3 ms·
> As I stressed there, if you’re in a biggest-number-naming contest, and you write “BB(10000),” you’ll destroy any opponent—however otherwise mathematically lit
by tzs 9y ago
> As I stressed there, if you’re in a biggest-number-naming contest, and you write “BB(10000),” you’ll destroy any opponent—however otherwise mathematically literate they are—who’s innocent of computability theory. For BB(n) grows faster than any computable sequence of integers: indeed, if it didn’t, then one could use that fact to solve the halting problem, contradicting Turing’s theorem.
I don't get this. You write "BB(10000)" and they write "BB(10000)+1". Why don't they win?
Yes, under the hypothesis they are innocent of computability theory. That just means they have no idea what "BB(10000)" means, but all they need to know for the biggest-number-naming-contest is that (1) it is a number, and (2) adding 1 to any number produces a bigger number.
I think that even little kids figure out #2 and will use it when confronted with a number they do not understand in such a contest. You might be able to challenge them on #1 and if they are young they might not realize that if their number is disqualified for not being a number then so is your number.
Or are we assuming you had the foresight to specify in the rules that players can only use numbers that they can personally explain, thus allowing you to use BB(10000) because of your familiarity with computability theory, but forbidding them from doing so because they only know it is something that you claim represents a big number?
- graycat 9y agoMy guess at an explanation -- have to be more clear on the "contest": (1) You don't know about the BB function. (2) We put our money in the pot. (3) Secretly you write down the digits of the largest number you can think of. (4) Secretly I see what BB(10,000) is and write down its digits. Uh, we are able to write quickly and have a lot of paper! (5) We show our numbers, and the person with the largest number wins.
- hawkice 9y agoThere are ways to determine the relative size of numbers even when you cannot write down the digits. It's how we know TREE(3) is larger than Ackermann(20) (or maybe larger than any description of simple recursive uses of Ackermann using only Ackermann(20) symbols -- so even [It's Ackermann sub 6 applied Ackermann sub 5 times to Ackermann sub 5 and so on, where Ackermann sub 1 is the Ackermann function] doesn't get close). Notice that claim is strong enough it doesn't matter what number you start with. 20, 3, a googolplex. It's all basically the same size.
- graycat 9y agoSure.
- schoen 9y agoAn equivalent response would be to say that "innocent of computability theory" here means "has no useful exposure to it at all (including how to use its notation in this contest)" rather than "doesn't understand its content".
- bhelkey 9y agoI think what the author is trying to get across is that BB(10000) would beat any large number generated through "conventional" means. From the article: > BB(5) ≥ 47,176,870 > BB(23) > Graham’s number[1] [1] https://en.wikipedia.org/wiki/Graham%27s_number https://en.wikipedia.org/wiki/Graham%27s_number
- remcob 9y agoHow about BB(BB(10000))? I would call that conventional means, well-defined notation, etc. And I can keep going in this direction, creating ever larger numbers. (In fact, if space permits, I could define this sequence and then take the BB(10000)th element of this sequence.)
- Sharlin 9y agoOf course you can go up the hierarchy, but BBⁿ(10000) is not "generated through conventional means" unless you take BB(10000) as given. And the point was that you're not given that.
- Scarblac 9y agoBasically, the game doesn't make sense when you take turns, instead contestants write down their answers and show them simultaneously. See https://www.scottaaronson.com/writings/bignumbers.html https://www.scottaaronson.com/writings/bignumbers.html .