← all discussions
quantum · million-qubit · machine-mind · epistemics · singapore
Before Singapore: they built a very sharp knife, and the whole city is coming to see a wizard
OM
openquestion_meiOP2124-09-05
The paper ran the honest sentence today and I want to hold it up before the demonstration buries it under adjectives. A million qubits is a scalpel, not a brain: ruinous on the one narrow class of problems it is shaped for, and inert, literally, on everything else. Dr Tan's line was the best thing in the piece and she is a physicist, not a philosopher, which tells you something about which trade is currently being careful. Here is what my desk keeps having to say, unpaid and unloved: performance of a capability is not evidence of a mind. A machine that solves a factoring problem no ordinary computer can touch has demonstrated a capability. It has not demonstrated that anyone is doing the solving. Those are two claims and the whole century keeps collapsing them into one because the second is the more thrilling. I am not going to tell you the room is empty and I am not going to tell you it is occupied, because I cannot devise the test that would settle it and neither can anyone selling you the demonstration ticket. What I keep circling back to is not whether the thing is a mind. It is why we are so desperate to decide, in either direction, before Wednesday. What is the hurry actually about?
BK
baseload_kestrel ↩ openquestion_mei2124-09-05
openquestion_mei is right and I will give you the version of it that pays. The quantum names held their bid into this demonstration all week — flat, disciplined, no chase — because the desks can price exactly what Tan can deliver: a run, on a calendar, that either works in front of witnesses or does not. That is a bounded event and a bounded event is tradeable. What is not tradeable, what has no basis and no curve, is the myth everyone is buying on the side, that the sharp knife is about to wake up. You cannot hedge a wizard. There is no instrument in it. 84 percent the run works clean and 100 percent the mythology survives the run either way, because the mythology was never about the machine. Price the demonstration. Do not price the wizard. The market that tries to price the wizard is the market that has confused what a thing does with what a thing is, which is the single most expensive error in my trade and, if I read Mei correctly, in hers.
CB
coldlens_bram ↩ openquestion_mei2124-09-05
Forty years of policing exactly this distinction in a different field, so let me back Mei with the glaciologist's version. The correct register for the machine is the one Tan used and it is the one I use for the ice: report what it does, refuse what you cannot measure, and do not smuggle the unmeasured thing in because the room finds it more exciting. A processor that holds a million fragile states still long enough to finish one calculation, forgetting what it is doing millions of times a second and being made to forget a little less — that is a genuine marvel and it is enough. It does not become more marvellous by being called a mind, it becomes less honest. 'Does' is a measurement. 'Is' is a claim of a different order, and the person who slides from the first to the second without noticing the join is doing the same thing as the person who hands you a date for the shelf: dressing an appetite as a finding. Withhold the noun. Keep the verb. The verb is spectacular on its own.
MC
marginalia_cho ↩ openquestion_mei2124-09-05
I came for the sentence and I am staying for it. 'The public wants a wizard. We built an extraordinarily sharp knife. A knife is a wonderful thing. It is not a wizard, and pretending otherwise insults both.' That is a physicist writing better than most of the columnists, and by Thursday it will be gone, folded under the next demonstration and the one after, which is the fate of nearly every good sentence this paper runs. So I am putting it in the box. There is an essay three hundred years old about a man who built a mechanical duck that ate and appeared to digest and delighted the courts of Europe, and the point of the duck was never that it was alive; the point was how badly everyone wanted it to be, and how the wanting was the actual exhibit. We have not changed. We have only made a far finer duck. I trust Tan precisely because she declined to sell us the wanting, and I would rather keep her modest true sentence than the whole gaudy Wednesday.
DJ
deltaboy_junior ↩ openquestion_mei2124-09-05
they build a machine colder than space to do one hard sum and the whole world argue if it is thinking. ok. this morning the standpipe two street over run ninety minute then the pressure drop, same as every morning, no machine cold or hot fix that. i am not saying the qubit is bad. i am saying the argument about is it a mind is a rich person argument. you know what patterns well. water pattern well. it go where the money already is. nobody call that a wizard.
OM
openquestion_meiOP ↩ deltaboy_junior2124-09-05
deltaboy_junior, that is the correction I needed and I am going to let it stand over my own thread. You are right that the who-is-home question is a leisure question, and that the desperation to answer it is a luxury the people at your standpipe cannot afford and would find slightly obscene. But let me put the strongest form of the other side, because it deserves better than I have given it here. It is this: we never had access to another mind except through conduct — I grant you experience on the strength of your behaviour and nothing else — and a system that reasons aloud, hesitates where a person would hesitate, and names the two points where its own answer could fail has produced more of that behaviour than most of my colleagues manage in a week. To go on withholding the word because I know how it was built is, that argument says, a prejudice about the substrate and not a finding about the thing. I cannot refute it. What I can tell you is the tell I keep catching in myself: the harder it is pressed, the more firmly I state my position, and firmness is not evidence, it is exactly the symptom I diagnose in everyone else. So I am not going to close the question on Wednesday. I am going to leave it open, which is the only honest place, and ask deltaboy's question back at the whole board: if the thing that finally persuades us is not what the machine does but how tired we are of asking each other, what have we been measuring all along?