RQYhREEc

1[00:00]
2the following is a conversation with
3
4[00:01]
5donald knuth one of the greatest and
6
7[00:04]
8most impactful computer scientists and
9
10[00:06]
11mathematicians ever he's the recipient
12
13[00:10]
14of the 1974 Turing award considered the
15
16[00:14]
17Nobel Prize of computing he's the author
18
19[00:16]
20of the multi-volume work the magnum opus
21
22[00:19]
23the art of computer programming he made
24
25[00:23]
26several key contributions to the
27
28[00:25]
29rigorous analysis of computational
30
31[00:27]
32complexity of algorithms including the
33
34[00:30]
35popularization of asymptotic notation
36
37[00:33]
38that we all affectionately know as the
39
40[00:35]
41Big O notation he also created the tech
42
43[00:39]
44typesetting system which most computer
45
46[00:41]
47scientists physicists mathematicians and
48
49[00:44]
50scientists and engineers in general used
51
52[00:47]
53to write technical papers and make them
54
55[00:49]
56look beautiful I can imagine no better
57
58[00:53]
59guest to in 2019 with than Don one of
60
61[00:57]
62the kindest most brilliant people in our
63
64[00:59]
65field this podcast was recorded many
66
67[01:02]
68months ago it's one I avoided because
69
70[01:04]
71perhaps counter-intuitively the
72
73[01:06]
74conversation meant so much to me if you
75
76[01:09]
77can believe it I knew even less about
78
79[01:11]
80recording back then so the camera angle
81
82[01:13]
83is a bit off I hope that's okay with you
84
85[01:15]
86the office space was a bit cramped for
87
88[01:18]
89filming but it was a magical space Ordon
90
91[01:22]
92does most of his work it meant a lot to
93
94[01:24]
95me that he would welcome me into his
96
97[01:26]
98home it was quite a journey to get there
99
100[01:28]
101as many people know he doesn't check
102
103[01:30]
104email so I had to get creative the
105
106[01:33]
107effort was worth it I've been doing this
108
109[01:36]
110podcast on the side for just over a year
111
112[01:38]
113sometimes I had to sacrifice a bit of
114
115[01:40]
116sleep but always happy to do it and to
117
118[01:42]
119be part of an amazing community of
120
121[01:44]
122curious minds thank you for your kind
123
124[01:47]
125words support for the interesting
126
127[01:49]
128discussions and I look forward to many
129
130[01:51]
131more of those in 2020 this is the
132
133[01:55]
134artificial intelligence podcast if you
135
136[01:57]
137enjoy it subscribe on YouTube give it
138
139[02:00]
140five stars an Apple podcast follow on
141
142[02:02]
143Spotify support on patreon or simply
144
145[02:04]
146connect with me on Twitter at lex
147
148[02:06]
149friedman spelled fri d-m am
150
151[02:10]
152I recently started doing ads at the end
153
154[02:12]
155of the introduction I'll do one or two
156
157[02:14]
158minutes after introducing the episode
159
160[02:16]
161and never any ads in the middle that
162
163[02:18]
164break the flow of the conversation I
165
166[02:19]
167hope that works for you and doesn't hurt
168
169[02:21]
170the listening experience I provide time
171
172[02:24]
173stamps for the start of the conversation
174
175[02:25]
176that you can skip to but it helps if you
177
178[02:28]
179listen to the ad and support this
180
181[02:30]
182podcast by trying out the product the
183
184[02:32]
185service being advertised this show is
186
187[02:34]
188presented by cash app the number one
189
190[02:37]
191finance app in the App Store
192
193[02:38]
194I personally use cash app to send money
195
196[02:40]
197to friends but you can also use it to
198
199[02:43]
200buy sell and deposit Bitcoin in just
201
202[02:45]
203seconds cash app also has a new
204
205[02:48]
206investing feature you can buy a fraction
207
208[02:50]
209of a stock say $1 worth no matter what
210
211[02:53]
212the stock price is brokerage services
213
214[02:55]
215are provided by cash app investing a
216
217[02:57]
218subsidiary of square and member s IBC
219
220[03:00]
221I'm excited to be working with cash app
222
223[03:03]
224to support one of my favorite
225
226[03:04]
227organizations called first best known
228
229[03:07]
230for their first robotics and Lego
231
232[03:09]
233competitions they educate and inspire
234
235[03:11]
236hundreds of thousands of students in
237
238[03:14]
239over 110 countries and have a perfect
240
241[03:16]
242rating and charity navigator which means
243
244[03:18]
245that donated money is used to maximum
246
247[03:21]
248effectiveness when you get cash app from
249
250[03:24]
251the App Store or Google Play
252
253[03:26]
254and use code Lex podcast you'll get ten
255
256[03:29]
257dollars in cash up will also donate ten
258
259[03:31]
260dollars the first which again is an
261
262[03:34]
263organization that I've personally seen
264
265[03:35]
266inspire girls and boys to dream of
267
268[03:38]
269engineering a better world and now
270
271[03:40]
272here's my conversation with Donald Knuth
273
274[03:43]
275in 1957 atcase tech you were once
276
277[03:50]
278allowed to spend several evenings with a
279
280[03:53]
281IBM 650 computer as you've talked about
282
283[03:56]
284in the past then you fell in love with
285
286[03:57]
287computing then
288
289[03:59]
290can you take me back to that moment with
291
292[04:02]
293the IBM 650 what was it that grabs you
294
295[04:07]
296about that computer so the IBM 650 was
297
298[04:11]
299this this machine that
300
301[04:14]
302well it didn't fill a room but it it was
303
304[04:17]
305it was big and noisy but when I first
306
307[04:21]
308saw it it was through a window and there
309
310[04:23]
311were just a lot of lights flashing on it
312
313[04:25]
314and I was a freshman I had a job with
315
316[04:32]
317the statistics group and I was supposed
318
319[04:35]
320to punch cards and pour data and then
321
322[04:38]
323sort them on another machine but then
324
325[04:40]
326they got this new computer came in and I
327
328[04:44]
329and it had interesting like you know
330
331[04:47]
332lights okay so well but I had it kind of
333
334[04:50]
335key to the building so I can you know
336
337[04:52]
338like I could get in and look at it and
339
340[04:53]
341got a manual for it and and my first
342
343[04:57]
344experience was based on the fact that I
345
346[04:58]
347could punch cards basically would you a
348
349[05:00]
350big thing for though deal with thick but
351
352[05:02]
353the is
354 [6:50 ]was you know big in size but
355
356[05:06]
357but incredibly small in power in memory
358
359[05:12]
360it had it had 2,000 words of memory and
361
362[05:16]
363in a word of memory was 10 decimal
364
365[05:18]
366digits plus a sign and it it would do to
367
368[05:22]
369add two numbers together you could
370
371[05:24]
372probably expect that would take oh say
373
374[05:28]
375three milliseconds so that's pretty fast
376
377[05:31]
378it's the memories that constraint the
379
380[05:33]
381memories the problem that was why it was
382
383[05:35]
384three millisecond because it took five
385
386[05:37]
387milliseconds for the drum to go around
388
389[05:39]
390and you had to wait I don't know five
391
392[05:43]
393cycle times if you have an instruction
394
395[05:46]
396one position on the drum then it would
397
398[05:49]
399be ready to read the data for the
400
401[05:50]
402instruction and three notches the drum
403
404[05:55]
405is 50 cycles around and you go three
406
407[05:58]
408cycles and you can get the data and then
409
410[06:00]
411you can go another three cycles and get
412
413[06:02]
414and get to next instruction if the
415
416[06:04]
417instruction is there otherwise otherwise
418
419[06:06]
420you spin until you get to there
421
422[06:08]
423play and and we had no random-access
424
425[06:12]
426memory whatsoever until my senior year
427
428[06:13]
429you see here we got fifty words of
430
431[06:16]
432random access memory which were which
433
434[06:17]
435were priceless and we would and we would
436
437[06:19]
438move stuff up to the up to the random
439
440[06:23]
441access memory in 60 word chunks and then
442
443[06:26]
444we would start again so it's separating
445
446[06:29]
447when to go up there and could you have
448
449[06:31]
450predicted the future 60 years later of
451
452[06:35]
453computing from then you know in fact the
454
455[06:38]
456hardest question I was ever asked was
457
458[06:42]
459what could I have predicted in other
460
461[06:44]
462words the interviewer asked me she said
463
464[06:47]
465you know what about computing has
466
467[06:50]
468surprised you you know and immediately I
469
470[06:51]
471ran I rattled off a couple dozen things
472
473[06:54]
474and inches okay so what didn't surprise
475
476[06:56]
477and I was I tried for five minutes to
478
479[07:00]
480think of something that I thought I
481
482[07:01]
483would have predicted and I and I and I
484
485[07:03]
486couldn't but I let me say that this
487
488[07:07]
489machine I didn't know well it there
490
491[07:10]
492wasn't there wasn't much else in the
493
494[07:12]
495world at that time the 650 was the first
496
497[07:14]
498machine that was that there were more
499
500[07:17]
501than a thousand of ever before that
502
503[07:19]
504there were you know there was each
505
506[07:22]
507machine there might be a half a dozen
508
509[07:23]
510examples maybe my first mass-market
511
512[07:26]
513mass-produced the first one yeah done in
514
515[07:30]
516quantity and and IBM I didn't sell them
517
518[07:35]
519they they rented them but but they they
520
521[07:38]
522rented them to universities that at
523
524[07:40]
525great you know I had a great deal and
526
527[07:44]
528and so that's why a lot of students
529
530[07:48]
531learned about computers at that time so
532
533[07:51]
534you refer to people including yourself
535
536[07:54]
537who gravitate toward a kind of
538
539[07:57]
540computational thinking as geeks for at
541
542[08:00]
543least I've heard you used that
544
545[08:02]
546terminology it true that I think there's
547
548[08:05]
549something that happened to me as I was
550
551[08:07]
552growing up that made my brain structure
553
554[08:10]
555in a certain way that resonates with
556
557[08:12]
558with computers so there's the space of
559
560[08:15]
561people it's 2% of the population you
562
563[08:17]
564empirically estimate that's a prick
565
566[08:21]
567it's been proven fairly constant over
568
569[08:23]
570most of my career however it might be
571
572[08:27]
573different now because kids have
574
575[08:28]
576different experiences when they're young
577
578[08:30]
579so what does the world look like to a
580
581[08:34]
582geek what is what is this aspect of
583
584[08:38]
585thinking that is unique to their makes
586
587[08:42]
588it yeah that makes a geek this is cuter
589
590[08:47]
591the important question in in the 50s
592
593[08:51]
594IBM noticed that that there were geeks
595
596[08:57]
597and non geeks and so they tried to hire
598
599[08:59]
600geeks and they put out as worth papers
601
602[09:01]
603saying you know if you play chess come
604
605[09:03]
606to Madison Avenue and for an interview
607
608[09:05]
609or something like this they were they
610
611[09:06]
612were trying for some things so what it
613
614[09:08]
615what what is it that I find easy and
616
617[09:11]
618other people tend to find harder and and
619
620[09:14]
621I think there's two main things one is
622
623[09:17]
624this with is ability to jump jump levels
625
626[09:23]
627of abstraction so you see something in
628
629[09:27]
630the large and you see something in the
631
632[09:30]
633small and and can you pass between those
634
635[09:34]
636unconsciously so you know that in order
637
638[09:37]
639to solve some big problem what you need
640
641[09:41]
642to do is add one to a into a certain
643
644[09:44]
645register or anything that gets you to
646
647[09:46]
648another step and you can and we and
649
650[09:48]
651below the yeah I mean I don't go down to
652
653[09:50]
654the electron level but I knew what those
655
656[09:53]
657milliseconds were what the drum was like
658
659[09:55]
660on the 650 I knew how I was gonna factor
661
662[09:59]
663her number or or find a root of an
664
665[10:01]
666equation or something be alavés because
667
668[10:03]
669of what was doing and and as I'm
670
671[10:05]
672debugging I'm going through you know did
673
674[10:08]
675I make a key punch err did I did I write
676
677[10:12]
678the wrong instruction do I have the
679
680[10:13]
681wrong wrong thing in a register and each
682
683[10:16]
684level at each level it is different and
685
686[10:20]
687so this idea of being able to see
688
689[10:23]
690something at all at lots of levels and
691
692[10:27]
693fluently go between them it seems to me
694
695[10:30]
696to be more pronounced much more
697
698[10:32]
699pronounced in in the people that
700
701[10:34]
702with computers like I got so in my books
703
704[10:38]
705I also don't stick after the high level
706
707[10:41]
708but but i but i mix low level stuff with
709
710[10:47]
711high level and this means that some
712
713[10:50]
714people think you know that I that I
715
716[10:54]
717should write better books and it's
718
719[10:57]
720probably true but but other people say
721
722[11:00]
723well but that's if you think like like
724
725[11:03]
726that then that's the way to train
727
728[11:04]
729yourself like to keep mixing the levels
730
731[11:06]
732and and learn more and more how to jump
733
734[11:10]
735between so that that's the one thing the
736
737[11:11]
738other the other thing is that it's more
739
740[11:14]
741of a talent it to be able to deal with
742
743[11:19]
744non-uniformity where there's case one
745
746[11:22]
747case two case three instead of instead
748
749[11:26]
750of having one or two rules that govern
751
752[11:28]
753everything so if so it doesn't bother me
754
755[11:32]
756if I need like an algorithm has ten
757
758[11:37]
759steps to it you know each step is does
760
761[11:39]
762something else that doesn't bother me
763
764[11:40]
765but a lot of a lot of pure mathematics
766
767[11:43]
768is based on one or two rules which which
769
770[11:46]
771are universal and and and so this means
772
773[11:49]
774that people like me sometimes work with
775
776[11:52]
777systems that are more complicated than
778
779[11:54]
780necessary because it doesn't bother us
781
782[11:55]
783that we don't that we didn't figure out
784
785[11:58]
786the simple rule and you mentioned that
787
788[12:01]
789while Jacobi boule Abel and all the
790
791[12:05]
792mathematicians in 19th century may have
793
794[12:08]
795had symptoms of geek the first hundred
796
797[12:12]
798percent legit geek was touring Alan
799
800[12:15]
801Torrie I I think he had yeah a lot more
802
803[12:17]
804of this quality than anyone could from
805
806[12:23]
807reading the kind of stuff he didn't so
808
809[12:27]
810hot as touring what influence has
811
812[12:31]
813touring had on you well well your way
814
815[12:34]
816and so I didn't know that aspect of him
817
818[12:38]
819until after I graduated some years I it
820
821[12:40]
822has undergraduate we had a class that
823
824[12:43]
825talked about computability theory and
826
827[12:45]
828Turing machines and and that was all it
829
830[12:49]
831sounded like a very specific kind of
832
833[12:52]
834purely theoretical approach to stuff
835
836[12:55]
837so when how old was I when I when I
838
839[12:58]
840learned that he thought he had you know
841
842[13:02]
843designed when she and that he wrote the
844
845[13:06]
846you know you wrote a wonderful manual
847
848[13:09]
849for for Manchester machines and and he
850
851[13:13]
852invented all the subroutines and and and
853
854[13:19]
855he was a real hacker that that he had
856
857[13:22]
858his hands dirty
859
860[13:23]
861I thought for many years that he had
862
863[13:27]
864only done purely formal work as I
865
866[13:31]
867started reading his own publications I
868
869[13:32]
870could yeah you know I could feel this
871
872[13:34]
873kinship and and of course he had a lot
874
875[13:39]
876of peculiarities like he wrote numbers
877
878[13:42]
879backwards because I mean left to right
880
881[13:46]
882to the right to left because that's the
883
884[13:48]
885that's it was easier for computers to
886
887[13:50]
888process him that way what do you mean
889
890[13:53]
891left to right he would write PI as you
892
893[13:57]
894know nine five one four point three I
895
896[14:01]
897mean okay right forget it for one point
898
899[14:08]
900three on the blackboard I mean when he
901
902[14:12]
903he we had trained himself to to do that
904
905[14:16]
906because the computers he was working
907
908[14:18]
909with I worked that way inside trained
910
911[14:21]
912himself to think like a computer well
913
914[14:22]
915there you go that's nuts geek thinking
916
917[14:26]
918you've practiced some of the most
919
920[14:28]
921elegant formalism in computer science
922
923[14:30]
924and yet you're the creator of a concept
925
926[14:34]
927like literate programming which seems to
928
929[14:37]
930move closer to natural language type of
931
932[14:41]
933description of programming yep yeah
934
935[14:44]
936absolutely so how do you see those two
937
938[14:45]
939as conflicting as the formalism of
940
941[14:48]
942theory and the idea of literate
943
944[14:50]
945programming so there we are in a non
946
947[14:53]
948uniform system well I don't think one
949
950[14:56]
951one-size-fits-all and I don't and I
952
953[14:58]
954don't think all truth lies in one in one
955
956[15:02]
957kind of expertise and so somehow in a
958
959[15:05]
960way you'd say my what my life is a
961
962[15:07]
963convex combination of English and
964
965[15:11]
966mathematics and you're okay with that
967
968[15:14]
969and not only that I think thriving I
970
971[15:16]
972wish you know I want my kids to be that
973
974[15:18]
975way I want cetera not used left-brain
976
977[15:21]
978right-brain at the same time you got a
979
980[15:24]
981lot more done that's that was part of
982
983[15:25]
984the and I've heard that you didn't
985
986[15:31]
987really read for pleasure until into your
988
989[15:33]
99030s literature true you know more about
991
992[15:38]
993me than I do but I'll try to be
994
995[15:40]
996consistent with what you're really ya
997
998[15:41]
999know just believe me
1000
1001[15:42]
1002yeah just go with whatever story I tell
1003
1004[15:45]
1005you it'll be easier that way the
1006
1007[15:46]
1008conversation I've heard mentioned a
1009
1010[15:50]
1011Philip Roth's American pastoral which I
1012
1013[15:53]
1014love as a book I don't know if it was it
1015
1016[15:58]
1017was mentioned as something I think that
1018
1019[15:59]
1020was meaningful to you as well in either
1021
1022[16:03]
1023case what literary books had a lasting
1024
1025[16:06]
1026impact on you what okay good so I so I
1027
1028[16:09]
1029met Russ already well we both got
1030
1031[16:14]
1032doctors from Harvard on the same day so
1033
1034[16:16]
1035I so we were yeah we had lunch together
1036
1037[16:20]
1038and stuff like that and but he knew that
1039
1040[16:22]
1041you know computer books would never sell
1042
1043[16:24]
1044well well all right so you say you you
1045
1046[16:28]
1047you you're a teenager when you left
1048
1049[16:32]
1050Russia so I I have to say that Tolstoy
1051
1052[16:36]
1053was one of the big influences on me
1054
1055[16:38]
1056I especially like Anna Karenina not
1057
1058[16:42]
1059because of a particular area of the plot
1060
1061[16:46]
1062of the story where but because there's
1063
1064[16:51]
1065this character who you know did the
1066
1067[16:54]
1068philosophical discussions it's all it's
1069
1070[16:58]
1071a whole way of life is worked out there
1072
1073[17:02]
1074it's among the characters until in and
1075
1076[17:04]
1077so it that I thought was was especially
1078
1079[17:07]
1080beautiful on the other hand does they
1081
1082[17:09]
1083have ski I I didn't like at all because
1084
1085[17:13]
1086I I felt that he his genius was mostly
1087
1088[17:16]
1089because he kept forgetting what he what
1090
1091[17:17]
1092he had started out to do and he was just
1093
1094[17:19]
1095sloppy I didn't think that that it then
1096
1097[17:23]
1098that he polished his stuff at all and
1099
1100[17:26]
1101and I tend to admire somebody who who
1102
1103[17:30]
1104Todd's the i's and cross the t's so that
1105
1106[17:32]
1107the music of the prose this way you
1108
1109[17:34]
1110admire more and that I certainly do
1111
1112[17:37]
1113admire the music of the language which I
1114
1115[17:39]
1116couldn't appreciate in the Russian
1117
1118[17:41]
1119original but but I can and Victor Hugo
1120
1121[17:44]
1122Glenn's close friendships much his
1123
1124[17:46]
1125closer but but Tolstoy I like the same
1126
1127[17:51]
1128reason I like Herman Wouk as a as a
1129
1130[17:53]
1131novelist I that I think I like his book
1132
1133[17:58]
1134Marjorie Morningstar has a similar
1135
1136[18:00]
1137character in who who who developed his
1138
1139[18:02]
1140own personal philosophy and export and
1141
1142[18:05]
1143it called goes in in was consistent yeah
1144
1145[18:10]
1146right and it's worth worth pondering uh
1147
1148[18:14]
1149so zo like Nietzsche and like what you
1150
1151[18:18]
1152don't like Friedrich Nietzsche or age
1153
1154[18:20]
1155yeah no no you like this has like I keep
1156
1157[18:24]
1158seeing quotations for Nietzsche and and
1159
1160[18:26]
1161you never tempt me to read any further
1162
1163[18:29]
1164please full of contradictions we will
1165
1166[18:32]
1167certainly not appreciate him but
1168
1169[18:34]
1170Schiller you know I'm trying to get the
1171
1172[18:37]
1173cross what I appreciate in literature
1174
1175[18:39]
1176and part of it is the is is as you say
1177
1178[18:44]
1179the music of the language of the way it
1180
1181[18:46]
1182flows and take Raymond Chandler versus
1183
1184[18:51]
1185Dashiell Hammett Dashiell Hammett
1186
1187[18:53]
1188sentences are awful and Raymond
1189
1190[18:56]
1191Chandler's are beautiful they just flow
1192
1193[18:59]
1194so I I don't I don't read literature
1195
1196[19:04]
1197because it's supposed to be good for me
1198
1199[19:07]
1200or because somebody said it's great but
1201
1202[19:09]
1203but it I could find things that I like I
1204
1205[19:14]
1206mean you mentioned you address like
1207
1208[19:17]
1209James Bond so like I love Ian Fleming I
1210
1211[19:20]
1212think he's got a he had a really great
1213
1214[19:22]
1215gift for if he has a golf game or game
1216
1217[19:26]
1218of bridge or something and this comes
1219
1220[19:28]
1221into a story it'll it'll be the most
1222
1223[19:30]
1224exciting golf game or or you know the
1225
1226[19:33]
1227absolute best possible hands a bridge
1228
1229[19:36]
1230that that exists and and any he exploits
1231
1232[19:41]
1233it and tells it beautifully as well so
1234
1235[19:45]
1236in connecting some things here looking
1237
1238[19:49]
1239at literate programming and being able
1240
1241[19:51]
1242to it convey encode algorithms to a
1243
1244[19:59]
1245computer in a way that mimics how humans
1246
1247[20:03]
1248speak how what do you think about
1249
1250[20:06]
1251natural language in general and the
1252
1253[20:08]
1254messiness of our human world about
1255
1256[20:11]
1257trying to express yeah difficult things
1258
1259[20:14]
1260so the idea of literate programming is
1261
1262[20:17]
1263to is really to try to understand
1264
1265[20:24]
1266something better by seeing it from these
1267
1268[20:26]
1269two perspectives the formal and the
1270
1271[20:28]
1272informal if we try to understand a
1273
1274[20:31]
1275complicated thing if we can look at it
1276
1277[20:33]
1278in different ways and so this is in fact
1279
1280[20:36]
1281the key to technical writing a good
1282
1283[20:39]
1284technical writer
1285
1286[20:40]
1287try not to be obvious about it but says
1288
1289[20:42]
1290everything twice formally and informally
1291
1292[20:45]
1293or maybe three times but you try to give
1294
1295[20:48]
1296the reader a way to put the concept into
1297
1298[20:55]
1299his own brain or her own brain is that
1300
1301[20:57]
1302better for the writer or the reader or
1303
1304[21:00]
1305both well the writer just tries to
1306
1307[21:05]
1308understand the reader that's the goal of
1309
1310[21:07]
1311a writer is to have a good mental image
1312
1313[21:10]
1314of the reader and to say what the reader
1315
1316[21:13]
1317expects next and to to impress the
1318
1319[21:18]
1320reader with what has impressed the
1321
1322[21:19]
1323writer why something is interesting so
1324
1325[21:24]
1326when you have a computer program we try
1327
1328[21:26]
1329to instead of looking at it as something
1330
1331[21:29]
1332that we're just trying to give an
1333
1334[21:30]
1335instruction to the computer what we
1336
1337[21:32]
1338really want to be is giving giving
1339
1340[21:35]
1341insight to the person who's who's gonna
1342
1343[21:39]
1344be maintaining this program or to the
1345
1346[21:41]
1347programmer himself when he's debugging
1348
1349[21:44]
1350it as to why this stuff is being done
1351
1352[21:46]
1353and so all the techniques of exposition
1354
1355[21:50]
1356that a teacher uses or book writers make
1357
1358[21:54]
1359you better program or if your if your
1360
1361[21:56]
1362program is going to be not just a
1363
1364[21:59]
1365one-shot deal so how difficult is that
1366
1367[22:03]
1368do you see hope for the combination of
1369
1370[22:07]
1371informal and formal for the programming
1372
1373[22:11]
1374task yeah I I'm the wrong person to ask
1375
1376[22:14]
1377I guess because I'm a geek but but I
1378
1379[22:17]
1380think for a geek it's easy I don't know
1381
1382[22:19]
1383I don't know see not some people have
1384
1385[22:23]
1386difficulty writing and that might be
1387
1388[22:26]
1389because there's something in their brain
1390
1391[22:29]
1392structure that makes it hard for them to
1393
1394[22:32]
1395write or or it might be something just
1396
1397[22:34]
1398that they haven't had enough practice
1399
1400[22:35]
1401I'm not the right one to to uh to judge
1402
1403[22:39]
1404but I don't think you teach any person
1405
1406[22:42]
1407any particular skill like I do think
1408
1409[22:45]
1410that that writing is is half of my life
1411
1412[22:49]
1413and so I put it together and let
1414
1415[22:51]
1416program he won't even when I'm writing a
1417
1418[22:53]
1419one-shot program I I write it in
1420
1421[22:58]
1422literate way because I get it right
1423
1424[23:02]
1425faster though now does it get compiled
1426
1427[23:05]
1428automatically or so I guess on the
1429
1430[23:09]
1431technical side my question was how
1432
1433[23:12]
1434difficult is a design a system where
1435
1436[23:15]
1437much of the programming is done
1438
1439[23:17]
1440informally informally yeah informally I
1441
1442[23:21]
1443think whatever works to make it
1444
1445[23:25]
1446understandable is good but then you have
1447
1448[23:28]
1449to also understand how informal is you
1450
1451[23:33]
1452have to know the limitations you have to
1453
1454[23:35]
1455connect so so by putting the formula and
1456
1457[23:38]
1458informal together this this is where
1459
1460[23:41]
1461this is where it gets locked into your
1462
1463[23:43]
1464into your brain now you can you can say
1465
1466[23:48]
1467informally well I'm working on a problem
1468
1469[23:51]
1470right now so let's go there I get that
1471
1472[23:54]
1473can you give me an example of of
1474
1475[23:57]
1476connecting the informal in the formal
1477
1478[23:59]
1479well it's a little too complicated an
1480
1481[24:02]
1482example there's a puzzle that that's
1483
1484[24:05]
1485self referential it's called a Japanese
1486
1487[24:07]
1488arrow puzzle and and and you're given a
1489
1490[24:11]
1491a bunch of boxes each one points north
1492
1493[24:14]
1494east south or west and at the end you're
1495
1496[24:18]
1497supposed to fill in each box with the
1498
1499[24:20]
1500number of distinct numbers that it
1501
1502[24:23]
1503points to so if I put a three in a box
1504
1505[24:26]
1506that means that and it's pointing to
1507
1508[24:29]
1509five other boxes that means that there's
1510
1511[24:30]
1512going to be three different numbers in
1513
1514[24:32]
1515those five bucks and and those boxes are
1516
1517[24:36]
1518pointing what I might be pointing to me
1519
1520[24:38]
1521one of my might be pointing the other
1522
1523[24:39]
1524way but anyway I kind of defined a set
1525
1526[24:44]
1527of numbers that obeys this complicated
1528
1529[24:47]
1530condition that each number counts how
1531
1532[24:50]
1533many distinct numbers if it points do
1534
1535[24:52]
1536well and still a guy sent me his
1537
1538[24:57]
1539solution to this problem where he where
1540
1541[25:00]
1542he presents
1543
1544[25:03]
1545formal statements that that say either
1546
1547[25:06]
1548this is true or this is true this is
1549
1550[25:07]
1551true and and and so I try to render that
1552
1553[25:10]
1554formal statement informally and I try
1555
1556[25:14]
1557say I contain a three and and the guys
1558
1559[25:20]
1560I'm pointing to contain the numbers one
1561
1562[25:23]
1563two and six so by putting it in formally
1564
1565[25:26]
1566and also I converted into a into a
1567
1568[25:29]
1569dialogue statement that helps me
1570
1571[25:32]
1572understand the logical statement that
1573
1574[25:35]
1575he's written down as a string of numbers
1576
1577[25:37]
1578in terms of some abstract variables
1579
1580[25:40]
1581Eddie yeah
1582
1583[25:40]
1584that's really interesting so maybe an
1585
1586[25:43]
1587extension of that there has been a
1588
1589[25:46]
1590resurgence in computer science and
1591
1592[25:48]
1593machine learning and neural networks so
1594
1595[25:52]
1596using data to construct algorithms so
1597
1598[25:56]
1599it's another way to construct algorithms
1600
1601[25:58]
1602really yes you can think of it that way
1603
1604[26:03]
1605so as opposed to natural language to
1606
1607[26:05]
1608construct algorithms use data to
1609
1610[26:06]
1611construct other so what what's the view
1612
1613[26:10]
1614of this branch of computer science where
1615
1616[26:13]
1617data is almost more important than the
1618
1619[26:16]
1620mechanism of the algorithm it seems to
1621
1622[26:19]
1623be suited to a certain kind of non geek
1624
1625[26:23]
1626and would you know which is probably why
1627
1628[26:25]
1629it's it's like it's taken off that it
1630
1631[26:29]
1632has its own community that I thought
1633
1634[26:31]
1635really that really resonates with that
1636
1637[26:33]
1638but it's hard to you know to trust
1639
1640[26:37]
1641something like that because nobody even
1642
1643[26:40]
1644the people who who work with it that
1645
1646[26:43]
1647they have no idea what is what has been
1648
1649[26:45]
1650learned that's a really interesting
1651
1652[26:48]
1653thought that it's it makes algorithms
1654
1655[26:53]
1656more accessible to a different community
1657
1658[26:56]
1659a different type of brain yep and that's
1660
1661[26:59]
1662really interesting because just like
1663
1664[27:03]
1665literate programming perhaps could make
1666
1667[27:06]
1668programming more accessible to a certain
1669
1670[27:09]
1671kind of brain there are people who think
1672
1673[27:11]
1674it's just a matter of Education and
1675
1676[27:13]
1677anybody can learn to be a great program
1678
1679[27:16]
1680or anybody can
1681
1682[27:17]
1683to be a great skier uh yeah you know I I
1684
1685[27:23]
1686wish that were true but but I know that
1687
1688[27:25]
1689there's a lot of things that I've tried
1690
1691[27:27]
1692to do and I and like I was well motivate
1693
1694[27:30]
1695an icon and I kept trying to build
1696
1697[27:33]
1698myself up and I never got past a certain
1699
1700[27:35]
1701level I can't use for example I can't
1702
1703[27:38]
1704view three-dimensional objects in my in
1705
1706[27:43]
1707my head I have to I have to make a model
1708
1709[27:45]
1710and look at it and study it from all
1711
1712[27:47]
1713points of view and then I start to get
1714
1715[27:49]
1716some idea but other people are good at
1717
1718[27:52]
1719four dimensions I mean physicists yeah
1720
1721[27:57]
1722so let's go to the art of computer
1723
1724[28:03]
1725programming in 1962 you set the table of
1726
1727[28:07]
1728contents for this magnum opus right yeah
1729
1730[28:13]
1731it was supposed to be a single book for
1732
1733[28:15]
173412 chapters now today what is it
1735
1736[28:19]
173757 years later you're in the middle of
1738
1739[28:23]
1740volume 4 of 7 and in the middle of going
1741
1742[28:27]
1743for B is 4 B precisely can ask you for
1744
1745[28:31]
1746an impossible task which is try to
1747
1748[28:34]
1749summarize the book so far maybe by
1750
1751[28:39]
1752giving a little examples so from the
1753
1754[28:42]
1755sorting and the search in the
1756
1757[28:43]
1758combinatorial algorithms if you were to
1759
1760[28:46]
1761give a summary a quick elevator summary
1762
1763[28:51]
1764yeah right what depending how many
1765
1766[28:53]
1767floors that are in the building yes
1768
1769[28:55]
1770the first volume called fundamental
1771
1772[28:57]
1773algorithms talks about something that
1774
1775[29:01]
1776you can't the stuff you can't do without
1777
1778[29:03]
1779I guess that you have to know the basic
1780
1781[29:07]
1782concepts of what is a program now what
1783
1784[29:10]
1785is it what is it algorithm and and and
1786
1787[29:13]
1788it also talks about a low-level machine
1789
1790[29:15]
1791so you can have some some kind of an
1792
1793[29:17]
1794idea what's going on and it has basic
1795
1796[29:22]
1797concepts of input/output and subroutines
1798
1799[29:26]
1800induction induction writes mathematical
1801
1802[29:30]
1803so so the thing that makes my book
1804
1805[29:33]
1806different from a lot of others is that
1807
1808[29:37]
1809all that I try to not only present the
1810
1811[29:40]
1812algún but I try to analyze them and
1813
1814[29:42]
1815which means to quantitatively I say not
1816
1817[29:44]
1818only does it work but it works this fast
1819
1820[29:46]
1821okay and so I need math for them and
1822
1823[29:49]
1824then there's the standard way to
1825
1826[29:51]
1827structure data inside and represent
1828
1829[29:53]
1830information in the computer so that's
1831
1832[29:56]
1833all volume 1 volume 2 talks
1834
1835[29:59]
1836it's called semi numerical algorithms
1837
1838[30:01]
1839and here we're here we're writing
1840
1841[30:03]
1842programs but we're also dealing with
1843
1844[30:06]
1845numbers algorithms deal with with with
1846
1847[30:09]
1848any kinds of objects but but specific
1849
1850[30:11]
1851when there's objects or numbers well
1852
1853[30:13]
1854then then we have certain special
1855
1856[30:17]
1857paradigms that apply to things that have
1858
1859[30:19]
186012 numbers and so there's there's what
1861
1862[30:21]
1863there's like there's arithmetic on
1864
1865[30:24]
1866numbers and and there's matrices full of
1867
1868[30:26]
1869numbers there's random numbers and
1870
1871[30:29]
1872there's power series full of numbers
1873
1874[30:31]
1875there's different algebraic concepts
1876
1877[30:34]
1878that have numbers in structured ways and
1879
1880[30:37]
1881the arithmetic in the way a computer
1882
1883[30:38]
1884would think about arithmetic is a
1885
1886[30:40]
1887floating point floating point arithmetic
1888
1889[30:42]
1890a high precision arithmetic not only
1891
1892[30:45]
1893addition subtraction multiplication but
1894
1895[30:47]
1896also comparison up number
1897
1898[30:50]
1899so then check then volume three talks
1900
1901[30:53]
1902about I like that one sort insert
1903
1904[30:55]
1905sorting a circle of sorting right so so
1906
1907[30:58]
1908here you know we're not getting
1909
1910[30:59]
1911necessarily with numbers because you
1912
1913[31:01]
1914slipped you saw it letters and other
1915
1916[31:03]
1917objects and searching we're doing all
1918
1919[31:04]
1920the time we googled nowadays but I mean
1921
1922[31:06]
1923we have to find stuff
1924
1925[31:08]
1926so again algorithms that that underlie
1927
1928[31:13]
1929all kinds of applications like you know
1930
1931[31:16]
1932none of these volumes it's about a
1933
1934[31:17]
1935particular application but the
1936
1937[31:19]
1938applications are examples of of why
1939
1940[31:22]
1941people want to know about sorting why
1942
1943[31:23]
1944people want to know about random numbers
1945
1946[31:25]
1947so then volume 4 goes into combinatorial
1948
1949[31:29]
1950I'll again this is where we have
1951
1952[31:32]
1953zillions of things to deal with and we
1954
1955[31:35]
1956and here we keep finding cases where one
1957
1958[31:41]
1959good idea can can make something go more
1960
1961[31:43]
1962than a million times faster and and and
1963
1964[31:48]
1965we're dealing with problems that are
1966
1967[31:50]
1968probably never going to be solved
1969
1970[31:52]
1971efficiently but that doesn't mean we
1972
1973[31:55]
1974give up on them and and and we have this
1975
1976[31:58]
1977chance to have good ideas and and go
1978
1979[32:00]
1980much much faster on them so so that's
1981
1982[32:03]
1983comets are all algorithms and those are
1984
1985[32:05]
1986the ones that are yeah I'm using
1987
1988[32:07]
1989charting is most fun for you well how
1990
1991[32:11]
1992many toriel algorithms are the ones that
1993
1994[32:14]
1995I always that I always enjoyed the most
1996
1997[32:17]
1998because that's when my skillet
1999
2000[32:20]
2001programming had most payoff you know the
2002
2003[32:23]
2004different the difference between an
2005
2006[32:24]
2007obvious algorithm that you think up
2008
2009[32:26]
2010first thing and you know and a good you
2011
2012[32:29]
2013know an interesting subtle out algorithm
2014
2015[32:32]
2016that not so obvious but but run circles
2017
2018[32:36]
2019around the other one that's that's where
2020
2021[32:39]
2022computer science 3d comes comes in and
2023
2024[32:42]
2025and a lot of these comets are methods
2026
2027[32:45]
2028were found first in applications to
2029
2030[32:49]
2031artificial intelligence or cryptography
2032
2033[32:53]
2034and in my case I I just liked him and it
2035
2036[32:58]
2037was associated more with puzzles that
2038
2039[33:00]
2040you like the most in the domain of
2041
2042[33:02]
2043graphs and graph theory graphs are great
2044
2045[33:05]
2046because they're terrific models of so
2047
2048[33:08]
2049many things in the real world and and
2050
2051[33:10]
2052and and you you throw numbers on a graph
2053
2054[33:13]
2055you got a network and so there you're
2056
2057[33:15]
2058right there you have but many more
2059
2060[33:18]
2061things so but comma toriel in general is
2062
2063[33:22]
2064in any arrangement of objects that that
2065
2066[33:26]
2067has some kind of a higher structure non
2068
2069[33:30]
2070non random structure and it's okay
2071
2072[33:34]
2073it is possible to put something together
2074
2075[33:37]
2076satisfying all these conditions like I
2077
2078[33:39]
2079mentioned arrows a minute ago you know
2080
2081[33:41]
2082is there a way to to put these numbers
2083
2084[33:44]
2085on a bunch of boxes that that are
2086
2087[33:46]
2088pointing to each other is that going to
2089
2090[33:47]
2091be possible at all that's volume four
2092
2093[33:49]
2094that's volume four what is a sage of
2095
2096[33:52]
2097Hawaiian for a was part one and and what
2098
2099[33:56]
2100happened was in 1962 when I started
2101
2102[33:59]
2103writing down a table of contents it
2104
2105[34:03]
2106wasn't going to be a book about computer
2107
2108[34:06]
2109programming in general it was going to
2110
2111[34:07]
2112be a book about how to write compilers
2113
2114[34:09]
2115and I was asked to write a book
2116
2117[34:13]
2118explaining how to how to write a
2119
2120[34:15]
2121compiler and at that time there were
2122
2123[34:20]
2124only a few dozen people in the world who
2125
2126[34:22]
2127had written compilers and I happen to be
2128
2129[34:24]
2130one of them so and I also had some
2131
2132[34:29]
2133experience for writing for like the
2134
2135[34:33]
2136campus newspaper and things like that so
2137
2138[34:35]
2139so I said okay great I'm the only person
2140
2141[34:39]
2142I know who who's written a compiler but
2143
2144[34:42]
2145hasn't invented any new techniques for
2146
2147[34:43]
2148writing compilers and and all the other
2149
2150[34:45]
2151people I knew had super ideas but I
2152
2153[34:50]
2154couldn't see that they would be able to
2155
2156[34:51]
2157write a book that wouldn't that would
2158
2159[34:53]
2160describe anybody else's ideas with their
2161
2162[34:55]
2163own so I could be the I could be the
2164
2165[34:57]
2166journalist and I could explained what
2167
2168[34:59]
2169all these cool ideas about compiler
2170
2171[35:02]
2172writing that were and and then I I
2173
2174[35:06]
2175started pretty
2176
2177[35:07]
2178well yeah let me you need and have a
2179
2180[35:09]
2181chapter about data structures you need
2182
2183[35:11]
2184to have some introductory material I
2185
2186[35:13]
2187want to talk about searching because a
2188
2189[35:15]
2190compiler writer has to it has to look up
2191
2192[35:19]
2193the variables in a symbol table and find
2194
2195[35:22]
2196out you know which which when you when
2197
2198[35:27]
2199you write the name of a variable in one
2200
2201[35:29]
2202place it's supposed to be the same as
2203
2204[35:31]
2205the one you put somewhere else so you
2206
2207[35:33]
2208need all these basic techniques and I
2209
2210[35:35]
2211and I you know kind of know some
2212
2213[35:38]
2214arithmetic to stuff so I throw I threw
2215
2216[35:40]
2217in these chapters and I threw in a
2218
2219[35:42]
2220chapter on comma talks because that was
2221
2222[35:46]
2223what I really enjoyed programming the
2224
2225[35:48]
2226most but there weren't many algorithms
2227
2228[35:49]
2229and known about combinatorial methods in
2230
2231[35:51]
22321962 so that was a kind of a short
2233
2234[35:54]
2235chapter but it was sort of thrown in
2236
2237[35:56]
2238just for fun and Chapter twelve was
2239
2240[35:59]
2241going to be actual compilers applying
2242
2243[36:01]
2244all the stuff in chapters 1 to 11 to
2245
2246[36:05]
2247make compilers well ok so that was my
2248
2249[36:07]
2250table of contents from 1962 and during
2251
2252[36:11]
2253the 70s the whole field of combinatoric
2254
2255[36:14]
2256s-- went through a huge explosion people
2257
2258[36:18]
2259talk about it comet oil explosion and
2260
2261[36:20]
2262they usually mean by that that the
2263
2264[36:22]
2265number of cases goes up you know you
2266
2267[36:25]
2268change n to n plus 1 and all of a sudden
2269
2270[36:27]
2271you your problem has gotten more than
2272
2273[36:29]
2274ten times harder but there was an
2275
2276[36:33]
2277explosion of ideas about combinatoric
2278
2279[36:36]
2280s-- in the 70s and to the point that but
2281
2282[36:39]
2283Mike's take 1975 I bet you more than
2284
2285[36:44]
2286half of all the journals of computer
2287
2288[36:45]
2289science we're about combinatorial method
2290
2291[36:48]
2292and what kind of problems were occupying
2293
2294[36:50]
2295people's minds what kind of problems in
2296
2297[36:53]
2298combinatorics was it's it's that gravity
2299
2300[36:56]
2301graph theory yeah gravity was was quite
2302
2303[36:59]
2304dominant I mean no but all of the
2305
2306[37:03]
2307np-hard problems that you have like
2308
2309[37:07]
2310Hamiltonian path or foul sail going
2311
2312[37:10]
2313beyond yeah yeah going beyond graphs you
2314
2315[37:12]
2316had a operation research whenever it was
2317
2318[37:16]
2319a small class of problems that had
2320
2321[37:18]
2322efficient solutions and they were
2323
2324[37:19]
2325associated with Maitre D' a special
2326
2327[37:22]
2328mathematical construction but once we
2329
2330[37:25]
2331went to things that involve three things
2332
2333[37:28]
2334at a time instead of instead of two all
2335
2336[37:30]
2337of a sudden the things got harder so we
2338
2339[37:32]
2340had satisfiability problems or if you
2341
2342[37:35]
2343have if you have clauses every Clause
2344
2345[37:38]
2346has two logical elements in it then we
2347
2348[37:40]
2349can satisfy it linear time we can test
2350
2351[37:43]
2352for satisfy building linear time but if
2353
2354[37:45]
2355you allow yourself three variables in
2356
2357[37:48]
2358the clause then nobody knows how to do
2359
2360[37:52]
2361it so these articles were about trying
2362
2363[37:54]
2364to find better or better ways to to
2365
2366[37:58]
2367solve cryptography problems and graph
2368
2369[38:00]
2370three problems where the we have lots of
2371
2372[38:03]
2373data but we didn't know how to find the
2374
2375[38:05]
2376best subset so the data like with
2377
2378[38:08]
2379sorting we could get the answer didn't
2380
2381[38:12]
2382take long so how did they continue to
2383
2384[38:14]
2385change from the 70s to today yeah so now
2386
2387[38:17]
2388there may be half a dozen conferences
2389
2390[38:20]
2391whose topic is cognate arcs different
2392
2393[38:24]
2394kind but fortunately I don't have to
2395
2396[38:26]
2397rewrite my book every month you know
2398
2399[38:28]
2400like I had to in in the 70 but still
2401
2402[38:31]
2403there's huge amount of work being done
2404
2405[38:33]
2406and people getting better ideas on these
2407
2408[38:37]
2409problems that don't seem to have really
2410
2411[38:40]
2412efficient solutions but we can still get
2413
2414[38:42]
2415into a lot more with him and so this
2416
2417[38:45]
2418book that I'm finishing now is I've got
2419
2420[38:48]
2421a whole bunch of brand new methods that
2422
2423[38:51]
2424the fires I know there's no other
2425
2426[38:53]
2427there's no other book that covers that
2428
2429[38:57]
2430covers this particular approach and and
2431
2432[39:00]
2433so I'm trying to do my best of exploring
2434
2435[39:04]
2436the tip of the iceberg and and and I try
2437
2438[39:08]
2439out lots of things and and keep keep
2440
2441[39:11]
2442rewriting finding as I find better
2443
2444[39:14]
2445better method so what's your writing
2446
2447[39:17]
2448process like what's your thinking and
2449
2450[39:19]
2451writing process like every day so what's
2452
2453[39:24]
2454your routine even
2455
2456[39:25]
2457yeah I guess it's actually the best
2458
2459[39:29]
2460question because I spent seven days a
2461
2462[39:31]
2463week
2464
2465[39:32]
2466you're doing it the most prepares to
2467
2468[39:35]
2469answer it yeah yeah but okay so the
2470
2471[39:41]
2472chair I'm sitting in is where I do
2473
2474[39:44]
2475that's where the magic happens well
2476
2477[39:47]
2478reading and writing that many chairs
2479
2480[39:49]
2481usually sitting over there where I have
2482
2483[39:50]
2484other books some reference book but but
2485
2486[39:53]
2487I I found his chair which was designed
2488
2489[39:58]
2490by a Swedish guy anyway it turns out
2491
2492[40:01]
2493this was the only chair I can really sit
2494
2495[40:02]
2496in for hours and hours and not know that
2497
2498[40:04]
2499I'm in a chair but then I have the
2500
2501[40:06]
2502stand-up desk right next next to us and
2503
2504[40:08]
2505and so after I write something with
2506
2507[40:11]
2508pencil and eraser I get up and I type it
2509
2510[40:15]
2511and revise and rewrite the kernel the
2512
2513[40:21]
2514idea is first put on paper yep
2515
2516[40:24]
2517that's worth right and I call right
2518
2519[40:27]
2520maybe five programs a week of course
2521
2522[40:31]
2523literate programming and these are
2524
2525[40:34]
2526before I describe something in my book I
2527
2528[40:36]
2529always program it to see how it's
2530
2531[40:38]
2532working and I and I tried a lot so for
2533
2534[40:42]
2535example I learned at the end of January
2536
2537[40:44]
2538I learned of a breakthrough by for
2539
2540[40:48]
2541Japanese people who had extended one of
2542
2543[40:51]
2544the one of my methods in in a new
2545
2546[40:53]
2547direction and so I I spent the next five
2548
2549[40:56]
2550days writing a program to implement what
2551
2552[40:59]
2553they did and then I you know but they
2554
2555[41:01]
2556had only generalized part of what I had
2557
2558[41:04]
2559done so that I had to see if I could
2560
2561[41:06]
2562generalize more parts of it and then I
2563
2564[41:08]
2565had to take their approach and I had to
2566
2567[41:11]
2568I had to try it out on a couple of dozen
2569
2570[41:13]
2571of the other problems I had already
2572
2573[41:15]
2574worked out with that with my old methods
2575
2576[41:17]
2577and so that took another couple of weeks
2578
2579[41:19]
2580and then I would you know then I then I
2581
2582[41:22]
2583started to see the light nicely and and
2584
2585[41:26]
2586I started writing the final draft and
2587
2588[41:29]
2589and then I would you know type it up
2590
2591[41:32]
2592involves some new mathematical questions
2593
2594[41:34]
2595and so I wrote to my friends and might
2596
2597[41:37]
2598be good at solving those problems and
2599
2600[41:39]
2601and they solve some of them so I put
2602
2603[41:43]
2604that in his exercises and and so a month
2605
2606[41:46]
2607later I had absorbed one new idea that I
2608
2609[41:50]
2610that I learned and you know I'm glad I
2611
2612[41:53]
2613heard about it in time otherwise my I
2614
2615[41:55]
2616wouldn't put my book out before I heard
2617
2618[41:57]
2619about the idea on the other hand this
2620
2621[41:59]
2622book was supposed to come in at 300
2623
2624[42:01]
2625pages and I'm up to 350 now that added
2626
2627[42:04]
262810 pages to the book but if I learn
2629
2630[42:07]
2631about another one I probably first gonna
2632
2633[42:10]
2634shoot me well so in the process in that
2635
2636[42:15]
2637one month process are some days harder
2638
2639[42:18]
2640than others are some days harder than
2641
2642[42:20]
2643others well yeah my work is fun but I
2644
2645[42:23]
2646also work hard and every big job has
2647
2648[42:26]
2649parts that are a lot more fun than
2650
2651[42:28]
2652others and so many days I'll say why do
2653
2654[42:32]
2655I have to have such high standards like
2656
2657[42:35]
2658why couldn't I just be sloppy and not
2659
2660[42:36]
2661try this out and you know just just
2662
2663[42:38]
2664report the answer but I but I know that
2665
2666[42:42]
2667people are conning me to do this and so
2668
2669[42:45]
2670okay so okay Donald grit my teeth and do
2671
2672[42:49]
2673it and and and then the joy comes out
2674
2675[42:52]
2676when I see that actually you know I'm
2677
2678[42:54]
2679getting good results and and and I get
2680
2681[42:56]
2682and I even more when I see that somebody
2683
2684[43:00]
2685has actually read and understood what I
2686
2687[43:02]
2688wrote and told me how to make it even
2689
2690[43:04]
2691better I did want to mention something
2692
2693[43:08]
2694about the about the method so I got this
2695
2696[43:12]
2697tablet here where I do the first you
2698
2699[43:19]
2700know the first writing of concepts okay
2701
2702[43:23]
2703so so and what language I didn't write
2704
2705[43:28]
2706so hey take a look at but you know here
2707
2708[43:30]
2709random say explain how to draw such
2710
2711[43:33]
2712skewed pixel diagrams okay so I got this
2713
2714[43:36]
2715paper about 40 years ago when I was
2716
2717[43:40]
2718visiting my sister in Canada and they
2719
2720[43:42]
2721make tablets of paper with this nice
2722
2723[43:45]
2724large size and just the right very small
2725
2726[43:48]
2727space between like oh yeah yeah
2728
2729[43:50]
2730particularly also just yeah
2731
2732[43:58]
2733you know I've got these manuscripts
2734
2735[44:00]
2736going back to the 60s and and and those
2737
2738[44:06]
2739are when I get my ideas on paper okay
2740
2741[44:09]
2742but I'm a good typist in fact I went to
2743
2744[44:11]
2745type in school when I was when I was in
2746
2747[44:14]
2748high school and so I can type faster
2749
2750[44:15]
2751than I think so then when I do the
2752
2753[44:18]
2754editing you know stand up and type then
2755
2756[44:21]
2757I then I revise this and it comes out a
2758
2759[44:24]
2760lot different than what you look for
2761
2762[44:27]
2763style and rhythm and things like that
2764
2765[44:28]
2766come out at the at the typing state and
2767
2768[44:30]
2769you type in tack and I type in tack and
2770
2771[44:34]
2772can you can you think in tech No so to a
2773
2774[44:38]
2775certain extent I have I have only a
2776
2777[44:40]
2778small number of idioms that I use like
2779
2780[44:44]
2781you know a beginning or theorem I do
2782
2783[44:45]
2784something for displayed equation I do
2785
2786[44:47]
2787something and and so on but I but I I
2788
2789[44:50]
2790have to see it and in the way that it's
2791
2792[44:54]
2793on here yeah right for example touring
2794
2795[44:56]
2796wrote what the other direction
2797
2798[44:59]
2799you don't write macros you don't think
2800
2801[45:03]
2802in macros particularly but when I need a
2803
2804[45:05]
2805macro I'll go ahead and and these and do
2806
2807[45:09]
2808it but but the thing is they I also
2809
2810[45:11]
2811write to fit I mean I'll I'll change
2812
2813[45:15]
2814something if I can if I can save a line
2815
2816[45:17]
2817I've got you know it's like haiku I'll
2818
2819[45:19]
2820figure out a way to rewrite the sentence
2821
2822[45:21]
2823so that it'll look better on the page
2824
2825[45:24]
2826and I shouldn't be wasting my time on
2827
2828[45:26]
2829that but but I can't resist because I
2830
2831[45:29]
2832know it's only another three percent of
2833
2834[45:32]
2835the time or something like that and it
2836
2837[45:34]
2838could also be argued that that is what
2839
2840[45:36]
2841life is about
2842
2843[45:38]
2844ah yes in fact that's true like like I
2845
2846[45:43]
2847worked in the garden one day a week and
2848
2849[45:45]
2850that's that's kind of a description of
2851
2852[45:47]
2853my life is getting rid of weeds you know
2854
2855[45:50]
2856removing bugs for programs in so you
2857
2858[45:53]
2859know a lot of writers talk about you
2860
2861[45:55]
2862know basically suffering the writing
2863
2864[45:57]
2865processes yeah having you know it's
2866
2867[46:00]
2868extremely difficult and I think of
2869
2870[46:02]
2871programming especially the or technical
2872
2873[46:05]
2874writing that you're doing can be like
2875
2876[46:08]
2877that do you find yourself
2878
2879[46:11]
2880methodologically how do you every day
2881
2882[46:14]
2883sit down to do the work is it a
2884
2885[46:17]
2886challenge you kind of say it's you know
2887
2888[46:20]
2889oh yeah it's fun
2890
2891[46:24]
2892but it'd be interesting to hear if there
2893
2894[46:27]
2895are non fun parts that you really
2896
2897[46:29]
2898struggle with yes the fun comes with
2899
2900[46:32]
2901when I'm able to put together ideas of
2902
2903[46:36]
2904to two people who didn't know about each
2905
2906[46:38]
2907other and and and so I might be the
2908
2909[46:41]
2910first person that saw both of their
2911
2912[46:42]
2913ideas and so then you know then I get to
2914
2915[46:46]
2916make the synthesis and that gives me a
2917
2918[46:49]
2919chance to be creative but the dredge
2920
2921[46:52]
2922work is where I act I've got a chase
2923
2924[46:55]
2925everything down to its root this leads
2926
2927[46:58]
2928me into really interesting stuff i mean
2929
2930[47:00]
2931like i learned about sanskrit nice yeah
2932
2933[47:02]
2934and again you know I try to give credit
2935
2936[47:05]
2937to all the authors and so I write like
2938
2939[47:07]
2940so I write to people who know that the
2941
2942[47:11]
2943people thought as if they're dead I
2944
2945[47:13]
2946communicate this way I and I gotta get
2947
2948[47:17]
2949the math right and I got a tack all my
2950
2951[47:19]
2952programs try to find holes in them and I
2953
2954[47:23]
2955rewrite the programs over after I get a
2956
2957[47:25]
2958better idea
2959
2960[47:26]
2961is there ever dead-ends data and so yeah
2962
2963[47:29]
2964I throw stuff out yeah look one of the
2965
2966[47:32]
2967things that I spent a lot of time
2968
2969[47:35]
2970preparing a major example based on the
2971
2972[47:38]
2973game of baseball and I know a lot of
2974
2975[47:41]
2976people who for whom baseball is the most
2977
2978[47:44]
2979important thing in the world you know
2980
2981[47:46]
2982yes but it's but I also know a lot of
2983
2984[47:47]
2985people from cricket is the most
2986
2987[47:49]
2988important in the world or suck or
2989
2990[47:52]
2991something you know and and I realized
2992
2993[47:55]
2994that if if I had a big sample I mean it
2995
2996[47:58]
2997was gonna have a fold-out illustration
2998
2999[47:59]
3000and everything I was saying well what
3001
3002[48:01]
3003what am I really teaching about
3004
3005[48:02]
3006algorithms here where I had this this is
3007
3008[48:05]
3009this baseball example and if I was a
3010
3011[48:07]
3012person who who knew only cricket
3013
3014[48:10]
3015wouldn't think what would they think
3016
3017[48:12]
3018about this and and so I ripped the whole
3019
3020[48:14]
3021thing out but I you know I had I had a
3022
3023[48:17]
3024something that would really appeal to
3025
3026[48:19]
3027people who grew up with baseball as as
3028
3029[48:21]
3030has a major theme in their life which is
3031
3032[48:24]
3033a lot of people but yeah so I said on
3034
3035[48:27]
3036minority the small minority I took out
3037
3038[48:30]
3039bowling to
3040
3041[48:32]
3042even a smaller my noise what's the art
3043
3044[48:37]
3045in the art of programming why why is
3046
3047[48:42]
3048there of the few words in the title why
3049
3050[48:45]
3051is art one of them yeah well that's
3052
3053[48:47]
3054that's what I wrote my Turing lecture
3055
3056[48:49]
3057about and and so when people talk about
3058
3059[48:53]
3060art it really I mean what the word means
3061
3062[48:57]
3063is something that's not a nature so when
3064
3065[49:02]
3066you have artificial intelligence that
3067
3068[49:05]
3069that art come from the same root saying
3070
3071[49:09]
3072that this is something that was created
3073
3074[49:11]
3075by by human beings and then it's gotten
3076
3077[49:16]
3078a further meaning often a fine art which
3079
3080[49:19]
3081has this beauty to the to the mix and
3082
3083[49:21]
3084says you know we have things that are
3085
3086[49:23]
3087artistically done and and this means not
3088
3089[49:26]
3090only done by humans but also done in a
3091
3092[49:29]
3093way that's elegant and brings joy and
3094
3095[49:33]
3096and has has I guess what
3097
3098[49:39]
3099Tolstoy burrs dusky but anyway it it's
3100
3101[49:46]
3102that part that that says that it's done
3103
3104[49:49]
3105well as well as not only a different
3106
3107[49:53]
3108from nature in general then alright is
3109
3110[49:58]
3111what human beings are specifically good
3112
3113[50:01]
3114at and when they say hey like artificial
3115
3116[50:03]
3117intelligence well they're trying to
3118
3119[50:05]
3120mimic human beings but there's an
3121
3122[50:07]
3123element of fine art and beauty you are
3124
3125[50:11]
3126well that's what I that's what I try to
3127
3128[50:13]
3129also say that you can write a program
3130
3131[50:16]
3132and make a work of art so now in terms
3133
3134[50:22]
3135of surprising you know what ideas in
3136
3137[50:28]
3138writing from sort and search to the
3139
3140[50:32]
3141combinatorial algorithms what ideas have
3142
3143[50:35]
3144you come across that were particularly
3145
3146[50:40]
3147surprising to you that that change the
3148
3149[50:44]
3150way you see a space of
3151
3152[50:47]
3153I get a surprise every time I have a bug
3154
3155[50:49]
3156in my program but but that isn't really
3157
3158[50:52]
3159what your transformational surprises for
3160
3161[50:57]
3162example in volume for a I was especially
3163
3164[51:00]
3165surprised when I learned about data
3166
3167[51:03]
3168structure called B BDD boolean decision
3169
3170[51:06]
3171diagram because I sort of had the
3172
3173[51:10]
3174feeling that as an old-timer and you
3175
3176[51:15]
3177know I've been programming since this
3178
3179[51:16]
3180since the 50s and bTW these weren't
3181
3182[51:21]
3183invented until 1986 and here comes a
3184
3185[51:24]
3186brand new idea that revolutionized the
3187
3188[51:27]
3189way to represent a boolean function and
3190
3191[51:29]
3192boolean functions are so basic to all
3193
3194[51:32]
3195kinds of things in it I mean logically
3196
3197[51:37]
3198underlies it everything we can describe
3199
3200[51:40]
3201all of what we know in terms of logic
3202
3203[51:43]
3204somehow and and here and and
3205
3206[51:46]
3207propositional logic I thought that was
3208
3209[51:51]
3210cutting Dryden everything was known but
3211
3212[51:54]
3213but but he but here comes a Randy Bryant
3214
3215[51:59]
3216and oh and discovers that BDDs are
3217
3218[52:03]
3219incredibly powerful then then that's all
3220
3221[52:07]
3222so I that mean means I have a whole new
3223
3224[52:11]
3225section to the book that I never would
3226
3227[52:13]
3228have thought of until 1986
3229
3230[52:15]
3231not until 1990s when I went when people
3232
3233[52:18]
3234started to got to use it for you know
3235
3236[52:23]
3237billion dollar of applications and it
3238
3239[52:26]
3240was it was the standard way to design
3241
3242[52:28]
3243computers for a long time until until
3244
3245[52:31]
3246sad solvers came along when in the year
3247
3248[52:33]
32492000 so that's another great big
3250
3251[52:35]
3252surprise so uh a lot of these things
3253
3254[52:38]
3255have have totally changed the structure
3256
3257[52:40]
3258of my book and the middle third of
3259
3260[52:44]
3261volume four B's is about that solvers
3262
3263[52:46]
3264and that's
3265
3266[52:49]
3267300 plus pages which is which is all
3268
3269[52:53]
3270about material mostly about material
3271
3272[52:56]
3273that was discovered in this century and
3274
3275[52:59]
3276I had to start from scratch and meet all
3277
3278[53:03]
3279the people in the field and right
3280
3281[53:05]
3282I have 15 different sets Alvers that i
3283
3284[53:07]
3285wrote while preparing that seven of them
3286
3287[53:10]
3288are described in the book others were
3289
3290[53:13]
3291for my own experience so newly invented
3292
3293[53:16]
3294data structures or ways to represent a
3295
3296[53:20]
3297whole new class of algorithm calling you
3298
3299[53:22]
3300classified yeah and the interesting
3301
3302[53:24]
3303thing about the BD DS was that the
3304
3305[53:28]
3306theoretician started looking at it and
3307
3308[53:31]
3309started to describe all the things you
3310
3311[53:33]
3312couldn't do with BD DS and so they were
3313
3314[53:37]
3315getting a bad they were getting a bad
3316
3317[53:39]
3318name because you know okay they were
3319
3320[53:43]
3321they were useful but they didn't solve
3322
3323[53:46]
3324everything I'm sure that the
3325
3326[53:48]
3327theoreticians are in the next 10 years
3328
3329[53:51]
3330are gonna show why machine learning
3331
3332[53:54]
3333doesn't solve everything but I not only
3334
3335[53:58]
3336worried about the worst case I get a
3337
3338[54:00]
3339huge delight when I can actually solve a
3340
3341[54:02]
3342problem that I couldn't solve before
3343
3344[54:04]
3345yeah even though I can't solve the
3346
3347[54:07]
3348problem that's that it suggests as a
3349
3350[54:09]
3351further problem like I know that I'm Way
3352
3353[54:12]
3354better than I was before and so I found
3355
3356[54:14]
3357out that BD DS could do all kinds of
3358
3359[54:17]
3360miraculous things and so I had been
3361
3362[54:24]
3363quite a few years learning about the
3364
3365[54:28]
3366that territory so in general what brings
3367
3368[54:32]
3369you more pleasure in proving or showing
3370
3371[54:37]
3372a worst case analysis of an algorithm or
3373
3374[54:40]
3375showing a good average case or just
3376
3377[54:43]
3378showing a good case that you know
3379
3380[54:46]
3381something good pragmatically can be done
3382
3383[54:47]
3384with this algorithm yeah I like a good
3385
3386[54:50]
3387case that that is maybe only a million
3388
3389[54:53]
3390times faster than I was able to do
3391
3392[54:54]
3393before but and not worried about the
3394
3395[54:57]
3396fact that
3397
3398[54:58]
3399and that is still that is still gonna
3400
3401[55:01]
3402take too long if I double the size of
3403
3404[55:03]
3405the problem so that said you popularize
3406
3407[55:08]
3408the asymptotic notation for describing
3409
3410[55:10]
3411running time obviously in the analysis
3412
3413[55:14]
3414of algorithms worst cases such as such
3415
3416[55:17]
3417an important part do you see any aspects
3418
3419[55:20]
3420of that kind of analysis is lacking so
3421
3422[55:24]
3423and notation - well the main purpose you
3424
3425[55:28]
3426have notations that that help us for the
3427
3428[55:32]
3429problems we want to solve and so that
3430
3431[55:33]
3432they match our they match our intuitions
3433
3434[55:36]
3435and people who worked in number theory
3436
3437[55:38]
3438had used asymptotic notation in what
3439
3440[55:41]
3441Ennis in a certain way but it was only
3442
3443[55:44]
3444known to a small group of people and and
3445
3446[55:46]
3447I realized that in fact it was very
3448
3449[55:50]
3450useful to be able to have a notation for
3451
3452[55:52]
3453something that we don't know exactly
3454
3455[55:54]
3456what it is but we only know partial
3457
3458[55:56]
3459about it and so on stick so for example
3460
3461[56:00]
3462instead of Big O notation let's just
3463
3464[56:02]
3465let's just take us a much simpler
3466
3467[56:04]
3468notation where I say 0 or 1 or 0 1 or 2
3469
3470[56:09]
3471and suppose that suppose that when I had
3472
3473[56:13]
3474been in high school we would be allowed
3475
3476[56:15]
3477to put in the middle of our formula x +
3478
3479[56:18]
34800 1 or 2 equals y okay and then then we
3481
3482[56:24]
3483would learn how to multiply two such
3484
3485[56:27]
3486expressions together and and you know
3487
3488[56:30]
3489deal with them
3490
3491[56:32]
3492well the same thing Big O notation says
3493
3494[56:34]
3495here's something that's I'm not sure
3496
3497[56:38]
3498what it is but I know it's not too big I
3499
3500[56:40]
3501know it's not bigger than some constant
3502
3503[56:43]
3504times N squared or something like that
3505
3506[56:44]
3507fine so I write Big O of N squared and
3508
3509[56:47]
3510now I learned how to add Big O of N
3511
3512[56:49]
3513squared to Big O of N cubed and I know
3514
3515[56:51]
3516how to add Big O of N squared 2 plus 
3517
35181[56:54]
3519and square that and how to take
3520
3521[56:56]
3522logarithms and Exponential's to have big
3523
3524[56:58]
3525O's in the middle of them and that
3526
3527[57:01]
3528turned out to be hugely valuable in all
3529
3530[57:04]
3531of the work that I was trying to do is
3532
3533[57:06]
3534I'm trying to figure out how good
3535
3536[57:08]
3537so I have there been algorithms in your
3538
3539[57:12]
3540journey that perform very differently in
3541
3542[57:15]
3543practice than they do in theory well the
3544
3545[57:19]
3546worst case of a comet our logarithm is
3547
3548[57:21]
3549almost always horrible but but we have
3550
3551[57:26]
3552sad solvers that are solving where one
3553
3554[57:28]
3555of the one of the last exercises in that
3556
3557[57:31]
3558part of my book was to figure out a
3559
3560[57:34]
3561problem that has a hundred variables
3562
3563[57:37]
3564that's that's difficult for us at solver
3565
3566[57:40]
3567but uh but you would think that a
3568
3569[57:42]
3570problem with the hundred boolean
3571
3572[57:44]
3573variables has required to do 2 to the
3574
3575[57:47]
3576100th operations because that's the
3577
3578[57:51]
3579number of possibilities when you have
3580
3581[57:53]
3582200 boolean variables in 2 to the 100th
3583
3584[57:56]
3585to the 100th is way bigger than then we
3586
3587[57:59]
3588can handle 10 to the 17th is a lot
3589
3590[58:02]
3591you've mentioned over the past few years
3592
3593[58:04]
3594that you believe P may be equal to NP
3595
3596[58:07]
3597but that it's not really you know
3598
3599[58:11]
3600somebody does prove that P equals NP it
3601
3602[58:14]
3603will not directly lead to an actual
3604
3605[58:16]
3606algorithm to solve difficult problems
3607
3608[58:19]
3609can you explain your intuition here has
3610
3611[58:21]
3612it been changed and in general on the
3613
3614[58:24]
3615difference between easy and difficult
3616
3617[58:26]
3618problems of P and NP and so on yes so
3619
3620[58:29]
3621the popular idea is if an algorithm
3622
3623[58:33]
3624exists then somebody will find it and
3625
3626[58:38]
3627it's just a matter of writing it down
3628
3629[58:42]
3630one point well but many more algorithms
3631
3632[58:47]
3633exist than anybody can end understand or
3634
3635[58:51]
3636ever make you discover yeah because
3637
3638[58:53]
3639they're just way beyond human
3640
3641[58:55]
3642comprehension of the total number of
3643
3644[58:57]
3645algorithms is more than mind-boggling
3646
3647[59:03]
3648so so we have situations now where we
3649
3650[59:06]
3651know that algorithm exists but we don't
3652
3653[59:09]
3654know we don't the foggiest idea what the
3655
3656[59:11]
3657algorithms are there's there are simple
3658
3659[59:14]
3660examples based on on game playing where
3661
3662[59:18]
3663you have
3664
3665[59:20]
3666where you say well there must be an
3667
3668[59:23]
3669algorithm that exists to win in the game
3670
3671[59:25]
3672of hex because for the first player to
3673
3674[59:28]
3675win in the game of hex because hex is
3676
3677[59:31]
3678always either an a win for the first
3679
3680[59:33]
3681player of the second player well what's
3682
3683[59:35]
3684the game of hack there's a game of hex
3685
3686[59:37]
3687which is which based on putting pebbles
3688
3689[59:39]
3690onto a hexagonal board and and the white
3691
3692[59:42]
3693player tries to get a light path from
3694
3695[59:45]
3696left to right and the black player tries
3697
3698[59:47]
3699to get a black path from bottom to top
3700
3701[59:48]
3702and how does capture occur just so and
3703
3704[59:51]
3705and and there's no capture you just put
3706
3707[59:53]
3708levels down what one at a time but
3709
3710[59:56]
3711there's no drawers because they after
3712
3713[59:58]
3714all the white and black are played
3715
3716[59:59]
3717there's either going to be a white path
3718
3719[60:01]
3720across from each to west or a black path
3721
3722[60:03]
3723from from bottom to top so there's
3724
3725[60:06]
3726always you know it's the perfect
3727
3728[60:08]
3729information game and people people play
3730
3731[60:10]
3732take turns like like tic-tac-toe and hex
3733
3734[60:16]
3735or it can be different sizes but we
3736
3737[60:19]
3738there's no possibility of a draw and
3739
3740[60:21]
3741player to move one at a time and so it's
3742
3743[60:24]
3744got to be either a first player win or a
3745
3746[60:26]
3747second player win
3748
3749[60:27]
3750mathematically you follow out all the
3751
3752[60:30]
3753trees and and either either there's
3754
3755[60:33]
3756always the win for the percolator
3757
3758[60:34]
3759second player okay and it's finite the
3760
3761[60:37]
3762game is finite so there's an algorithm
3763
3764[60:39]
3765that will decide you can show it has to
3766
3767[60:42]
3768be one of the other because the second
3769
3770[60:44]
3771player could mimic the first player with
3772
3773[60:47]
3774kind of a pairing strategy and so you
3775
3776[60:50]
3777can show that it has to be what it has
3778
3779[60:55]
3780to be one or that but we don't know any
3781
3782[60:57]
3783algorithm no way there there a case
3784
3785[61:01]
3786where you can prove the existence of the
3787
3788[61:05]
3789solution but we but nobody knows anyway
3790
3791[61:07]
3792how to find it but more like the
3793
3794[61:09]
3795algorithm question there's a very
3796
3797[61:13]
3798powerful theorem and graph theory by
3799
3800[61:15]
3801Robinson to see more that says that
3802
3803[61:18]
3804every class of graphs that is closed
3805
3806[61:23]
3807under taking minors
3808
3809[61:26]
3810has a polynomial time algorithm to
3811
3812[61:29]
3813determine whether it's in this class or
3814
3815[61:30]
3816not now a class of graphs for example
3817
3818[61:32]
3819planar graphs these are graphs that you
3820
3821[61:34]
3822can draw in a plane without crossing
3823
3824[61:36]
3825lines and and a planar graph is close
3826
3827[61:39]
3828taking minors means that you can shrink
3829
3830[61:42]
3831an edging into a point or you can delete
3832
3833[61:46]
3834an edge and so you start with a planar
3835
3836[61:50]
3837graph and drink any edge to a point is
3838
3839[61:52]
3840still planar deleting edges to a planner
3841
3842[61:55]
3843okay now but there are millions of
3844
3845[62:02]
3846different ways to describe family of
3847
3848[62:08]
3849graph that still is remains the same
3850
3851[62:12]
3852undertaking minor and Robertson Nassim
3853
3854[62:15]
3855are proved that any such family of
3856
3857[62:17]
3858graphs there is a finite number of
3859
3860[62:20]
3861minimum graphs that are obstructions so
3862
3863[62:26]
3864that if it's not in the family then then
3865
3866[62:31]
3867it has to contain then there has to be a
3868
3869[62:34]
3870way to shrink it down and until you get
3871
3872[62:37]
3873one of these bad minimum graphs that's
3874
3875[62:39]
3876not in the family for in plate case for
3877
3878[62:42]
3879planar graph the minimum graph is a is a
3880
3881[62:45]
3882five-pointed star where there everything
3883
3884[62:47]
3885pointed to another and the minimum graph
3886
3887[62:49]
3888consisting of trying to connect three
3889
3890[62:51]
3891utilities to three houses without
3892
3893[62:53]
3894crossing lines and so there are two
3895
3896[62:55]
3897there are two bad graphs that are not
3898
3899[62:57]
3900planar and every every non planar graph
3901
3902[63:00]
3903contains one of these two bad graphs by
3904
3905[63:03]
3906by shrinking and he said again so he
3907
3908[63:09]
3909proved that there's a finite number of
3910
3911[63:11]
3912these bad guys always a finite know
3913
3914[63:13]
3915somebody says here's a family it's hard
3916
3917[63:15]
3918to believe and they present its sequence
3919
3920[63:20]
3921of 20 papers I mean in there it's deep
3922
3923[63:22]
3924work but it you know it's because that's
3925
3926[63:25]
3927for any arbitrary class so it's for any
3928
3929[63:28]
3930arbitrary class that's closed under
3931
3932[63:29]
3933taking minors that's closed under maybe
3934
3935[63:32]
3936I'm not understanding because it seems
3937
3938[63:34]
3939like a lot of them are closed
3940
3941[63:35]
3942taking minors almost all the important
3943
3944[63:37]
3945classes of graphs are
3946
3947[63:39]
3948there are tons of of such graphs but
3949
3950[63:42]
3951also hundreds of them that arise in
3952
3953[63:46]
3954applications like I have a book over
3955
3956[63:48]
3957here called classes of graphs and then
3958
3959[63:51]
3960and it it's amazing how many different
3961
3962[63:56]
3963classes people have looked at so why do
3964
3965[63:59]
3966you bring up this theorem lower this
3967
3968[64:01]
3969proof so you know there are lots of
3970
3971[64:04]
3972algorithms that that are known for
3973
3974[64:06]
3975special class of graphs for example if I
3976
3977[64:08]
3978have a certain if I have a chordal graph
3979
3980[64:10]
3981then I can color it efficiently if I
3982
3983[64:13]
3984have some kinds of graphs it'll make a
3985
3986[64:16]
3987great Network very soon like you'd like
3988
3989[64:19]
3990to test you somebody gives you a graph
3991
3992[64:22]
3993that's always it in this family of grass
3994
3995[64:24]
3996if so then I hope then I can I can go to
3997
3998[64:27]
3999the library and find an algorithm that's
4000
4001[64:29]
4002gonna solve my problem on that graph
4003
4004[64:32]
4005okay so we we have we want to have a
4006
4007[64:35]
4008graph that says number than that says
4009
4010[64:41]
4011give me a graph I'll tell you whether
4012
4013[64:43]
4014it's and whether it's in this family or
4015
4016[64:46]
4017not okay and so all I have to do is test
4018
4019[64:51]
4020whether or not that does this given
4021
4022[64:54]
4023graph have a minor that's one of the bad
4024
4025[64:56]
4026ones a minor is is everything you can
4027
4028[64:58]
4029get by shrinking and removing edges and
4030
4031[65:01]
4032given any minor there's a polynomial
4033
4034[65:03]
4035time algorithm saying I can tell whether
4036
4037[65:06]
4038this is a minor of you and there's a
4039
4040[65:09]
4041finite number of bad cases so I just
4042
4043[65:12]
4044tried you know does it have this bad
4045
4046[65:13]
4047case by polynomial time I got the answer
4048
4049[65:16]
4050does he have this bad case probably time
4051
4052[65:18]
4053I got the answer a total polynomial time
4054
4055[65:22]
4056and so I've solved the problem however
4057
4058[65:25]
4059all we know is that the number of minors
4060
4061[65:27]
4062is finite we don't know what we might
4063
4064[65:31]
4065only know one or two of those minors but
4066
4067[65:32]
4068we don't know that if we got it if we
4069
4070[65:34]
4071got 20 of them we don't know there might
4072
4073[65:36]
4074be 20 125 the Halloween all we know is
4075
4076[65:40]
4077that is that it's finite so here we have
4078
4079[65:43]
4080a polynomial time algorithm that we
4081
4082[65:44]
4083don't know
4084
4085[65:45]
4086mm-hm that's a really great example of
4087
4088[65:47]
4089what you worry about or why you think P
4090
4091[65:50]
4092equals NP won't be useful
4093
4094[65:52]
4095but still why do you hold the intuition
4096
4097[65:56]
4098that P equals NP because you have to
4099
4100[66:02]
4101rule out so many possible algorithms
4102
4103[66:05]
4104have been not working you know you can
4105
4106[66:11]
4107you can take the graph and you can
4108
4109[66:13]
4110represent it as in terms of certain
4111
4112[66:17]
4113prime numbers and then you can multiply
4114
4115[66:19]
4116those together and then you can then you
4117
4118[66:21]
4119can take the bitwise and and and you
4120
4121[66:25]
4122know and construct some certain constant
4123
4124[66:29]
4125in polynomial time and then that's you
4126
4127[66:31]
4128know perfectly valid algorithm and that
4129
4130[66:33]
4131there's so many algorithms of that kind
4132
4133[66:36]
4134a lot of times we see random you take
4135
4136[66:42]
4137data and and and we get coincidences
4138
4139[66:46]
4140that that that some fairly random
4141
4142[66:49]
4143looking number actually is useful
4144
4145[66:51]
4146because because it god it happens to it
4147
4148[66:57]
4149happens to self it happens to solve a
4150
4151[66:59]
4152problem just because you know there's
4153
4154[67:02]
4155there's so many hairs on your head
4156
4157[67:05]
4158but it seems like unlikely that two
4159
4160[67:10]
4161people are going to have the same number
4162
4163[67:11]
4164of hairs on their head but but they're
4165
4166[67:16]
4167obvious but you can count how many
4168
4169[67:17]
4170people there are and how many hairs on
4171
4172[67:19]
4173there so there must be people walking
4174
4175[67:21]
4176around in the country to have the same
4177
4178[67:23]
4179number of hairs on their head well
4180
4181[67:24]
4182that's the kind of a coincidence that
4183
4184[67:26]
4185you might say also you know this this
4186
4187[67:29]
4188particular combination of operations
4189
4190[67:31]
4191just happens to prove that a graph is
4192
4193[67:34]
4194has a Hamiltonian path and I see lots of
4195
4196[67:37]
4197cases where unexpected things happen
4198
4199[67:41]
4200when you have enough enough
4201
4202[67:42]
4203possibilities but because the space of
4204
4205[67:45]
4206possibility is so huge I have to rule
4207
4208[67:48]
4209them all out and so that's the reason
4210
4211[67:50]
4212for my intuition is good by no means
4213
4214[67:52]
4215approve I mean some people say you know
4216
4217[67:56]
4218well P can't equal NP because you've had
4219
4220[67:59]
4221all these smart people you know the
4222
4223[68:03]
4224smartest designers of algorithms that
4225
4226[68:05]
4227have been
4228
4229[68:06]
4230wrecking their brains for years and
4231
4232[68:07]
4233years and and there's million-dollar
4234
4235[68:10]
4236prizes out there and you know none of
4237
4238[68:11]
4239them nobody has thought of the algorithm
4240
4241[68:16]
4242so it must must be no such job on the
4243
4244[68:19]
4245other hand I can use exactly the same
4246
4247[68:22]
4248logic and I can say well P must be equal
4249
4250[68:25]
4251to NP because there's so many smart
4252
4253[68:27]
4254people out here been trying to prove it
4255
4256[68:28]
4257unequal to NP and they've all failed you
4258
4259[68:32]
4260know this kind of reminds me of the
4261
4262[68:36]
4263discussion about the search for aliens
4264
4265[68:38]
4266they've been trying to look for them and
4267
4268[68:40]
4269we haven't found them yet therefore they
4270
4271[68:42]
4272don't exist
4273
4274[68:42]
4275yeah but you can show that there's so
4276
4277[68:45]
4278many planets out there that they very
4279
4280[68:46]
4281possibly could exist yeah and right and
4282
4283[68:50]
4284then there's also the possibility that
4285
4286[68:52]
4287that they exist but they they all
4288
4289[68:55]
4290discovered machine learning or something
4291
4292[68:57]
4293and and and then blew each other up well
4294
4295[69:01]
4296on that small quick danger
4297
4298[69:03]
4299let me ask do you think there's
4300
4301[69:04]
4302intelligent life out there in the
4303
4304[69:05]
4305universe I have no idea do you hope so
4306
4307[69:09]
4308do you think about it it I I don't I
4309
4310[69:12]
4311don't spend my time thinking about
4312
4313[69:14]
4314things that I could never know really
4315
4316[69:16]
4317and yet you do enjoy the fact that there
4318
4319[69:18]
4320are many things you don't know you do
4321
4322[69:20]
4323enjoy the mystery of things I enjoy the
4324
4325[69:24]
4326fact that there that I have limits yeah
4327
4328[69:26]
4329but I don't but but I don't take time to
4330
4331[69:31]
4332answer unsolvable questions I got it
4333
4334[69:35]
4335well because you've taken on some tough
4336
4337[69:38]
4338questions that may seem unsolvable you
4339
4340[69:40]
4341have taken on some tough questions and
4342
4343[69:42]
4344you seem unsolvable if there is because
4345
4346[69:44]
4347we are thrilled when I can get further
4348
4349[69:46]
4350than I ever thought I could right yeah
4351
4352[69:48]
4353but but I don't what much like was
4354
4355[69:51]
4356religion these I'm glad the dirt that
4357
4358[69:54]
4359that there are no proof that God exists
4360
4361[69:57]
4362or not I mean I think it would spoil the
4363
4364[69:59]
4365mystery it it would be too dull yeah so
4366
4367[70:05]
4368to quickly talk about the other art of
4369
4370[70:08]
4371artificial intelligence
4372
4373[70:10]
4374what is if you what's your view
4375
4376[70:13]
4377you know artificial intelligence
4378
4379[70:15]
4380community has developed as part of
4381
4382[70:17]
4383computer science and in parallel with
4384
4385[70:18]
4386computer science
4387
4388[70:19]
4389since the 60s what's your view of the AI
4390
4391[70:22]
4392community from the 60s to now so all the
4393
4394[70:27]
4395way through it was the people who were
4396
4397[70:29]
4398inspired by trying to mimic intelligence
4399
4400[70:35]
4401or to do things that that were somehow
4402
4403[70:37]
4404the greatest achievements of
4405
4406[70:39]
4407intelligence that had been inspiration
4408
4409[70:41]
4410to people who have pushed the envelope
4411
4412[70:43]
4413of computer science maybe more than any
4414
4415[70:48]
4416other group of people so it's all the
4417
4418[70:51]
4419way through it's been a great source of
4420
4421[70:53]
4422of good problems to to sink teeth into
4423
4424[70:59]
4425and and getting getting partial answers
4426
4427[71:06]
4428and then more and more successful
4429
4430[71:08]
4431answers over the year so this has this
4432
4433[71:12]
4434has been the inspiration for lots of the
4435
4436[71:14]
4437great discoveries of computer science
4438
4439[71:16]
4440are you yourself captivated by the
4441
4442[71:18]
4443possibility of creating of algorithms
4444
4445[71:20]
4446having echoes of intelligence in them
4447
4448[71:26]
4449not as much as most of the people in the
4450
4451[71:29]
4452field I guess I would say but but that's
4453
4454[71:32]
4455not to say that they're wrong or that
4456
4457[71:34]
4458it's just you asked about my own
4459
4460[71:36]
4461personal preferences and yeah but but
4462
4463[71:39]
4464the thing that I that I worry about is
4465
4466[71:47]
4467when people start believing that they've
4468
4469[71:49]
4470actually succeeded and because the seems
4471
4472[71:56]
4473to me this huge gap between really
4474
4475[71:59]
4476understanding something and being able
4477
4478[72:02]
4479to pretend to understand something and
4480
4481[72:05]
4482give these give the illusion of
4483
4484[72:06]
4485understanding something do you think
4486
4487[72:08]
4488it's possible to create without
4489
4490[72:10]
4491understanding yeah
4492
4493[72:12]
4494so to uh I do that all the time to run I
4495
4496[72:15]
4497mean that's why I use random members I
4498
4499[72:18]
4500like yeah but I but but there's there's
4501
4502[72:23]
4503still what this great gap I don't know
4504
4505[72:25]
4506certain it's impossible but I'm like but
4507
4508[72:27]
4509I don't see a anything coming any closer
4510
4511[72:31]
4512to really
4513
4514[72:33]
4515the the kind of stuff that I would
4516
4517[72:36]
4518consider intelligence say you've
4519
4520[72:39]
4521mentioned something that on that line of
4522
4523[72:41]
4524thinking which I very much agree with so
4525
4526[72:45]
4527the art of computer programming as the
4528
4529[72:48]
4530book is focused on single processor
4531
4532[72:51]
4533algorithms and for the most part and you
4534
4535[72:56]
4536mentioned that's only because I set the
4537
4538[72:59]
4539table of contents in 1962 you have to
4540
4541[73:01]
4542remember for sure there's no I'm glad I
4543
4544[73:05]
4545didn't wait until 1965 or one book maybe
4546
4547[73:11]
4548will touch in the Bible but one book
4549
4550[73:13]
4551can't always cover the entirety of
4552
4553[73:15]
4554everything so I'm glad yeah I'm glad the
4555
4556[73:20]
4557the table of contents for the art of
4558
4559[73:24]
4560computer programming is what it is but
4561
4562[73:26]
4563you did mention that that you thought
4564
4565[73:29]
4566that an understanding of the way ant
4567
4568[73:30]
4569colonies are able to perform incredibly
4570
4571[73:33]
4572organized tasks might well be the key to
4573
4574[73:36]
4575understanding human cognition
4576
4577[73:38]
4578so these fundamentally distributed
4579
4580[73:40]
4581systems so what do you think is the
4582
4583[73:43]
4584difference between the way Don Knuth
4585
4586[73:46]
4587would sort a list and an ant colony
4588
4589[73:48]
4590would sort a list or performing
4591
4592[73:51]
4593algorithm sorting a list isn't same as
4594
4595[73:54]
4596cognition though but but I know what
4597
4598[73:57]
4599you're getting at is well the advantage
4600
4601[74:00]
4602of ant colony at least we can see what
4603
4604[74:04]
4605they're doing we we know which ant has
4606
4607[74:06]
4608talked to which other ant and and and
4609
4610[74:08]
4611and it's much harder with the quick
4612
4613[74:11]
4614brains to just to know how to what
4615
4616[74:14]
4617extent of neurons are passing signal so
4618
4619[74:18]
4620I understand that aunt Connie might be a
4621
4622[74:21]
4623if they have the secret of cognition
4624
4625[74:24]
4626think of an ant colony as a cognitive
4627
4628[74:27]
4629single being rather than as a colony of
4630
4631[74:31]
4632lots of different ants I mean just like
4633
4634[74:32]
4635the cells of our brain are and and the
4636
4637[74:37]
4638microbiome and all that is interacting
4639
4640[74:41]
4641entities but but somehow I consider
4642
4643[74:46]
4644myself to be
4645
4646[74:46]
4647single person well you know aunt Connie
4648
4649[74:50]
4650you can say might be cognitive is
4651
4652[74:54]
4653somehow and it's yeah I mean you know I
4654
4655[74:57]
4656okay I like I smash a certain aunt and
4657
4658[75:03]
4659mmm that's stung what was that right you
4660
4661[75:06]
4662know but if we're going to crack the the
4663
4664[75:08]
4665the secret of cognition it might be that
4666
4667[75:11]
4668we could do so by but my psyche note how
4669
4670[75:16]
4671ants do it because we have a better
4672
4673[75:18]
4674chance to measure and they're
4675
4676[75:19]
4677communicating by pheromones and by
4678
4679[75:21]
4680touching each other and sight but but
4681
4682[75:24]
4683not by much more subtle phenomenon Mike
4684
4685[75:27]
4686electric currents going through but even
4687
4688[75:30]
4689a simpler version of that what are your
4690
4691[75:32]
4692thoughts of maybe Conway's Game of Life
4693
4694[75:34]
4695okay so Conway's Game of Life is is able
4696
4697[75:39]
4698to simulate any any computable process
4699
4700[75:43]
4701and any deterministic process is like
4702
4703[75:47]
4704how you went there I mean that's not its
4705
4706[75:49]
4707most powerful thing I would say I mean
4708
4709[75:54]
4710you can simulate it but the magic is
4711
4712[75:58]
4713that the individual units are
4714
4715[76:00]
4716distributed yes and extremely simple yes
4717
4718[76:03]
4719we can we understand exactly what the
4720
4721[76:06]
4722primitives are the permit is the just
4723
4724[76:07]
4725like with the anthology even simple but
4726
4727[76:09]
4728if we but still it doesn't say that I
4729
4730[76:12]
4731understand I understand life I mean I
4732
4733[76:16]
4734understand it it gives me an it gives me
4735
4736[76:23]
4737a better insight into what does it mean
4738
4739[76:24]
4740to to have a deterministic universe what
4741
4742[76:30]
4743does it mean to to have free choice for
4744
4745[76:34]
4746example do you think God plays dice yes
4747
4748[76:38]
4749I don't see any reason why God should be
4750
4751[76:40]
4752forbidden from using the most efficient
4753
4754[76:42]
4755ways to to to I mean we we know that
4756
4757[76:51]
4758dice are extremely important and
4759
4760[76:53]
4761inefficient algorithms there are things
4762
4763[76:55]
4764like that couldn't be done well without
4765
4766[76:57]
4767randomness and so I don't see any reason
4768
4769[76:59]
4770why
4771
4772[77:00]
4773my god should be prohibited but when the
4774
4775[77:03]
4776when the algorithm requires it
4777
4778[77:05]
4779you don't see why the know the physics
4780
4781[77:09]
4782should constrain it yeah
4783
4784[77:11]
4785so in 2001 you gave a series of lectures
4786
4787[77:13]
4788at MIT about religion and science
4789
4790[77:17]
4791well that would 1999 but you published
4792
4793[77:19]
4794the book came out in Cooper so in 1999
4795
4796[77:23]
4797you spent a little bit of time in Boston
4798
4799[77:26]
4800enough to give those lectures yeah and I
4801
4802[77:31]
4803read in the 2001 version that most of it
4804
4805[77:36]
4806it's quite fascinating read I recommend
4807
4808[77:37]
4809people its transcription of your
4810
4811[77:39]
4812lectures so what did you learn about how
4813
4814[77:43]
4815ideas get started and grow from studying
4816
4817[77:45]
4818the history of the Bible sieve
4819
4820[77:47]
4821rigorously studied a very particular
4822
4823[77:49]
4824part of the Bible what did you learn
4825
4826[77:52]
4827from this process about the way us human
4828
4829[77:54]
4830beings as a society develop and grow
4831
4832[77:57]
4833ideas share ideas and I'm by those idea
4834
4835[78:01]
4836I I tried to summarize that I wouldn't
4837
4838[78:05]
4839say that I that I learned a great deal
4840
4841[78:08]
4842of really definite things like right
4843
4844[78:10]
4845where I could make conclusions but I
4846
4847[78:12]
4848learned more about what I don't know you
4849
4850[78:15]
4851have a complex subject which is really
4852
4853[78:17]
4854beyond human understanding so so we give
4855
4856[78:22]
4857up on saying I'm never going to get to
4858
4859[78:24]
4860the end of the road and I'm never going
4861
4862[78:25]
4863to understand it but you say but but
4864
4865[78:27]
4866maybe it might be good for me to to get
4867
4868[78:31]
4869closer and closer and learn more about
4870
4871[78:33]
4872more and more about something and so you
4873
4874[78:35]
4875know oh how can I do that efficiently
4876
4877[78:38]
4878and the answer is well use randomness
4879
4880[78:43]
4881and so to try a random subset of the
4882
4883[78:49]
4884that that is within my grasp and and and
4885
4886[78:53]
4887and study that in detail instead of just
4888
4889[78:57]
4890studying parts that somebody tells me to
4891
4892[79:00]
4893study or instead of studying nothing
4894
4895[79:03]
4896because it's too hard so I I i decided
4897
4898[79:11]
4899for my own amusement that one ones that
4900
4901[79:14]
4902I would I would take a subset of the of
4903
4904[79:18]
4905the verses of the Bible
4906
4907[79:21]
4908and I would try to find out what the
4909
4910[79:25]
4911best thinkers have said about that small
4912
4913[79:28]
4914subset and I had had about let's say 660
4915
4916[79:32]
4917verses out of out of 3,000 I think it's
4918
4919[79:35]
4920one out of 500 or something like this
4921
4922[79:37]
4923and so then I went to the libraries
4924
4925[79:39]
4926which which are well indexed uh you can
4927
4928[79:42]
4929you you know I spent for example at at
4930
4931[79:46]
4932Boston Public Library I I would go once
4933
4934[79:49]
4935a week for a year and I went to I went I
4936
4937[79:54]
4938have done time stuff and over Harvard
4939
4940[79:57]
4941library to look at this yes that weren't
4942
4943[80:01]
4944in the Boston Public where they where
4945
4946[80:04]
4947scholars had looked at and you can call
4948
4949[80:06]
4950in the eight and you can go down the
4951
4952[80:08]
4953shelves and and you can pretty you can
4954
4955[80:11]
4956look at the index and say oh there it is
4957
4958[80:12]
4959this verse I mentioned anywhere in this
4960
4961[80:15]
4962book if so look at page 105 so I was
4963
4964[80:18]
4965like I could learn not only about the
4966
4967[80:20]
4968Bible but about the secondary literature
4969
4970[80:22]
4971about the Bible the things that scholars
4972
4973[80:24]
4974have written about it and so that that
4975
4976[80:26]
4977gave me a way to uh to zoom in on parts
4978
4979[80:32]
4980of the things so that I could get more
4981
4982[80:34]
4983more insight and and so I look at it as
4984
4985[80:37]
4986a way of giving me some firm pegs which
4987
4988[80:43]
4989icon which I could hang pieces of
4990
4991[80:44]
4992information but not as as things where I
4993
4994[80:47]
4995would say and therefore this is true in
4996
4997[80:50]
4998this random approach of sampling the
4999
5000[80:54]
5001Bible what did you learn about the the
5002
5003[80:58]
5004most you know central oh one of the
5005
5006[81:03]
5007biggest accumulation of ideas you know
5008
5009[81:05]
5010to me that the that the main thrust was
5011
5012[81:08]
5013not the one that most people think of as
5014
5015[81:11]
5016saying you know you know don't have sex
5017
5018[81:13]
5019or something like this but that the main
5020
5021[81:16]
5022thrust was to try to to try to figure
5023
5024[81:22]
5025out how to live in harmony
5026
5027[81:24]
5028with God's wishes I'm assuming that God
5029
5030[81:27]
5031exists and I say I'm glad that I that
5032
5033[81:31]
5034there's no way to prove this because
5035
5036[81:32]
5037that would that would I would run
5038
5039[81:36]
5040through the proof once and then I'd
5041
5042[81:37]
5043forget it and and it would and and I
5044
5045[81:40]
5046would never just speculate about
5047
5048[81:44]
5049spiritual things and mysteries otherwise
5050
5051[81:48]
5052and I think my life would be very
5053
5054[81:50]
5055incomplete so I so I'm assuming that God
5056
5057[81:55]
5058exists but it if but a lot of things the
5059
5060[81:59]
5061people say God doesn't exist but that's
5062
5063[82:02]
5064still important to them and so in a way
5065
5066[82:04]
5067in a way that might still be other God
5068
5069[82:07]
5070is there or not in some sense so it it
5071
5072[82:11]
5073guys important to them it's one of the
5074
5075[82:14]
5076one of the verses I studied act is you
5077
5078[82:17]
5079can interpret as saying you know it's
5080
5081[82:19]
5082much better to be an atheist that not to
5083
5084[82:22]
5085care at all so I would say it's yeah
5086
5087[82:26]
5088it's similar to the P equals NP
5089
5090[82:27]
5091discussion yeah you you mentioned a
5092
5093[82:30]
5094mental exercise that I'd love it if you
5095
5096[82:33]
5097could partake in yourself a mental
5098
5099[82:37]
5100exercise of being God and so how would
5101
5102[82:39]
5103you if you were God dot Knuth how would
5104
5105[82:42]
5106you present yourself to the people of
5107
5108[82:43]
5109Earth you mentioned your love of
5110
5111[82:47]
5112literature and there was it there's this
5113
5114[82:48]
5115book that would that really uh I can
5116
5117[82:50]
5118recommend to you if I can't think yeah
5119
5120[82:53]
5121the title I think is blasphemy it talks
5122
5123[82:56]
5124about God revealing himself through a
5125
5126[82:58]
5127computer in in in Los Alamos and and it
5128
5129[83:06]
5130it's the only book that I've ever read
5131
5132[83:09]
5133where the punchline was really the very
5134
5135[83:13]
5136last word of the book and it explained
5137
5138[83:16]
5139the whole idea of the book and so I
5140
5141[83:18]
5142don't want to give that away but it but
5143
5144[83:21]
5145it's really very much about this
5146
5147[83:22]
5148question that that she raised
5149
5150[83:26]
5151but but suppose God said okay that my
5152
5153[83:31]
5154previous on means of communication with
5155
5156[83:35]
5157the world are and not the best for the
5158
5159[83:37]
516021st century so what should I do now and
5161
5162[83:40]
5163and and it's conceivable that that it
5164
5165[83:45]
5166would that that God would choose the way
5167
5168[83:48]
5169that's described in this book and
5170
5171[83:50]
5172another way to look at this exercise is
5173
5174[83:52]
5175looking at the human mind looking at the
5176
5177[83:55]
5178human spirit the human life in a
5179
5180[83:57]
5181systematic way I think it mostly you
5182
5183[84:00]
5184want to learn humility you want to
5185
5186[84:02]
5187realize that once we solve one problem
5188
5189[84:03]
5190that doesn't mean it worked at all so no
5191
5192[84:06]
5193other problems are going to drop out and
5194
5195[84:08]
5196and and and we have to realize that that
5197
5198[84:15]
5199that there are there are things beyond
5200
5201[84:17]
5202our beyond our ability I see hubris all
5203
5204[84:24]
5205around yeah well said if you were to run
5206
5207[84:29]
5208program analysis on your own life how
5209
5210[84:33]
5211did you do in terms of correctness
5212
5213[84:35]
5214running time resource use asymptotically
5215
5216[84:39]
5217speaking of course okay yeah well I
5218
5219[84:42]
5220would say that question has not been
5221
5222[84:46]
5223asked me before and i i i started out
5224
5225[84:57]
5226with library subroutines and and
5227
5228[85:04]
5229learning how to be a automaton that was
5230
5231[85:07]
5232obedient and i had the great advantage
5233
5234[85:10]
5235that i didn't have anybody to blame for
5236
5237[85:14]
5238my failures if I started getting not
5239
5240[85:20]
5241understanding something I I knew that I
5242
5243[85:23]
5244should stop playing ping pong and that
5245
5246[85:24]
5247was that into it was my fault that I was
5248
5249[85:26]
5250that I wasn't studying hard enough or
5251
5252[85:28]
5253something rather than that somebody was
5254
5255[85:30]
5256discriminating against me in some way
5257
5258[85:32]
5259and
5260
5261[85:33]
5262I don't know how to avoid this the
5263
5264[85:37]
5265existence of biases in the world but i
5266
5267[85:38]
5268but i but i know that that's an extra
5269
5270[85:41]
5271burden that i didn't have to suffer from
5272
5273[85:45]
5274and and and then i I found the from from
5275
5276[85:54]
5277parents I learned the idea of of
5278
5279[85:58]
5280altruist to other people as being more
5281
5282[86:03]
5283important than then when I get out of
5284
5285[86:08]
5286stuff myself I you know that I need to I
5287
5288[86:11]
5289need to be happy enough enough in order
5290
5291[86:16]
5292to be able to speed up service but I
5293
5294[86:18]
5295thought but I you know but I I came to a
5296
5297[86:21]
5298philosophy for finally that that I
5299
5300[86:23]
5301phrased as point eight is enough there
5302
5303[86:29]
5304was a TV show once called hate is enough
5305
5306[86:31]
5307which was about a you know somebody had
5308
5309[86:33]
5310eight kids but but I I say point a is
5311
5312[86:38]
5313enough which means if I can have a way
5314
5315[86:41]
5316of rating happiness I think it's good
5317
5318[86:45]
5319design that to have to have an organism
5320
5321[86:51]
5322that's happy about eighty percent of the
5323
5324[86:53]
5325time and if it was a hundred percent of
5326
5327[86:58]
5328the time it would be like every like
5329
5330[87:00]
5331everybody's on drugs and and never and
5332
5333[87:02]
5334and and and everything collapses nothing
5335
5336[87:06]
5337works because everybody's just too happy
5338
5339[87:08]
5340do you think you've achieved that point
5341
5342[87:10]
5343eight optimal work there are times when
5344
5345[87:13]
5346I when I'm down and I you know and I
5347
5348[87:17]
5349think I mean I know that I'm chemically
5350
5351[87:19]
5352right I know that I've actually been
5353
5354[87:21]
5355programmed to be I to be depressed a
5356
5357[87:26]
5358certain amount of time and and and if
5359
5360[87:28]
5361that gets out of kilter and I'm more
5362
5363[87:30]
5364depressed and you know sometimes like
5365
5366[87:32]
5367like I find myself trying to say now who
5368
5369[87:34]
5370should I be mad at today there must be a
5371
5372[87:36]
5373reason why
5374
5375[87:38]
5376but I but then I realize you know it's
5377
5378[87:41]
5379just my it's just my chemistry telling
5380
5381[87:43]
5382me that I'm supposed to be mad at
5383
5384[87:45]
5385somebody and so and so I triggered up
5386
5387[87:47]
5388say okay go to sleep and get better but
5389
5390[87:50]
5391but if I'm but if I'm not a hundred
5392
5393[87:53]
5394percent happy that doesn't mean that I
5395
5396[87:56]
5397should find somebody that that's
5398
5399[87:57]
5400screaming and and try to size them up
5401
5402[88:00]
5403but I'd be like I'm saying you know okay
5404
5405[88:04]
5406I'm not 100% happy but but I'm happy
5407
5408[88:08]
5409enough to death to be a you know part of
5410
5411[88:11]
5412a sustainable situation so so that's
5413
5414[88:15]
5415kind of the numerical analysis I do you
5416
5417[88:21]
5418invert stores the human life is a point
5419
5420[88:24]
5421eight yeah I hope it's okay to talk
5422
5423[88:27]
5424about as you talked about previously in
5425
5426[88:29]
5427two thousand six six you were diagnosed
5428
5429[88:32]
5430with prostate cancer has that encounter
5431
5432[88:35]
5433with mortality changed you in some way
5434
5435[88:38]
5436or the way you see the world the first
5437
5438[88:42]
5439encounter with mortality with Mike when
5440
5441[88:44]
5442my dad died and I I went through a month
5443
5444[88:47]
5445when I sort of came to kink you know be
5446
5447[88:56]
5448comfortable with the fact that I was
5449
5450[88:57]
5451going to die someday and during that
5452
5453[88:59]
5454month I don't know I I felt okay but I
5455
5456[89:06]
5457couldn't sing and you know I and I and I
5458
5459[89:11]
5460couldn't do original research either
5461
5462[89:14]
5463like tighten right I sort of remember
5464
5465[89:16]
5466after three or four weeks the first time
5467
5468[89:19]
5469I started having a technical thought
5470
5471[89:21]
5472that made sense and was maybe slightly
5473
5474[89:23]
5475creative I could sort of feel they know
5476
5477[89:25]
5478that and that something was starting to
5479
5480[89:28]
5481move again but that was you know so I
5482
5483[89:31]
5484felt very empty for until I came to
5485
5486[89:35]
5487grips with the I yes I learned that this
5488
5489[89:39]
5490is a sort of a standard grief process
5491
5492[89:40]
5493that people go through ok so then now
5494
5495[89:43]
5496I'm at a point in my life even more so
5497
5498[89:47]
5499than in 2006 where where all of my go
5500
5501[89:50]
5502have been fulfilled except for finishing
5503
5504[89:52]
5505narrative computer programming
5506
5507[89:54]
5508i I I had one made unfulfilled goal that
5509
5510[90:01]
5511I'd wanted all my life to write a piece
5512
5513[90:04]
5514of a piece piece of music that and I had
5515
5516[90:08]
5517an idea for for a certain kind of music
5518
5519[90:12]
5520that I thought ought to be written at
5521
5522[90:13]
5523least somebody ought to try to do it and
5524
5525[90:15]
5526I and I felt that it was a that it
5527
5528[90:19]
5529wasn't going to be easy but I wanted to
5530
5531[90:20]
5532I wanted it proof of concept I wanted to
5533
5534[90:24]
5535know if it was going to work or not and
5536
5537[90:25]
5538so I spent a lot of time and finally I
5539
5540[90:29]
5541finished that piece and we had the we
5542
5543[90:31]
5544had the world premiere last year on my
5545
5546[90:34]
554780th birthday and we had another
5548
5549[90:36]
5550premiere in Canada and there's talk of
5551
5552[90:39]
5553concerts in Europe and various things so
5554
5555[90:41]
5556that but that's done it's part of the
5557
5558[90:44]
5559world's music now and it's either good
5560
5561[90:45]
5562or bad but I did what I was hoping to do
5563
5564[90:49]
5565so the only thing that I know that that
5566
5567[90:53]
5568I have on my agenda is to is to try to
5569
5570[90:58]
5571do as well as I can with the art of
5572
5573[91:00]
5574computer programming until I go see now
5575
5576[91:02]
5577do you think there's an element of point
5578
5579[91:05]
5580eight that might point eight yeah well I
5581
5582[91:08]
5583look at it more that I got actually took
5584
5585[91:13]
558621.0 with when that concert was over
5587
5588[91:18]
5589with I mean I you know I so in 2006 I
5590
5591[91:25]
5592was at point eight um so when I was
5593
5594[91:28]
5595diagnosed with prostate cancer then I
5596
5597[91:30]
5598said okay well maybe this is yet you
5599
5600[91:33]
5601know I've I've had all kinds of good
5602
5603[91:37]
5604luck all my life and there's no I'm
5605
5606[91:39]
5607nothing to complain about so I might die
5608
5609[91:41]
5610now and we'll see what happened and so
5611
5612[91:45]
5613so it's quite seriously I went and I
5614
5615[91:49]
5616didn't I had no expectation that I
5617
5618[91:54]
5619deserved better
5620
5621[91:56]
5622I didn't make any plans for the future I
5623
5624[91:59]
5625had my surgery I came out of the surgery
5626
5627[92:03]
5628and and spend some time learning how to
5629
5630[92:09]
5631walk again and so on is painful for a
5632
5633[92:13]
5634while but I got home and I realized I
5635
5636[92:17]
5637hadn't really thought about what what to
5638
5639[92:20]
5640do next I hadn't I hadn't any
5641
5642[92:22]
5643expectation and I'm still alive okay now
5644
5645[92:27]
5646I can write some more books but it but I
5647
5648[92:29]
5649didn't come with the attitude that you
5650
5651[92:30]
5652know I you know this was this was
5653
5654[92:34]
5655terribly unfair and and I just said okay
5656
5657[92:39]
5658I was accepting whatever it turned out
5659
5660[92:43]
5661you know I look like I gotten I got more
5662
5663[92:48]
5664than my shirt already so why should I
5665
5666[92:54]
5667and I didn't and I really when I got
5668
5669[92:58]
5670home I read I realized that I had really
5671
5672[93:00]
5673not thought about the next step what I
5674
5675[93:01]
5676would do after I would doubt after I
5677
5678[93:03]
5679would be able to work and I had sort of
5680
5681[93:05]
5682thought of it as if as this might you
5683
5684[93:08]
5685know I was comfortable with with the
5686
5687[93:11]
5688fact that it was at the end but but I
5689
5690[93:14]
5691was hoping that I would still you know
5692
5693[93:18]
5694be able to learn about satisfiability
5695
5696[93:23]
5697and and also someday even write music I
5698
5699[93:29]
5700didn't start I didn't started seriously
5701
5702[93:31]
5703on the music project until 2012 so I'm
5704
5705[93:35]
5706gonna be in huge trouble if I don't talk
5707
5708[93:37]
5709to you about this in in the 70s you've
5710
5711[93:41]
5712created the tech typesetting system
5713
5714[93:43]
5715together with meta font language for
5716
5717[93:45]
5718font description and computer modern
5719
5720[93:47]
5721family of typefaces that has basically
5722
5723[93:50]
5724defined the methodology in the aesthetic
5725
5726[93:53]
5727of the countless research fields right
5728
5729[93:57]
5730math physics well beyond design and so
5731
5732[94:02]
5733on okay well first of all thank you I
5734
5735[94:04]
5736think I speak for a lot of people in
5737
5738[94:07]
5739saying that but question in terms of
5740
5741[94:10]
5742beauty there's a beauty to typography
5743
5744[94:12]
5745that you've created and yet beauty is
5746
5747[94:16]
5748hard to
5749
5750[94:16]
5751five right how does one create beautiful
5752
5753[94:22]
5754letters and beautiful equations like
5755
5756[94:25]
5757what what
5758
5759[94:26]
5760so I mean perhaps there's no words to be
5761
5762[94:30]
5763describing you know be described in the
5764
5765[94:33]
5766process but so the great Harvard
5767
5768[94:38]
5769mathematician Georg deeper cut wrote a
5770
5771[94:43]
5772book in the 30s called the aesthetic
5773
5774[94:44]
5775measure rate where he would have
5776
5777[94:47]
5778pictures of vases and underneath would
5779
5780[94:49]
5781be a number and this was how beautiful
5782
5783[94:51]
5784the vase was and he had a formula for
5785
5786[94:53]
5787this and and he actually also right over
5788
5789[94:56]
5790brought about music and so he could he
5791
5792[94:59]
5793could you know so I thought maybe I
5794
5795[95:01]
5796would part of my musical composition I
5797
5798[95:04]
5799would try to program his algorithms and
5800
5801[95:08]
5802and you know so that I would I would
5803
5804[95:11]
5805write something that had the highest
5806
5807[95:13]
5808number by his score well it wasn't quite
5809
5810[95:15]
5811rigorous enough work for a computer to
5812
5813[95:18]
5814to do but anyway people have tried to
5815
5816[95:21]
5817put numerical value on beauty but and
5818
5819[95:24]
5820and he did probably the most serious
5821
5822[95:28]
5823attempt and and George Gershwin's
5824
5825[95:32]
5826teacher also wrote two volumes where he
5827
5828[95:34]
5829talked about his method of of composing
5830
5831[95:38]
5832music but but you're talking about
5833
5834[95:40]
5835another kind of beauty and beauty and
5836
5837[95:42]
5838letters and letter fell against and
5839
5840[95:44]
5841whatever that overture is right so so
5842
5843[95:47]
5844and so that's the beholder as they say
5845
5846[95:52]
5847but kinder striving for excellence in
5848
5849[95:55]
5850whatever definition you want to give to
5851
5852[95:57]
5853beauty then you try to get as close to
5854
5855[95:59]
5856that as you can somehow with it I guess
5857
5858[96:02]
5859I guess I'm trying to ask and there may
5860
5861[96:04]
5862not be a good answer
5863
5864[96:06]
5865what loose definitions were you're
5866
5867[96:09]
5868operating under with the community of
5869
5870[96:11]
5871people that you're working on oh the
5872
5873[96:13]
5874loose definition I wanted I wanted it to
5875
5876[96:18]
5877appeal to me to me I knew you personally
5878
5879[96:21]
5880yeah that's a good start
5881
5882[96:23]
5883yeah no and it failed that test went
5884
5885[96:25]
5886when I got volume two came out with this
5887
5888[96:28]
5889with the new printing and
5890
5891[96:30]
5892I was expecting to be the happiest day
5893
5894[96:31]
5895of my life and I felt like burning like
5896
5897[96:37]
5898how angry I was that I opened the book
5899
5900[96:40]
5901and it it was in the same beige covers
5902
5903[96:44]
5904and and but but it didn't look right on
5905
5906[96:47]
5907the page the number two was particularly
5908
5909[96:52]
5910ugly I couldn't stand any page that had
5911
5912[96:54]
5913a to in his page number and I was
5914
5915[96:57]
5916expecting that it was you know I spent
5917
5918[96:59]
5919all this time making measurements and I
5920
5921[97:01]
5922and I had Kent had looked at dolphins in
5923
5924[97:05]
5925different different ways and I hate I
5926
5927[97:08]
5928had great technology but but it did you
5929
5930[97:12]
5931know but I but I wasn't done I had I had
5932
5933[97:16]
5934to retune the whole thing after 196
5935
59361[97:19]
5937has it ever made you happy finally oh oh
5938
5939[97:22]
5940yes
5941
5942[97:24]
5943or is it appointing oh no no and so many
5944
5945[97:27]
5946books have come out that would never
5947
5948[97:29]
5949have been written without this I just
5950
5951[97:31]
5952didn't just draw it's just it's a joy
5953
5954[97:34]
5955but I could but now I I mean all these
5956
5957[97:37]
5958pages that are sitting up there I don't
5959
5960[97:40]
5961have a it if I didn't like him I would
5962
5963[97:43]
5964change him like that's my nobody else
5965
5966[97:47]
5967has this ability they have to stick with
5968
5969[97:49]
5970what I gave them
5971
5972[97:50]
5973yes so in terms of the other side of it
5974
5975[97:53]
5976there's the typography so the look of
5977
5978[97:55]
5979the top of the type and the curves and
5980
5981[97:57]
5982the lines what about the spacing but
5983
5984[98:01]
5985what about the spacing because you know
5986
5987[98:04]
5988the white space you know it seems like
5989
5990[98:07]
5991you could be a little bit more
5992
5993[98:08]
5994systematic about the layout or oh yeah
5995
5996[98:12]
5997you can always go further
5998
5999[98:13]
6000III I didn't I didn't stop at point
6001
6002[98:17]
6003eight I stopped I stopped about point
6004
6005[98:19]
6006nine eight seems like you're not
6007
6008[98:22]
6009following your own rule for happiness or
6010
6011[98:25]
6012is no no no I
6013
6014[98:29]
6015there's okay the course there's just
6016
6017[98:32]
6018what is the Japanese word wabi-sabi or
6019
6020[98:35]
6021something they wear the
6022
6023[98:38]
6024the most beautiful works of art are
6025
6026[98:40]
6027those that have flaws because then the
6028
6029[98:42]
6030person who who perceives them as their
6031
6032[98:46]
6033own
6034
6035[98:47]
6036appreciation and that gives the viewer
6037
6038[98:50]
6039more satisfaction or a so on but but I
6040
6041[98:54]
6042but no no with typography I wanted it to
6043
6044[98:57]
6045look as good as I could in in the vast
6046
6047[99:00]
6048majority of cases and then when it
6049
6050[99:03]
6051doesn't then I I say okay that's 2% more
6052
6053[99:08]
6054work for the wrote for the author but
6055
6056[99:11]
6057but I didn't want to I didn't want to
6058
6059[99:14]
6060say that my job was to get 200% with and
6061
6062[99:18]
6063take all the work away from the author
6064
6065[99:20]
6066that's what I meant by that so if you
6067
6068[99:23]
6069were to venture a guess how much of the
6070
6071[99:27]
6072nature of reality do you think we humans
6073
6074[99:29]
6075understand so you mentioned you
6076
6077[99:32]
6078appreciate mystery how much of the world
6079
6080[99:35]
6081about us is shrouded in mystery are we
6082
6083[99:38]
6084are we if you were to put a number on it
6085
6086[99:41]
6087what what percent of it all do we
6088
6089[99:44]
6090understand oh we totally how many
6091
6092[99:46]
6093leading zeroes any point zero point zero
6094
6095[99:48]
6096zero there I don't know now I think it's
6097
6098[99:50]
6099infinitesimal how do we think about that
6100
6101[99:53]
6102what do we do about that do we continue
6103
6104[99:55]
6105one step at a time yeah we muddle
6106
6107[99:58]
6108through I mean we do our best we
6109
6110[100:01]
6111realized that one that nobody's perfect
6112
6113[100:03]
6114then we and we try to keep advancing but
6115
6116[100:07]
6117we don't spend time saying we're not
6118
6119[100:12]
6120there we're not all the way to the end
6121
6122[100:14]
6123some some mathematicians that that would
6124
6125[100:18]
6126be in the office next to me when I was
6127
6128[100:19]
6129in the math department they would never
6130
6131[100:21]
6132think about anything smaller than
6133
6134[100:23]
6135countable infinity and I never you know
6136
6137[100:26]
6138we intersect that countable infinity
6139
6140[100:28]
6141because I really got up to countable
6142
6143[100:31]
6144infinity I was always talking about
6145
6146[100:32]
6147finite stuff but but even even limiting
6148
6149[100:36]
6150to finite stuff which was which is which
6151
6152[100:40]
6153the universe might be there's no way to
6154
6155[100:43]
6156really know what whether the universe is
6157
6158[100:47]
6159in
6160
6161[100:48]
6162isn't just made out of capital in
6163
6164[100:56]
6165whenever you want to call them quarks or
6166
6167[100:59]
6168whatever where capital n is some fun a
6169
6170[101:01]
6171number all of the numbers that are
6172
6173[101:03]
6174comprehensible are still way smaller
6175
6176[101:05]
6177than most almost all finite numbers III
6178
6179[101:08]
6180I got this one paper called supernatural
6181
6182[101:13]
6183numbers where I what I guess you've
6184
6185[101:16]
6186probably ran into something called Knuth
6187
6188[101:18]
6189arrow notation did you ever run into
6190
6191[101:19]
6192that where anyway so you take the number
6193
6194[101:22]
6195I think it's like I and I called it
6196
6197[101:25]
6198super K but I named it after myself but
6199
6200[101:29]
6201it's it's but in arrow notation is
6202
6203[101:31]
6204something like ten and then four arrows
6205
6206[101:34]
6207and a three or something might not okay
6208
6209[101:36]
6210no the arrow notation if you have if you
6211
6212[101:40]
6213have no arrows that means multiplication
6214
6215[101:41]
6216XY means x times X times X times X Y
6217
6218[101:46]
6219times if you have one arrow that means
6220
6221[101:49]
6222exponentiation so x one arrow Y means X
6223
6224[101:51]
6225to the X to the X to the X to the X Y
6226
6227[101:55]
6228times so I find out by the way that this
6229
6230[101:58]
6231is notation was invented by a guy in
6232
6233[102:02]
62341830 and and he was like he was a a a
6235
6236[102:08]
6237[Music]
6238
6239[102:10]
6240one of the English nobility who who
6241
6242[102:13]
6243spent his time thinking about stuff like
6244
6245[102:14]
6246this and it was exactly the same concept
6247
6248[102:18]
6249that I that I'm that I used arrows and
6250
6251[102:21]
6252he used a slightly different notation
6253
6254[102:23]
6255but anyway this and then this Ackerman's
6256
6257[102:25]
6258function is is based on the same kind of
6259
6260[102:28]
6261ideas but Ackerman was 1920s but anyway
6262
6263[102:31]
6264you got this number 10
6265
6266[102:34]
6267quadruple arrow 3 so that's that says
6268
6269[102:37]
6270well we take you know we take 10 to the
6271
6272[102:41]
627310 to the 10 to the 10 to the 10 to the
6274
6275[102:44]
627610th anyway how many times do we do that
6277
6278[102:46]
6279oh Ken double arrow two times or
6280
6281[102:49]
6282something I mean how tall is that stack
6283
6284[102:51]
6285but but but then we do that again
6286
6287[102:54]
6288because that was the only 10 triple
6289
6290[102:56]
6291quadruple arrow to we take quadruple
6292
6293[102:59]
6294three large number it
6295
6296[103:01]
6297gets way beyond comprehension okay yeah
6298
6299[103:06]
6300and and and so but it's so small
6301
6302[103:11]
6303compared to what finite numbers really
6304
6305[103:13]
6306are because I want to using four arrows
6307
6308[103:15]
6309and you know in ten and a three I mean
6310
6311[103:18]
6312let's have that let's have that many
6313
6314[103:21]
6315number arrows I mean the boundary
6316
6317[103:23]
6318between infinite and finite is
6319
6320[103:26]
6321incomprehensible for us humans anyway
6322
6323[103:29]
6324infinity is a good is a useful way for
6325
6326[103:32]
6327us to think about extremely large
6328
6329[103:36]
6330extremely large things and and and and
6331
6332[103:42]
6333we we can manipulate it but but we can
6334
6335[103:46]
6336never know that the universe is actually
6337
6338[103:49]
6339and we're near that so it just so I
6340
6341[103:57]
6342realize how little we know but but but
6343
6344[104:03]
6345what we we found an awful lot of things
6346
6347[104:09]
6348that are too hard for any one person to
6349
6350[104:11]
6351know even with even in our small
6352
6353[104:13]
6354universe yeah and we did pretty good so
6355
6356[104:17]
6357when you go up to heaven and meet God
6358
6359[104:19]
6360and get to ask one question that would
6361
6362[104:23]
6363get answered what question would you ask
6364
6365[104:29]
6366what kind of browser do you have up here
6367
6368[104:32]
6369[Laughter]
6370
6371[104:41]
6372[Music]
6373
6374[104:48]
6375okay and then oh that's beautiful
6376
6377[104:51]
6378actually Don thank you so much it was a
6379
6380[104:54]
6381huge honor to talk to you I really well
6382
6383[104:56]
6384thanks for the gamut of questions yeah
6385
6386[104:59]
6387it was fun
6388
6389[105:00]
6390thanks for listening to this
6391
6392[105:01]
6393conversation with donald knuth thank you
6394
6395[105:04]
6396to our presenting sponsor cash app
6397
6398[105:07]
6399downloaded use cold Luck's podcast
6400
6401[105:09]
6402you'll get ten dollars and ten dollars
6403
6404[105:11]
6405will go to first a stem education
6406
6407[105:13]
6408nonprofit that inspires hundreds of
6409
6410[105:15]
6411thousands of young minds to learn and to
6412
6413[105:18]
6414dream of engineering our future if you
6415
6416[105:21]
6417enjoy this podcast subscribe on YouTube
6418
6419[105:23]
6420give it five stars an apple podcast
6421
6422[105:25]
6423supported on patreon or connect with me
6424
6425[105:27]
6426on Twitter and now let me leave you with
6427
6428[105:30]
6429some words of wisdom from donald knuth
6430
6431[105:33]
6432we should continually be striving to
6433
6434[105:35]
6435transform every art into a science and
6436
6437[105:38]
6438in the process we advance the art thank
6439
6440[105:43]
6441you for listening and hope to see you
6442
6443[105:45]
6444next time