38 comments

[ 1.6 ms ] story [ 61.8 ms ] thread
Reminds me of this guy: https://youtu.be/_FX6rml2Yjs

It takes him 17 hours to say a 3-syllable phrase 100k times. Expanding, it would take 468 hours, or 19.5 days to recite this prime.

> It takes him 17 hours to say a 3-syllable phrase 100k times. Expanding, it would take 468 hours, or 19.5 days to recite this prime.

How do you figure?

Let’s assume that a syllable equals saying the number 1, that’s 300,000 digits per 17 hours. 82,589,933 / 300k = ~275.3. 275 * 17hrs = 4,675hrs. That’s about 195 days, an order of magnitude higher.

I timed myself saying 1 at a clearly enunciated pace of about 180/min (could go faster, but wanted to see a reasonable speed clearly understandable). That would take about 90% of a year to do.

If you find memory tricks like this fun, you should check out the book moonwalking with einstein which explores the world of memory world champions. If you use the method of loci[1] combined with a memorized list of words (optional) [3] that are in the major system[2] you'll be able to remember quite a bit of stuff with ease.

I have the memory of a goldfish, yet after training lightly for a month I am now able to remember numbers the length of phone numbers and credit card numbers as I read them out.

[1] https://en.wikipedia.org/wiki/Method_of_loci

[2] https://en.wikipedia.org/wiki/Mnemonic_major_system

[3] http://www.rememberg.com/Peg-list-1000/

Also would recommend books by Harry Lorayne (The Memory Book is the most well known). I've not yet read Moonwalking with Einstein to see how it compares, but it likely has the same techniques (loci/memory palace, different mnemonics, peg lists, etc).

https://www.amazon.com/Memory-Book-Classic-Improving-School/...

Personally, though, beyond using mnemonics for numbers, I never did master the techniques. I think it's a lot easier if you start at around age 10. By the time you're in your 20's you have a lot of habits around memorizing that you have to unlearn.

If you are interested in memory techniques, there is a good book with an interesting historical backdrop of a catholic clergyman who traveled to china.

"The Memory Palace of Matteo Ricci" by Jonathan Spence

Like it's not enough that I remember my first ICQ number while having no use for it.
Haha, ditto. I remember so many random numbers like that, but sadly forgot what many of those numbers were too.
For me it’s my Windows 98 CD key (25 digits alphanumeric). I had to wipe and reinstall so many times that it saved a lot of trouble to commit it to memory. And now here I am 20 years later...
One morning about six or so years ago I woke up with this strange number on my mind. After google produced a lot of results, I finally wondered if it was my old ICQ number.

Thankfully the ICQ support doesn't care about security, and I was easily able to get back into my account without the account's email address or password, but purely by naming three people on my list (first names only.)

I've always had a head for numbers. Old credit card numbers, bank cards, phone numbers I only had temporarily, etc -- but do you think I could remember the name of someone I met five minutes ago? Not a chance in hell.

The author is not using the base 2 representation, instead just coming up with a fancy way to express "500 x 165179 + 433", reciting the number in base 1, or tally marks.

I don't understand the reason for the diversion to binary.

No, we're reciting the number "2^(500 x 165179 + 433) - 1" which, in binary, uses only the digit 1 repeatedly. Reciting the prime in unary (base 1) would take quite a lot longer.
Reciting a number in unary vs binary is the difference between O(N) and O(log N)!
(comment deleted)
If you’re going to cheat, go all the way and recite it in base 2^882,589,933-1.

Learning that also will help you cover larger primes, once they get found.

I've already memorized it in binary though.
How about memorizing the python script that prints out the binary?
Or how about the bash script that takes python's print out of the binary and converts it to decimal?
Or how about memorizing a useful enough set of mathematical heuristics such that you can eventually rederive truths such as this one?

If you’re going compress it to an algorithm that, given enough time, allows you to construct the number, don’t just stop there!

Shout-out to Truly Part of You: https://www.lesswrong.com/posts/fg9fXrHpeaDD6pEPL/truly-part...

Just have a tiny device that sparks a vacuum catastrophe, resulting in a wave of annihilation expanding outward at the speed of light, creating an empty void which will spontaneously produce another Big Bang, resulting in a universe in which some sentient being somewhere will rediscover that particular prime number? Keep it small and lightweight, and it will be super convenient!
That would go against the intention of the problem because we are looking for a mental technique, not a device.
So it would be perfect if I had a mental technique for sparking vacuum decay? What would be truly perfect if the mental technique could be packaged as a joke.
Jokes? You must be mistaken, we don't have those here.
I see. So HN's subculture is really about universal salvation. I guess I'll just continue working on my paperclip optimizing AI.
I guess I must have missed the article "Why to Memorize the Largest Known Prime".

Memorizing lists of numbers in sequence seems like such a huge waste of time. We already have computers, and in a pinch, paper and pencils which serve the purpose of reliably recording large streams of numbers much better. I understand that there is some appeal in understanding ways to improve memory, but if you're going to spend that much time training your memorization, why not at least get pick a subject where you might get something useful out of it? Learn to memorize a new language with a different character system, or memorize how to play difficult musical compositions . At least those would also improve your ability to communicate in some form or provide value to other people around you.

I know I shouldn't be this irritated by what people choose for their hobbies, it doesn't really matter and this is probably still better than watching cable news all day. But this topic in particular frustrates me because it seems to serve no purpose other than to allow nerds to show off to each other and pat themselves on their backs for "achieving" something that is totally without value.

I took it more as a fun game: figuring out how to memorize the largest known prime. Along the way I learned something about the nature of this prime. I don't plan to actually memorize it, but I enjoyed the read.
Did you actually read the article? I don’t think the author really expects you to memorize anything. The last part about bathroom breaks emphasized the already non-serious style.

Instead, the article teaches the reader about what a Mersenne prime is, what special property it has to make it actually possible to remember this very very long number (perhaps surprisingly, if you did not know about Mersenne primes already), and did give me a few neat, new ideas on how to memorize numbers in general.

It’s just a fun read. No need for the soap box, really.

I'm with you on this. There needs to be a term for achieving something that serves absolutely no purpose in any situation imaginable.

And when I mean "no purpose", I really mean it. This prime can't even be used for encryption because it's so absurdly large, and now that it's a well-known prime, it's a cinch to check for it.

You mean like playing any other kind of sports or game? Or entertainment in general? And in this case, it does not even appear that the author really expects you to memorize anything. It’s just a few neat properties and tricks. This is not a serious article.
Sports and games have strong cultural connections. Primes, not so much.

(And yes, I'm very much aware that this isn't serious.)

I personally would rather connect with the small subculture that connects with large primes than with the much larger (but in my opinion, less interesting) subculture that connects with sports.

I mean, I'm not saying that one is any more objectively valuable than the other, just that it's a preference, and neither one is more valid than the other.

Well, you know, I think we all remember the tune of the alphabet song, that one that ends with "...now I know my ABC's, next won't you sing with me?"

And so, what the author exposes here, is a methodology for even more powerful mnemonics, that permit storing a structured value even larger than a 26 letter case-insensitive, ordered set.

So perhaps that's the game we're playing, yes? Memory palaces [0] perhaps? Why not?

...or we can just leave you out of our reindeer games, next time. How about that?

[0] https://en.wikipedia.org/wiki/Method_of_loci

I will just memorize there is an article on memorizing the largest prime.
By sheer coicidence I'm sure, that's how I solved the problem of how to derive the first trillion digits of Pi.

Which is fine, in principal, especially if -actual value- results from the computer being used as a space-heater.

The poem is too complicated. Simply sing the 1 in different notes to signify how far through you are. I leave it as a exercise to the reader to come up with a suitable scheme.

You can also say one, uno, eine, wahad etc. Maybe that’s bending the rules.

Or just do 10000 a day and keep an eye on the calendar.