3 ms·

I have mixed feelings when people fight so fiercily against the "trying all possibilities at the same time" explanation. It makes it sound false, when in fact i

by sasaf5 5y ago

I have mixed feelings when people fight so fiercily against the "trying all possibilities at the same time" explanation. It makes it sound false, when in fact it is only incomplete. "Trying all possibilities at the same time such as they all add up constructively to the right answer" would be a better way to fix it. From the article: The goal in devising an algorithm for a quantum computer is to choreograph a pattern of constructive and destructive interference so that for each wrong answer the contributions to its amplitude cancel each other out, whereas for the right answer the contributions reinforce each other.

latenightcoding 5y ago
Yeah, quantum parallelism is a resource we exploit. The next thing to explain is interference.
- btilly 5y ago
  I think that the next thing to explain is that what we can do has very limiting rules rules and they aren't what we are used to. Something like this: "In order to maintain a quantum superposition, we have to follow very strict rules imposed by quantum mechanics. We can't make operations depend on the state, so nothing like an if statement is available. Also all operations have to be reversible, so we don't have logical 'and' or 'or'. A relative handful of programs can follow these rules and get a huge speedup from quantum parallelism. The vast majority of useful computer programs can't be rewritten this way and get little or no speedup."
physicsAI 5y ago
I think Scott Aaronson is right to rail against the explanation of "trying all possibilities at the same time". The emphasis should be on *choreographing* the right interference, not on the fact that all possibilities get accounted for by a quantum computer (since each outcome has a measurable probability). In fact, "trying all possibilities at the same time" is a ubiquitous phenomenon in Nature. Light rays find their trajectory by tryin out all possible paths and only the one with the highest constructive interference survives (this is what Feynman discovered in Quantum Electrodynamics, which is referenced in the article). The real trick in quantum computing is manipulating qubits through quantum gates such that the outcome that gets assigned the highest probability is the correct one. This is indeed hard as the environment usually interferes and spreads out the probabilities across all outcomes, leading to a junky, random outcome.

by sasaf5 5y ago

latenightcoding 5y ago

Yeah, quantum parallelism is a resource we exploit. The next thing to explain is interference.

btilly 5y ago
I think that the next thing to explain is that what we can do has very limiting rules rules and they aren't what we are used to. Something like this: "In order to maintain a quantum superposition, we have to follow very strict rules imposed by quantum mechanics. We can't make operations depend on the state, so nothing like an if statement is available. Also all operations have to be reversible, so we don't have logical 'and' or 'or'. A relative handful of programs can follow these rules and get a huge speedup from quantum parallelism. The vast majority of useful computer programs can't be rewritten this way and get little or no speedup."

physicsAI 5y ago

I think Scott Aaronson is right to rail against the explanation of "trying all possibilities at the same time". The emphasis should be on *choreographing* the right interference, not on the fact that all possibilities get accounted for by a quantum computer (since each outcome has a measurable probability). In fact, "trying all possibilities at the same time" is a ubiquitous phenomenon in Nature. Light rays find their trajectory by tryin out all possible paths and only the one with the highest constructive interference survives (this is what Feynman discovered in Quantum Electrodynamics, which is referenced in the article). The real trick in quantum computing is manipulating qubits through quantum gates such that the outcome that gets assigned the highest probability is the correct one. This is indeed hard as the environment usually interferes and spreads out the probabilities across all outcomes, leading to a junky, random outcome.