QuestionsIs the world real?There is a one-in-billions chance that we are in base reality
Is the world real?Probably a simulation

There is a one-in-billions chance that we are in base reality

Musk's video-game argument; "one in billions"

Elon Musk born 1971Entrepreneur who made the simulation hypothesis famous with the general public in 2016.
See profile
What it claims

The odds that we are in base reality are one in billions: we are almost certainly in a simulation.

Why

Games will become indistinguishable from reality and run everywhere: simulated worlds would far outnumber the real one.

The same idea through history
Nick Bostrom2003Bostrom gives the rigorous version: three possibilities, one of which is that we are almost certainly simulated. He declines to choose between them.
Elon Musk2016In June 2016, Musk picks the third possibility and sums it up with the image of video games: "one in billions" odds of being in base reality.
In short

"There is a one-in-billions chance that we are in base reality" is an argument by Elon Musk (Elon Musk, Code Conference, 2016, 2016) on the question "Is the world real?" (answer: probably a simulation). Its claim: The odds that we are in base reality are one in billions: we are almost certainly in a simulation. According to AskPhilosophy's logic check (Z3 solver), the argument is valid only if one grants a premise the author leaves unstated (P4). AskPhilosophy presents 3 objections, 2 of them with a documented reply, and the original text behind each premise, checked word for word.

Elon Musk, Code Conference, 2016
Valid with an implicit premise

In June 2016, at the Code Conference, Elon Musk was asked whether we live in a simulation. He answered with what he calls the strongest argument for it. Forty years ago video games were two rectangles and a dot; today they are near-photorealistic 3D simulations played by millions of people, and they improve every year. Even if the rate slowed a great deal, in 10,000 years, a moment on the evolutionary scale, such games would be indistinguishable from reality, and billions of devices could run them. He concludes there is a one-in-billions chance that we are in base reality. A caveat: Musk takes up in popular form the reasoning of Nick Bostrom (2003); Bostrom's page gives the full version.

P1Forty years ago video games were just two rectangles and a dot. Today they are near-photorealistic 3D simulations played by millions of people at once, and they improve every year.
P2If games keep improving at any rate at all, even a thousand times slower than today, they will become indistinguishable from reality. Ten thousand years is nothing on the evolutionary scale.
P3These games could be played on any set-top box or PC, and there would probably be billions of such devices.
P4 · implicit premiseIf billions of simulated worlds, indistinguishable from reality, exist for a single base world, then an observer with no means of telling them apart is very probably in a simulated one.
Therefore the odds that we are in base reality are one in billions: we are almost certainly in a simulation.

click a premise to reject it

If you accept every premise, logic forces you to accept the conclusion. So the debate is about the premises.
Checked by a logic solver: the premises cannot all be true while the conclusion is false.
Without the implicit premise: Without P4, the computer finds a case where billions of simulations exist but their inhabitants have no experience: being "in a simulation" would not concern us.
Valid with P4. This hidden premise contains two things: that simulated beings would have experiences like ours, and that one should spread one's confidence equally among all possible observers. The objections target both, and also the assumption that humanity (or other civilizations) gets that far.
Analogy. A lottery where one winning ticket is mixed with billions of losing ones and you cannot read yours: you almost certainly hold a losing ticket. The "winning ticket" here is base reality. The image is ours.

Objections

3

Each objection is a conversation: click to see who replies and where the debate stands.

Nick Bostrom 2003 · attacks the conclusion

Bostrom, whose reasoning Musk takes up, does not share this confidence. He says he sees no obvious way to choose between his three possibilities (civilization dies out, or it never simulates, or we are simulated) and finds it sensible to spread one's credence roughly evenly among them.

Musk replies
see the argument
Elon Musk 2016 · Musk, in the same answer (2016)
Musk adds, right after, that we should almost hope it is a simulation: if civilization stops advancing, that may be due to a catastrophe that erases it. Otherwise, he says, we will either create simulations indistinguishable from reality or civilization ceases to exist.
The debate goes on (our summary) This passage does not say why Bostrom's second possibility (no civilization simulates) would be ruled out: it speaks of what to hope for, not of probability (our summary).
Where the debate stands (our summary)The disagreement is about how to spread one's credence. Musk picks Bostrom's third possibility; Bostrom defends the disjunction of the three. No decisive argument has separated the possibilities (our summary).
E
Ezra Klein 2016 · attacks premise 4

The journalist Ezra Klein, who reports the exchange for Vox, thinks Bostrom's three options are not exhaustive. He sees no reason to believe that even a very advanced civilization will easily manage to simulate consciousness.

a defender replies
Nick Bostrom defender of the argument · 2003
Bostrom assumes, and does not prove, that mental states can supervene on a broad class of physical substrates, not only in a biological brain. His argument depends on this "substrate-independence"; this text, written thirteen years earlier, does not answer Klein directly.
Where the debate stands (our summary)A disputed point: can a simulation produce consciousness? Supporters of simulation say yes on grounds of substrate-independence; critics ask for a reason to believe it. No proof settles it today (our summary).
David Chalmers 2003 · attacks the conclusion

Well before Musk, Chalmers argues that even if we were simulated, that would not ruin our knowledge. Living in a simulation is not a sceptical hypothesis but a hypothesis about the underlying nature of reality: our world would be real, made of computational processes.

no reply
see the argument
We have not yet found a direct reply by Musk in the texts.
Where the debate stands (our summary)Chalmers does not attack the probability Musk advances: he discusses its significance. One camp thinks "simulated" means "not really real"; the other that reality is the same whether or not a program lies behind it. No consensus (our summary).