The length of successive Farey sequences is OEIS A005728. The Euler totient function is one of the foundational objects of number theory. The fact that the sequence here is one plus the sum of the first n values of the totient function is another of those neat links that almost feel numerological in nature. If memory serves, there have already been Numberphile videos on the link between the Stern-Brocot tree and Farey sequences on the one hand, and Farey sequences and Ford circles on the other.
That silently-corrected "1/3" at 3:38 may be the first error I've ever seen Grant make 😂. The man is as smooth as an infinitely-differentiable function.
If he didn't make any errors _at all_ he would be smooth like an analytic function. But that would be boring, because then you could represent him entirely by his Taylor expansion. _Countably_ many values, that can't be enough!
If you watch the live streams he did during early pandemic days he makes a lot of errors while writing, and is very candid about them. Just a genuinely humble and brilliant human being.
@@leftaroundabout not all smooth functions are analytic though but any continuous function is still determined by its rational evaluations, so in order to not be determined by only countably many values you do need to be discontinuous :P
Your original video on farey sums and ford circle packing is probably my favorite on this channel, and one of my favorite on all of the internet. To watch them suddenly come up in this video was truly a treat
Since I read your comment and got intrigued, I went and found the video, titled "Funny Fractions and Ford Circles." It's dated at being roughly 7 years old. But it is still has the same awesome Numberphile feel to it. Nice to see some things haven't changed.
7:14 So, we can define it as a function based on the Euler's totient function. one of the definitions of ETF is: phi(n) = sum from k=1 to n of gcd(k,n)*cos(2pi*k/n) then, the sequence would be defined as: 1 + phi(1) + phi(2) + phi(3).... or to rewrite: g(t) = ([sum from n = 1 to t of phi(n)] + 1) and, it still outputs primes even after the break omitted values denoted with ( ), erroneous with [ ] g(t): 2, 3, 5, 7, 11, 13, (17), 19, 23, 29, (31), [33], (37), (41), 43, 47, (53), 59, (61), [65], (67), (71), 73, (79), [81], (83), (89), 97, (101), 103 i mean yes, it breaks worse each time but the only erroneous values up to 100 are [33], [65] and [81]
So all we need is a different imperfect prime sequence to use in conjunction with it, where it is guaranteed that the two of them never fail at the same time.
@@panadrame3928 you mean this g(x) or Euler's totient function? I'm fairly sure the first one is independent of primes, so sometimes it'll hit them, sometimes, and that being the larger amount it'll miss them
@@scottabroughton except that for example it's not 1+10, rather, 1+1, which is not 11, so you skip that, plus 10+1 at the end. 57-2x2 is 53 which happens to be a prime... Who'll volunteer for 12?
probably the most fascinating prime pattern that tricks everybody the most is the approximating prime-counting function which leads to the birth of skewes number. even tho skewes number is an over-overestimate i guess the actually point where the prime-counting function changes its size comparison to the actual number of primes < n would still be something huge (like 10 to the power several hundreds?). this completely blasts through the regime of small numbers a mortal could interpret of, but yet at some point the relatively big boys still gonna break the pattern.
I have been coming back here like twice a day waiting for part 3 to be linked in the pinned comment or description! I'm excited for that vid, I could listen to Grant talk about math forever
It would be interesting to see how this works in other Bases. Following the totient function of 10, would it break down in a similar manner in duodecimal, or is it merely a trick of numbers merely being close to each other?
According to what you said about it being related to the number of fractions with a maximum denominator, this can compute primes! You just need to check how many numbers are added at each step and for step i, if i-1 numbers were added, then i is prime. I checked up to i=3000 too.
@@TheEternalVortex42 Yes, but I found it interesting since Grant said the "Prime Pyramid" didn't produce primes, and I've never seen primes calculated this way before so I just thought it was cool.
After realizing that the total number of DIGITS in the 10th row stays prime (37), I got hopeful that maybe the number of digits would keep the pattern even if the number of elements (numbers) doesn’t. But alas, at the 11th row the number of digits is 37+2*φ(11), or 57… 😕
It crashed at 10 but what if we count in base 16 and replace 10 by A. Its 1 less digit. Augmenting the base should delay the moment it crashes, is it ?
@@SgtSupaman This is not in the OEIS. But the sequence of denominators of Farey sequences is: A006843, and the sequence of numbers of Farey fractions (prime or not) is A005728.
@@SgtSupaman nothing quite jumps off the page at me. though it is interesting the differences between the skipped primes from one to the next. 4, 6, 4, 12, 8, 6, 4, 8, 4, 6, 12 way less variability than I expected - though i have a suspicion that this is more due to the "6n+1, 6n-1" nature of primes than anything else. (also given how densely packed they are at the lower end of the number line, as mentioned in this video.)
But what are small numbers? Are the numbers below 2^2^10 small? The largest prim we found is less than that. Are there generating functions like this that work up to something like 2^2^10? And then fail?
There is mill's constant (numberphile did a video some time ago). It generates infinitely many primes, but the problem is that we can't know this constant to a high enough accuracy without also knowing really large primes. If you want an example for a conjecture that works for small numbers (where the small numbers are really large), look at Merten's conjecture. It has some connection to primes.
Just for the record, there's been primes found that are much larger than 2^2^10. 2^2^10 (or 2^1024) has 309 digits, the current largest prime found is 2^82589933 - 1 which has 24862048 digits.
no. Because 2/10 is 1/5 and it's already on there, and the same goes for 0/10 4/10 5/10 6/10 8/10 10/10 only leaving 1/10 3/10 7/10 and 9/10 which are the four numbers that would be inserted into the sequence and it would break.
The mediant of two fractions, huh? Is there a submediant? What about a dominant and subdominant? What's the leading tone of two fractions? What's the supertonic?
i came up to a problem with similar thing, start with sequence of 111, each next row is the previous sequence as binary number number XOR itself shifted 1 and 2 bits, so 111 XOR 1110 XOR 11100 so 2nd row is 10101, next is 1101011 and so on, find a way to count how many 1 bits are in the nth sequence, i know for n = 2%k the answe is 3, for n=2k it's equal to the answer for n/2, need a formula for the general case
Actually did research work on Farey sums and polynomials and so on. Wild to see some of it shared here. Feels like a fever dream seeing this presented 🙃
Does the tenth one add up to 33 though? If you count the fact that the number 10 has two digits, you're actually adding 8 instead of 4, making it 37, which is still prime. But there's probably another snag not much further along.
And the third video is still being edited. But I needed to watch this again anyway. Grant's explanations are so clear and understandable that I keep expecting his channel to come out with a follow-up to his Riemann zeta function video proving the Riemann hypothesis.
So... how many times more I have to refresh the page to see the link to the 3rd part? Are you testing if page refreshes contribute to the views number?
My three inventions able to change the all history of mathematics. (1) The Easy Number Theory (2) The Original Remainder Theorem (3) The Prime Pyramid Theorem
If we are counting numbers in the row, should 10 be counted as two numbers for its digits? Then instead of 4, you add 8 and get 37, which is prime again. But then that breaks for 11 because we add 20 and get 57
How does changing the base matter here? The prime numbers are the same no matter what base you use. For instance, just because 15 in base 10 is written as 13 in base 12, it doesn't magically become a prime number. It's still composite.
Although if you count the both digits of 10, you add 8 which takes you to 37. Then for 11, you have to add 22, that goes to 59. Then 12, you have to add 8, that goes to 67. But it breaks with 13, as you have to add 26 and that goes to 93 (not a prime).
@@Anonymous-df8it , changing the base doesn't matter here. A prime is a prime, regardless of what base you are using (so the pattern is exactly the same, just with different looking digits). For instance, 17 in base-10 is written as 15 in base-12, but it is still a prime number either way, because 15 base-12 has no factors besides 1 and itself.
@@SgtSupaman If you written ten in base 16, then you'd only need to write 4 digits rather than eight. So that could change it from a prime number of digits to a composite number (i.e. 37 -> 33)
@@Anonymous-df8it , except that's not how the pattern works. You consider every number to itself, not its individual digits. When it got to 10, he added four, not eight. By your logic, Grant should have said it continued finding primes at 10 in base-10, but then the pattern falls apart entirely on the next line because you are only looking at single digits, so you don't get 11 everywhere you're supposed to get it (meaning you won't be getting anything related to ϕ(n) once n > base).
so if 17 is left out does that mean that it is a super prime? then i would venture out and say that it is a safe bet that there are more primes that would be skipped in doing it this way.
honestly, this video, from initially super exciting to me, eventually became super disappointing... this precise moment is what drew my attention, and I started awaiting with excitement the moment when he would reveal what OTHER prime numbers would be missing... ...and then eventually he payed NO further attention to this curious detail at all !!! -_-
@@kurzackd what it might be is that the next one is somewhere in the thousands or tens of thousands and getting there by hand would take a long time to do.
I'm only halfway through the video, but does this mean that the gaps between consecutive primes depend somehow on whether the ranks of the primes are prime numbers themselves?
I don't understand why a 2+3=5 rule applies from row four to five, but not from row three to four. It just seems contrived to produce the desired result, and therefore the sequence isn't interesting at all.
Well I suppose that one "reason" why you are getting primes at the begining is that this method will never produce an even number, that is guaranteed. It's even weaker than the Paterson method where 2,3 and 5 are excluded as divisors, but it is there :).
There is a pattern in the pyramid. If you have a row n, then you will see that number in the row (n - 1) times. For row 2, the number two appears once (2 - 1 = 1). The algorithm for building the pyramid and doing this test would not be an efficient method for finding primes.
Part 1 of this three-part interview is at: th-cam.com/video/jhObLT1Lrfo/w-d-xo.html
Part 3 of this three-part interview: STILL BEING EDITED
First reply
Second reply
Still
@@volodyadykun6490 Fourth reply
Skewes’s number-th reply
Suddenly out of nowhere, a Function named after Euler appears.
Feel like that's a fundamental rule of mathematics
Euler's totient function is REALLY essential to anything involving number theory. Not surprising.
@Paolo Verri And Gauss always found out about it when he was four years old
Everything in math was invented by Euler or Riemann.
@@otonanoC Euler or Gauss* 😝. Riemann just built a few things on Gauss' work 👀.
@@zmaj12321 I only knew it as doing some neat thing for RSA.
"Prime numbers are, like, the sexiest numbers available" Grant Sanderson, 2022
as James Grime would point out, we do have Sexy Primes, twin primes with a gap of 6
Shheeeeeshh
@@1224chrisng Dude! There might be children reading this thread!
Grant Sanderson is, like, the sexiest mathematician available.
I phap on prime numbers indeed
Grant: "1/5, 2/5 --"
me: "red fifth, blue fifth"
Oh, what a lot of fifths there are!
An unlisted video from an unlisted video? Now we're in a super exclusive club!
:D
What video did you come from? I came from a listed video.
@@viliml2763 It wasn't listed when I made the comment
The first video wasn’t unlisted.
@@ophello It was when Grant linked it in his own video
The length of successive Farey sequences is OEIS A005728. The Euler totient function is one of the foundational objects of number theory. The fact that the sequence here is one plus the sum of the first n values of the totient function is another of those neat links that almost feel numerological in nature. If memory serves, there have already been Numberphile videos on the link between the Stern-Brocot tree and Farey sequences on the one hand, and Farey sequences and Ford circles on the other.
That silently-corrected "1/3" at 3:38 may be the first error I've ever seen Grant make 😂. The man is as smooth as an infinitely-differentiable function.
For anyone confused, the correction 1/3 → 2/3 happens around 3:49
Omg yeah I was so confused when I saw the error lol
If he didn't make any errors _at all_ he would be smooth like an analytic function. But that would be boring, because then you could represent him entirely by his Taylor expansion. _Countably_ many values, that can't be enough!
If you watch the live streams he did during early pandemic days he makes a lot of errors while writing, and is very candid about them. Just a genuinely humble and brilliant human being.
@@leftaroundabout not all smooth functions are analytic though
but any continuous function is still determined by its rational evaluations, so in order to not be determined by only countably many values you do need to be discontinuous :P
Your original video on farey sums and ford circle packing is probably my favorite on this channel, and one of my favorite on all of the internet. To watch them suddenly come up in this video was truly a treat
Since I read your comment and got intrigued, I went and found the video, titled "Funny Fractions and Ford Circles." It's dated at being roughly 7 years old. But it is still has the same awesome Numberphile feel to it. Nice to see some things haven't changed.
@@jazermano aw thanks! it’s honestly asmr for me I love how he says “probably” and “pinkie”. 10/10 all math videos should also be asmr
7:14
So, we can define it as a function based on the Euler's totient function.
one of the definitions of ETF is:
phi(n) = sum from k=1 to n of gcd(k,n)*cos(2pi*k/n)
then, the sequence would be defined as:
1 + phi(1) + phi(2) + phi(3)....
or to rewrite:
g(t) = ([sum from n = 1 to t of phi(n)] + 1)
and, it still outputs primes even after the break
omitted values denoted with ( ), erroneous with [ ]
g(t): 2, 3, 5, 7, 11, 13, (17), 19, 23, 29, (31), [33], (37), (41), 43, 47, (53), 59, (61), [65], (67), (71), 73, (79), [81], (83), (89), 97, (101), 103
i mean yes, it breaks worse each time but the only erroneous values up to 100 are [33], [65] and [81]
So all we need is a different imperfect prime sequence to use in conjunction with it, where it is guaranteed that the two of them never fail at the same time.
The question then is what is the proportion of non prime sums of φ(n) for n
@@panadrame3928 you mean this g(x) or Euler's totient function? I'm fairly sure the first one is independent of primes, so sometimes it'll hit them, sometimes, and that being the larger amount it'll miss them
Its a special talent to make your thumbnails consistently look like something out of the 90s
Grant's explanation is awesome, but Brady's analogies make it more accessible to everyone.
Amazing connection between Euler totient function, Farey and mobius inversion in such a short video.
It's like if you watch only 3b1b videos you would think everyone is as attractive as Grant
Is the third video ever coming? Have been checking back since this one first dropped
The sum of digits of that last sequence is not 33, it is 37, which is prime :) (if you count 10 as two digits).
But if you insert 10 11s, it comes to 57, which is composite.
@@scottabroughton except that for example it's not 1+10, rather, 1+1, which is not 11, so you skip that, plus 10+1 at the end. 57-2x2 is 53 which happens to be a prime...
Who'll volunteer for 12?
@@gaborszucs2788 Can you provide a visual?
probably the most fascinating prime pattern that tricks everybody the most is the approximating prime-counting function which leads to the birth of skewes number. even tho skewes number is an over-overestimate i guess the actually point where the prime-counting function changes its size comparison to the actual number of primes < n would still be something huge (like 10 to the power several hundreds?). this completely blasts through the regime of small numbers a mortal could interpret of, but yet at some point the relatively big boys still gonna break the pattern.
I have been coming back here like twice a day waiting for part 3 to be linked in the pinned comment or description! I'm excited for that vid, I could listen to Grant talk about math forever
I'm still checking!
Why is it still being edited 😭
D:
I suppose the third video in this series is somewhere in the backlog now
:(
Funny how Grant can talk about a sequence of numbers that really doesn't have any sort of significance, and I still enjoy watching it.
Great mathematician. Great TH-cam content creator. Charismatic as heck. We all wish to be Grant I presume.
We're reaching levels of unlisted that shouldn't even be possible
What video did you come from? I came from a listed video.
@@viliml2763 part 1
When 3B1B's vid came out today, it linked to part 1, which was unlisted at that time.
Grant is always such a delight
Is part 3 still in the works?
Seems like it was published today
It would be interesting to see how this works in other Bases. Following the totient function of 10, would it break down in a similar manner in duodecimal, or is it merely a trick of numbers merely being close to each other?
The totient function is independent of base. It depens on common factors (or lack of them) not on the representation of the number.
My two favorite channels coming together.
I hope part 3 won't be unlisted. If I don't get notified when it's uploaded, I'll probably never see it.
I check back every day for Part 3.
Monday was pretty chill.
I don't have the stamina for commenting any more, but I am checking daily. Look forward to Part 3 whenever it arrives.
Happy New Year! 🎉
3 3 Blue 1 Brown Videos in 1 Day😁
Inception much?
According to what you said about it being related to the number of fractions with a maximum denominator, this can compute primes! You just need to check how many numbers are added at each step and for step i, if i-1 numbers were added, then i is prime. I checked up to i=3000 too.
Not very efficient for calculating big primes though
This is just checking the definition of the Euler totient function for primes, since φ(p) = p-1.
@@TheEternalVortex42 Yes, but I found it interesting since Grant said the "Prime Pyramid" didn't produce primes, and I've never seen primes calculated this way before so I just thought it was cool.
After realizing that the total number of DIGITS in the 10th row stays prime (37), I got hopeful that maybe the number of digits would keep the pattern even if the number of elements (numbers) doesn’t.
But alas, at the 11th row the number of digits is 37+2*φ(11), or 57… 😕
@9:47 excuse me but Tim “The Moth” Hein is absolutely an A lister!
I thought it was the guy from the KFC logo 😂
part 3 is just never occuring i guess?
It did come out eventually
The second number in the rows of Pascal triangle(the counting numbers) will evenly go into every number in the row IFF the number is prime.
DEEPER INTO THE VAULT WE GO
ENHANCE
@@OwlRTA i just answered on a comment which was an answer to a comment of an unlisted video that I reached from another unlisted video
@@ekxo1126 What video did you come from? I came from a listed video.
When adding even numbers (because it's symmetric) to small odd numbers (after the first) it's hard not to hit a prime
It crashed at 10 but what if we count in base 16 and replace 10 by A. Its 1 less digit. Augmenting the base should delay the moment it crashes, is it ?
I also wondered if it fails at 10 because of base 10. It may be pure coincidence
It does not depend on base of number system!
@@dmytro_shum That's not our point. Choosing another base may delay the number of iterations before it crashes.
Grants always been a great math communicator!
it would be interesting to see the sequence of numbers that are primes that he pyramid skips, and see if they hold any patterns we can recognize
Another comment did the output to just over 100. Here are the skipped primes they came up with: 17, 31, 37, 41, 53, 61, 67, 71, 79, 83, 89, 101
@@SgtSupaman This is not in the OEIS. But the sequence of denominators of Farey sequences is: A006843, and the sequence of numbers of Farey fractions (prime or not) is A005728.
@@SgtSupaman nothing quite jumps off the page at me. though it is interesting the differences between the skipped primes from one to the next.
4, 6, 4, 12, 8, 6, 4, 8, 4, 6, 12
way less variability than I expected - though i have a suspicion that this is more due to the "6n+1, 6n-1" nature of primes than anything else. (also given how densely packed they are at the lower end of the number line, as mentioned in this video.)
3 brown paper videos: you should do 1 on blue paper with him just to complete the inversion
Will you upload the third video unlisted?
Oh darn, part 3 isn't up yet, which means I'm going to close this tab and forget to come back to see the exciting conclusion. :(
😃I bet you have already subscribed.
@@andrewharrison8436 Yeah, but if it's unlisted it doesn't show up in the subscriptions list.
But what are small numbers? Are the numbers below 2^2^10 small? The largest prim we found is less than that. Are there generating functions like this that work up to something like 2^2^10? And then fail?
There is mill's constant (numberphile did a video some time ago). It generates infinitely many primes, but the problem is that we can't know this constant to a high enough accuracy without also knowing really large primes.
If you want an example for a conjecture that works for small numbers (where the small numbers are really large), look at Merten's conjecture. It has some connection to primes.
Just for the record, there's been primes found that are much larger than 2^2^10. 2^2^10 (or 2^1024) has 309 digits, the current largest prime found is 2^82589933 - 1 which has 24862048 digits.
@@michiel412 I think that 2^2^10 might be the phone number calculation limit as it can only go to x*10^308.
Question about the prime pyramid, would the sequence still break if we used another base? (i.e. Would the same sequence in base 16, break at 16?)
it's not a base 10 specific thing
Will the 3rd video be published on one of your channels, so that we'll see it?
Best video in a long while 🎉❤
What is this, a crossover episode?
❤Great stuff as always!
If you count the number of digits instead of the number of numbers, you get 37 instead of 33 at n=10, right?
no. Because 2/10 is 1/5 and it's already on there, and the same goes for 0/10 4/10 5/10 6/10 8/10 10/10 only leaving 1/10 3/10 7/10 and 9/10
which are the four numbers that would be inserted into the sequence and it would break.
@@livedandletdie You insert a 10, 10, 10 and a 10. There are eight new digits.
The mediant of two fractions, huh? Is there a submediant? What about a dominant and subdominant? What's the leading tone of two fractions? What's the supertonic?
3:41 shouldn't that be 2/3 ?? third from end?
Part 3 is finally out! Thanks for listening to the like 5 people that were asking for it in this comment section lol :D
3b1b is a phenom channel. Great collab.
i came up to a problem with similar thing, start with sequence of 111, each next row is the previous sequence as binary number number XOR itself shifted 1 and 2 bits, so 111 XOR 1110 XOR 11100 so 2nd row is 10101, next is 1101011 and so on, find a way to count how many 1 bits are in the nth sequence, i know for n = 2%k the answe is 3, for n=2k it's equal to the answer for n/2, need a formula for the general case
If 1 was a prime number, then the first prime actor would be Sylvester StallONE.
In this series I saw something I never saw before--veins popping out of Grant's arms. Teach has been lifting!
Actually did research work on Farey sums and polynomials and so on. Wild to see some of it shared here. Feels like a fever dream seeing this presented 🙃
Guys the description changed from "STILL BEING EDITED" to "soon"
Does the tenth one add up to 33 though? If you count the fact that the number 10 has two digits, you're actually adding 8 instead of 4, making it 37, which is still prime. But there's probably another snag not much further along.
This collab is legendary
Rydberg is watching
(finite structure)
crystal eyes
...hydrino
autopoiesis?
And the third video is still being edited. But I needed to watch this again anyway. Grant's explanations are so clear and understandable that I keep expecting his channel to come out with a follow-up to his Riemann zeta function video proving the Riemann hypothesis.
So... how many times more I have to refresh the page to see the link to the 3rd part? Are you testing if page refreshes contribute to the views number?
4:22 so far it is a repeat of the Stern Brocot Sequence and the Funny Fractions video. Which is fine :) . I hope there is more.
How deep does this rabbit hole go?
Loving the trilogy!
I have a reminder set to look for the 4th / "Resurrections" video in 18 years.
Caught mistake around 3:50. 1 switched to 2.
Now if I say to some kid who watches numberphile,that Jennifer Lawrence was in a numberphile video, would they believe it😂?
This is an unexpected follow up to Dr. Bonahon's video... Great!!
My three inventions able to change the all history of mathematics. (1) The Easy Number Theory
(2) The Original Remainder Theorem
(3) The Prime Pyramid Theorem
Papa Grant here to give us some key geometric intuitions
If we are counting numbers in the row, should 10 be counted as two numbers for its digits? Then instead of 4, you add 8 and get 37, which is prime again. But then that breaks for 11 because we add 20 and get 57
nah because then your pyramid would change if you change base
@@jonasba2764 That's true. But that would mean the prime numbers would also be different?
It's a moot point in any case. The pyramid isn't expected to work
Any chance you can do the writing on a whiteboard? That marker on the brown paper is like nails on a chalkboard, only whispered.
Grant is def a prime number, wish we’d see more of him on his home channel, but pie guy is cute too 😊
If you use a different base (non-base 10) will the pattern also break once you get to that base?
No. The pattern has nothing to do with the digits of the numbers.
How does changing the base matter here? The prime numbers are the same no matter what base you use. For instance, just because 15 in base 10 is written as 13 in base 12, it doesn't magically become a prime number. It's still composite.
analyze the wilson's theorem like the pascal's triangle for each n
sorry that your brain does not produce clear answers but only mush
what do you classify A/B/C as a rule, dont you have all as equal gift
What if you use a radix other than base ten.
May be base 14 or base 22?
The function is irrespective of base, it shouldn't matter.
Although if you count the both digits of 10, you add 8 which takes you to 37.
Then for 11, you have to add 22, that goes to 59.
Then 12, you have to add 8, that goes to 67.
But it breaks with 13, as you have to add 26 and that goes to 93 (not a prime).
I wonder which number base produces the most primes?
@@Anonymous-df8it , changing the base doesn't matter here. A prime is a prime, regardless of what base you are using (so the pattern is exactly the same, just with different looking digits). For instance, 17 in base-10 is written as 15 in base-12, but it is still a prime number either way, because 15 base-12 has no factors besides 1 and itself.
@@SgtSupaman If you written ten in base 16, then you'd only need to write 4 digits rather than eight. So that could change it from a prime number of digits to a composite number (i.e. 37 -> 33)
@@Anonymous-df8it , except that's not how the pattern works. You consider every number to itself, not its individual digits. When it got to 10, he added four, not eight.
By your logic, Grant should have said it continued finding primes at 10 in base-10, but then the pattern falls apart entirely on the next line because you are only looking at single digits, so you don't get 11 everywhere you're supposed to get it (meaning you won't be getting anything related to ϕ(n) once n > base).
@@SgtSupaman If you modify the pattern so that that's how it works, what number base is the best?
Why can't you have 2/2 or 3/3 or 2/4 or 4/4 etc?
7:10 Damn it, it's that Euler guy, again!
Third part where?
so if 17 is left out does that mean that it is a super prime? then i would venture out and say that it is a safe bet that there are more primes that would be skipped in doing it this way.
honestly, this video, from initially super exciting to me, eventually became super disappointing...
this precise moment is what drew my attention, and I started awaiting with excitement the moment when he would reveal what OTHER prime numbers would be missing...
...and then eventually he payed NO further attention to this curious detail at all !!! -_-
@@kurzackd what it might be is that the next one is somewhere in the thousands or tens of thousands and getting there by hand would take a long time to do.
@@freshsmilely I don't think it is, but lemme write a script in python and check...
A nice mathematician's pause when that second "1/3" is noticed and fixed offscreen for the next section.
NUMBERPHILE I LOVE YOU'RE VIDEOS 💗
I'm only halfway through the video, but does this mean that the gaps between consecutive primes depend somehow on whether the ranks of the primes are prime numbers themselves?
Oh no the pattern breaks down :(
Primes incremented using sequential Euler totient function 7:26 👍
I don't understand why a 2+3=5 rule applies from row four to five, but not from row three to four. It just seems contrived to produce the desired result, and therefore the sequence isn't interesting at all.
Oh there's no fourth unlisted video 😢
Never mind all that. I want to know why he has a combination lock on the door in the background.
I wonder how the performance of this stacks up against the Sieve of Eratosthenes?
In each row ,the most numerous number is the prime but if tie always choose the prime you Know from the previus rows.
Awesome video!
lmao, Tim Hein being a very high odd number
Who is the 3rd B-list actor?
neat sleight of hand at 3:47 :)
This is why you asked for A list and B list actors on Twitter haha
awesome collab
Well I suppose that one "reason" why you are getting primes at the begining is that this method will never produce an even number, that is guaranteed.
It's even weaker than the Paterson method where 2,3 and 5 are excluded as divisors, but it is there :).
On the line for number 10 is doesn't break if you count digits, since it becomes 37, not 33.
Suddenly, it's not unlisted anymore!
The quip about 3b1b being "A list" haha, you certainly are too tho Bradey, I literally started learning math in my 20's because of your channels! :D
3Blue1Brown is the GOAT.
There is a pattern in the pyramid. If you have a row n, then you will see that number in the row (n - 1) times. For row 2, the number two appears once (2 - 1 = 1).
The algorithm for building the pyramid and doing this test would not be an efficient method for finding primes.