- Can't generally throw away short-distance physical processes and get long distances right
- Climate models example: can't resolve distances below ~10km
- Non-linear system: short-scale information propagates to large scales
- Must replace missing physics with suitable approximations
- Climate scientists succeed because they have observations to check approximations
- Simulation programmer can't do this verification
Final Assessment
- Believers make big assumptions about what natural laws can be reproduced with computer simulations
- Don't explain how this supposed to work
- Finding alternative explanations matching all observations to high precision is extremely difficult
- Not serious scientific argument
- Doesn't mean it's wrong, but requires faith not logic
