New World Record! 100 Trillion digits of π.

Check out Emma's blog post about the calculation:

Register for Emma's live seminar! It's on at 19:00 BST and different time zones are different.

Emma Haruka Iwao's records:
2019: 31,415,926,535,897 digits of π
2022: 100,000,000,000,000 digits of π

Chudnovsky algorithm

Bailey–Borwein–Plouffe formula

y-cruncher - A Multi-Threaded Pi-Program

CORRECTIONS:
- At 07:34 I say that the first column are the hex locations. They are not. Those are the base-10 locations in scientific notation. Sorry!
- A few people noticed that at 02:20 I say 128 when I mean 158. The on-screen number is correct.
- Let me know if you spot any other mistakes

Filming and editing by Alex Genn-Bash
Additional footage by Emma Haruka Iwao
Blah blah blah by Matt
Music by Howard Carter
Design by Simon Wright and Adam Robinson

MATT PARKER: Stand-up Mathematician

[thomasfrewer]

Google's automatic transcription is going to listen to Matt say "the last digit of Pi is zero" 0:27 and serve it up to at least one person who googles it.

[legohead]

I’m so glad that mathematicians have finally figured out 0% of pi

[janmelantu]

“My personal best for calculating Pi is 14 digits and turns out Pi’s personal best is 14 digits” That’s a Parker Circle Constant if I’ve ever seen one

[utOfSkill]

At 3:14 Matt says "classic π" and I refuse to believe that wasn't intentional.

[davidhutchins]

Delightful ... 2x better than one person talking about PI. Now if you could only get 3.14 people obsessed with PI in one video...

[ExtantThylacine]

Can we just take a moment to appreciate Matt's brilliant editing by timing the moment he says: "Classic pi!" to be at exactly 3:14 ?

[Jupiterninja]

100 trillion is 10^14, and if π is a normal number there's a 1 in 10^14 chance of a particular string of 14 digits, so I think finding a 14 digit string of meta π makes perfect sense!

[GrayBlood]

It's actually not too surprising that there's only 14 digits of mini-pi. Each successive correct digit of pi would be another order of magnitude less likely and would require another order of magnitude of regular pi to look through. 100 trillion is 10^14, so it makes perfect sense.

[AmaroqStarwind]

Scientists: “We can probably stop now.”
Engineers: “We should have stopped at ten digits.”
Mathematicians: “We need to keep going.”

[ClostridiumChampion]

I like the poster in the background with "Education works best when all the parts are working." Featuring 3 gears that physically cannot turn, because they're interlocked.

[kayodesalandy]

*Matt at 13% error rate*: "Not too bad!"
Emma: "I have a duty to precision."

[thomashirsch]

an excellent achievement! Minor correction: at 2:20, the voiceover says the calculation took almost 128 days, when you should have said "almost 158 days". Very understandable - it would have been pretty cool to do this in 2^7 days

[averyredpandy]

Regarding the recursive pi it actually lines up almost perfectly with the probability. Since any given digit has a 1/10 chance of being the correct one, the chance of finding a 14 length recursive pi is 1 in a hundred trillion. Also there's 10 13-length ones(including the 14-length) and 98 12-length, all of which are pretty spot on.

[Vini-BR]

If a person could somehow memorize 10 thousand digits per second, it would still take them 316 years to reach 100 trillion digits

[macdofglasgow]

BBP still blows my mind. Just seems counter-intuitive that you can pull out specific digits like that. Fun video as typical by Matt.

[Hallgrenoid]

0:27 "The last digit of pi is 0"
- Matt Parker, 2022

Teacher: Is π rational? 
Student: YES!
Teacher: wrong! Where did you hear that?
Student: shows resources. 0:14 "This is the last 100 digit of PI"

[Qazqi]

I'm loving the interviews. It might be cool to have Alexander Yee (the creator of y-cruncher and former record holder) on here sometime to go into some nerd talk about the program. I've had the pleasure of being able to chat with him outside of his legendary StackOverflow Q&A presence, and it's probably worth a shot if you're interested in having him on.

[SeniorPoteyto]

Just in time for 1M subs! Congrats Matt.

[dantrizz]

From what I remember there are 10^80 particles in the observable universe so, the probability of an error in the calculation being 1 in 10^80 is equivalent to Matt choosing a random particle in the observable universe, Emma also choosing a random particle in the observable universe, and those both happen to be the same particle.