1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
|
.\" observations transcript, episode 002.
.\" Generated from the recorder's own transcript, then corrected by hand.
.\" Format: observations-transcripts(5). Editing it: observations-transcripts(7).
.episode 002
.recorded 2026-08-31
.source branwen "Gwern Branwen" "Laws of Tech: Commoditize Your Complement"
.url branwen https://gwern.net/complement
.source spolsky "Joel Spolsky" "Strategy Letter V: The Economics of Open Source"
.url spolsky https://www.joelonsoftware.com/2002/06/12/strategy-letter-v/
.p
00:00:00.000 00:00:01.860 Hi, and welcome to observations.
00:00:03.300 00:00:04.620 Today, I want to talk about the
00:00:04.620 00:00:06.480 oldest trick in the tech industry,
00:00:06.480 00:00:08.700 and especially in how to make money
00:00:08.700 00:00:10.740 in the tech industry called
00:00:10.740 00:00:12.720 commoditizing your complement.
00:00:13.800 00:00:16.020 And I'm gonna feature two articles,
00:00:16.020 00:00:18.720 um, one by guern, which is, which I'm
00:00:18.720 00:00:20.400 just gonna read the intro Vibe because?
00:00:21.360 00:00:23.040 I mean, it's a one article, so it's
00:00:23.040 00:00:26.520 like, um, 10K words, but the intro is
00:00:26.520 00:00:27.660 pretty good, and then I'm gonna jump
00:00:27.660 00:00:30.300 to the strategy letter number five by
00:00:30.300 00:00:32.880 Joel and software by Joel spolsky,
00:00:32.880 00:00:34.440 which he wrote in 2002.
.p
00:00:35.640 00:00:37.680 After being at Microsoft and
00:00:37.680 00:00:38.820 discovering this pattern?
00:00:41.400 00:00:43.560 Spelling out this pattern in a blog post.
00:00:45.120 00:00:47.880 Observations is a format I want to
00:00:47.880 00:00:51.000 create, which is basically me talking
00:00:51.000 00:00:53.160 to a microphone. I'm just pulling out
00:00:53.160 00:00:54.900 my recorders. You can probably tell
00:00:54.900 00:00:56.520 from the audio quality, like my phone
00:00:56.520 00:00:57.300 recorder right now.
.p
00:00:57.300 00:01:00.120 I don't even have a microphone, so if
00:01:00.120 00:01:02.520 you want, if you have observations
00:01:02.520 00:01:03.720 and want to share.
00:01:05.220 00:01:07.380 Just send me a like, open your
00:01:07.380 00:01:08.880 recorder as well, and send me a voice
00:01:08.880 00:01:10.260 message, and I might feature you in
00:01:10.260 00:01:11.880 the next episode. Once I figure out
00:01:11.880 00:01:14.880 how audio recording works and how to
00:01:14.880 00:01:16.800 splice and stuff, or it will just
00:01:16.800 00:01:17.700 cite you.
.p
00:01:18.600 00:01:20.280 But yeah, um, maybe you can also
00:01:20.280 00:01:22.860 reply on Mastodon or wherever?
00:01:23.760 00:01:27.000 Um, I might have some website in the
00:01:27.000 00:01:28.560 future right now. It's just a list of
00:01:28.560 00:01:29.100 episodes.
00:01:30.480 00:01:31.020 And the audio.
00:01:31.980 00:01:33.060 Okay, let's jump in.
00:01:34.080 00:01:36.960 So, ongoing.net complement very nice
00:01:36.960 00:01:39.960 URL laws of tech. Commoditize your
00:01:39.960 00:01:42.840 component. Tags AI, economics,
00:01:42.840 00:01:44.880 economics, Insight, p***, Google.
.p
.quote branwen
00:01:46.680 00:01:48.060 A classic pattern in technology
00:01:48.060 00:01:50.820 economics identified virtual sportsky
00:01:50.820 00:01:52.560 as layers of the stack attempting to
00:01:52.560 00:01:54.660 become monopolies while turning other
00:01:54.660 00:01:56.220 layers into perfectly competitive
00:01:56.220 00:01:58.140 markets, which are then, which are
00:01:58.140 00:02:00.180 commoditized in order to harvest more
00:02:00.180 00:02:02.640 of the consumer surplus discussion
00:02:02.640 00:02:03.420 and examples.
.endquote
00:02:04.620 00:02:06.780 So, especially the example section is
00:02:06.780 00:02:08.220 going to be interesting, but I think
00:02:08.220 00:02:09.540 it's gonna be a bit too much for this
00:02:09.540 00:02:10.200 episode.
.p
00:02:11.640 00:02:14.760 This was first published in 2018 and
00:02:14.760 00:02:19.200 then update in 2022, so I bought a
00:02:19.200 00:02:19.800 three year period.
00:02:20.940 00:02:23.100 In which Gwen added more examples and
00:02:23.100 00:02:24.000 is considered finished by.
00:02:25.980 00:02:27.720 Um, certainty. Highly likely
00:02:27.720 00:02:29.640 importance. Five. I think it's like a
00:02:29.640 00:02:31.260 out of 10 system that they use.
00:02:32.280 00:02:32.580 Um.
.p
00:02:35.880 00:02:39.960 Introduction Joel spolskin 2020 in 2002.
00:02:41.040 00:02:42.240 For 24 years ago
.quote branwen
00:02:42.240 00:02:42.900 identified a major
00:02:42.900 00:02:44.760 pattern in technology, business, and
00:02:44.760 00:02:46.380 economics. The pattern of
00:02:46.380 00:02:47.940 commoditizing your complement?
.endquote
00:02:49.020 00:02:51.900 So as an interjection, there's like a
00:02:51.900 00:02:53.520 this word commodity right
00:02:53.520 00:02:56.760 commoditizes a verb based on the on
00:02:56.760 00:02:57.360 the known.
00:02:58.440 00:02:59.460 Commodity and?
00:03:01.380 00:03:02.940 Observation I made is that the German
00:03:02.940 00:03:03.900 language doesn't really have?
.p
00:03:05.520 00:03:07.140 Um, an equal word.
00:03:09.720 00:03:11.160 Commonly known, and I don't think
00:03:11.160 00:03:14.520 there's there's probably commodity or
00:03:14.520 00:03:16.020 something used in economic circles,
00:03:16.020 00:03:18.360 but we have different terms that mean.
00:03:19.680 00:03:23.340 Just a piece of like foundational
00:03:23.340 00:03:26.220 material, but doesn't mean commodity in.
00:03:27.780 00:03:30.120 Um, so, so I would say a commodity is
00:03:30.120 00:03:30.420 something that.
00:03:32.220 00:03:34.680 There's multiple Alternatives that
00:03:34.680 00:03:35.880 are about the same price.
.p
00:03:37.200 00:03:38.760 And take. They are replaceable. So if
00:03:38.760 00:03:40.380 you go to the gas station and you,
00:03:40.380 00:03:42.840 you tank something. And it's gonna be
00:03:42.840 00:03:45.660 roughly the same fuel every
00:03:45.660 00:03:48.120 everywhere you go, but the producers
00:03:48.120 00:03:49.380 of the fuel might be a different
00:03:49.380 00:03:52.320 company or even the gas stations, a
00:03:52.320 00:03:53.820 different gas station. But the fuel
00:03:53.820 00:03:54.960 serves a commodity, right?
.p
00:03:56.340 00:03:57.300 Same goes for.
00:03:58.500 00:03:59.880 You could say the browser is a
00:03:59.880 00:04:01.860 commodity, right? Firefox, chromium,
00:04:01.860 00:04:04.200 Safari, they all Implement more or
00:04:04.200 00:04:06.180 less the same API. You can display
00:04:06.180 00:04:08.280 all the web pages more or less the
00:04:08.280 00:04:10.980 same, so the browser itself.
00:04:13.440 00:04:15.480 It's not really a commodity because
00:04:15.480 00:04:17.460 there's only three of them, but you
00:04:17.460 00:04:18.960 get the idea like, right?
.p
00:04:19.980 00:04:20.040 Um.
00:04:21.180 00:04:23.220 Like Windows and buildings like
00:04:23.220 00:04:24.840 Commodities. There's multiple producers.
00:04:25.980 00:04:28.140 Um, who fight for the cheapest price
00:04:28.140 00:04:31.680 or the best value, so the price will also?
00:04:33.720 00:04:37.500 Go down towards low margins in in a
00:04:37.500 00:04:37.920 healthy environment.
.p
00:04:42.780 00:04:44.820 So, there's no really equivalent word
00:04:44.820 00:04:45.600 for this in German.
00:04:46.680 00:04:48.900 There's good, there's vava, which is
00:04:48.900 00:04:52.500 just generic for trading trading item
00:04:52.500 00:04:52.860 basically.
00:04:54.420 00:04:56.400 Um, there's resource resource.
00:04:57.300 00:04:58.920 But there's nothing like, oh, this
00:04:58.920 00:04:59.220 thing is like?
00:05:01.020 00:05:02.940 Commodified doesn't exist, right?
00:05:03.960 00:05:05.700 Um, yeah, that was an interesting
00:05:05.700 00:05:06.900 observation I made, and I feel like
00:05:06.900 00:05:09.360 that also influences how, how Germans
00:05:09.360 00:05:09.600 think about?
.p
00:05:10.980 00:05:12.360 Um, especially, like?
00:05:13.740 00:05:14.820 Day-To-Day.
00:05:16.050 00:05:17.310 People who who don't have an
00:05:17.310 00:05:19.050 economics background to think about?
00:05:20.310 00:05:21.691 Um, the world.
00:05:22.710 00:05:23.850 Or rather, they they don't really
00:05:23.850 00:05:25.890 think about how to commoditize stuff.
00:05:26.850 00:05:28.230 Not very much.
00:05:29.130 00:05:30.810 But maybe that's just me being in a
00:05:30.810 00:05:31.950 bubble. Okay, let's continue.
00:05:34.260 00:05:36.000 So, the pattern of commoditizing a
00:05:36.000 00:05:36.300 component.
.p
00:05:36.300 00:05:37.980 It's an alternative to Vertical
00:05:37.980 00:05:39.240 integration, and
.quote branwen
00:05:39.240 00:05:40.440 it companies seek to
00:05:40.440 00:05:42.660 secure a choke point or quasi
00:05:42.660 00:05:44.760 Monopoly in products composed of many
00:05:44.760 00:05:46.860 necessary and sufficient layers by
00:05:46.860 00:05:49.080 dominating one layer before string.
00:05:49.080 00:05:50.760 So much competition in another layer
00:05:50.760 00:05:52.860 above or below its layer that no
00:05:52.860 00:05:54.360 competing monopolus can emerge.
00:05:54.360 00:05:56.700 Prices are driven down to marginal
00:05:56.700 00:05:58.140 costs elsewhere in the stack.
.p
00:05:58.140 00:05:59.640 Total price drops and increases
00:05:59.640 00:06:02.040 demand, and the majority of the
00:06:02.040 00:06:03.840 consumer surplus of the final product.
00:06:04.200 00:06:06.720 Can be diverted to the quasi-monopolist.
00:06:07.500 00:06:09.240 No matter how valuable the original
00:06:09.240 00:06:11.640 may be or how one, how much one could
00:06:11.640 00:06:13.320 charge for it, it can be more
00:06:13.320 00:06:15.120 valuable to make it free if it
00:06:15.120 00:06:16.680 increases profits elsewhere.
.p
00:06:17.880 00:06:19.440 A classic sample is the
00:06:19.440 00:06:21.300 commodification of PC Hardware by the
00:06:21.300 00:06:22.920 Microsoft OS Monopoly to
.endquote
00:06:22.920 00:06:23.940 judgment of
00:06:23.940 00:06:25.680 IBM and benefits of Microsoft.
00:06:28.200 00:06:28.320 Um.
.p
00:06:31.320 00:06:33.360 So, I noticed that the intro is a bit.
00:06:34.740 00:06:37.320 Is a bit unclear. As in, it says,
00:06:37.320 00:06:38.880 this is an alternative to vertical
00:06:38.880 00:06:39.360 integration.
00:06:40.530 00:06:43.050 Semicolon init company scene. So, I
00:06:43.050 00:06:44.910 think the this whole paragraph was
00:06:44.910 00:06:47.250 about commoditizing your complement.
00:06:49.110 00:06:49.290 Um.
00:06:51.330 00:06:52.590 Whereas vertical integration is.
00:06:54.390 00:06:56.190 You have control over every part of
00:06:56.190 00:06:57.630 the stack and thus.
.p
00:06:58.470 00:07:01.110 You can decrease margins in every
00:07:01.110 00:07:02.610 single layer, and then, like what
00:07:02.610 00:07:04.470 Apple does with with their projects,
00:07:04.470 00:07:07.350 is they create products that span
00:07:07.350 00:07:09.870 from they design their own Hardware,
00:07:09.870 00:07:11.370 they go via the operating system.
00:07:11.370 00:07:13.650 They create their own devices, and
00:07:13.650 00:07:15.870 then they sell the whole ecosystem,
00:07:15.870 00:07:17.250 including the apps and everything.
.p
00:07:18.960 00:07:19.620 But I think that's going to be an
00:07:19.620 00:07:20.040 example about.
00:07:21.600 00:07:24.540 Apple and the App Store, and like app
00:07:24.540 00:07:26.220 developers as an example of
00:07:26.220 00:07:28.080 commoditizing their complement later.
00:07:30.390 00:07:33.510 Going on. This pattern explains many,
.quote branwen
00:07:33.510 00:07:35.490 many, otherwise odd, or apparently
00:07:35.490 00:07:38.790 self-sabotaging Ventures by locked by
00:07:38.790 00:07:40.530 large tech companies into apparently
00:07:40.530 00:07:43.290 irrelevant Fields such, such as the
00:07:43.290 00:07:44.910 high rate of releasing open source
00:07:44.910 00:07:46.530 contributions by many internet
00:07:46.530 00:07:48.570 companies or the intrusion of
00:07:48.570 00:07:50.610 advertising companies into smartphone
00:07:50.610 00:07:52.170 manufacturing and web browser
00:07:52.170 00:07:54.090 development, statistical software,
00:07:54.090 00:07:56.610 fiber optic networks, Municipal
00:07:56.610 00:07:59.610 Wi-Fi, radio Spectrum auctions, DNS.
.p
00:08:00.750 00:08:02.910 Google. They are preemptive attempts
00:08:02.910 00:08:04.350 to commodify another company
00:08:04.350 00:08:06.750 elsewhere in the stack or defenses
00:08:06.750 00:08:07.650 against it being done to them.
.p
00:08:11.040 00:08:11.160 Um.
.p
00:08:14.970 00:08:17.370 And then he goes on. X Microsoft
00:08:17.370 00:08:20.310 product manager George Bosque, 2002
00:08:20.310 00:08:22.530 strategy letter. 5 the economics of
00:08:22.530 00:08:25.350 Open Source discusses a patternly saw
00:08:25.350 00:08:26.610 in technology companies, software,
00:08:26.610 00:08:27.570 and Microsoft in particular.
00:08:28.950 00:08:30.390 While he does not cite, it's likely
00:08:30.390 00:08:32.550 he was influenced by Carl Shapiro and
00:08:32.550 00:08:34.890 Google economists held variants
00:08:34.890 00:08:37.350 bestselling 1999 technology economics
00:08:37.350 00:08:39.870 book information rules, a strategic
00:08:39.870 00:08:41.670 guide to the network economy, which
00:08:41.670 00:08:43.590 discusses the idea extensively.
.p
00:08:45.510 00:08:47.970 A definition. Make yourself a
00:08:47.970 00:08:49.470 monopoly by growing the markets
00:08:49.470 00:08:49.710 around you.
00:08:50.910 00:08:51.210 I
.endquote
00:08:51.210 00:08:52.770 accept the definition and examples
00:08:52.770 00:08:55.470 below. Emphasis original and mostly
00:08:55.470 00:08:56.250 most links added.
00:08:58.530 00:09:02.130 Okay, so I think now is an excerpt
00:09:02.130 00:09:04.410 for Charles polskis.
00:09:04.950 00:09:06.450 Articles, so I would say, let's
00:09:06.450 00:09:07.050 switch over to.
00:09:09.150 00:09:09.210 Um.
00:09:10.170 00:09:12.210 To the strategy data and then start.
.p
00:09:13.350 00:09:15.390 Thinking about this a little bit. So,
00:09:15.390 00:09:18.450 June 12, 2002 by Joel sposky,
00:09:18.450 00:09:19.470 strategy letter number five.
00:09:20.670 00:09:22.770 Tax CEO Tech
00:09:22.770 00:09:22.890 news.
.quote spolsky
00:09:23.790 00:09:26.010 When I was in college, I took two
00:09:26.010 00:09:27.570 intro-economics courses,
00:09:27.570 00:09:30.330 macroeconomics, and microeconomics.
00:09:30.330 00:09:33.810 Macro rule of theories, like low end
00:09:33.810 00:09:35.670 unemployment, causes inflation that
00:09:35.670 00:09:37.950 never quite stood up to reality, but
00:09:37.950 00:09:40.290 the microstaff was both cool and useful.
.p
00:09:40.290 00:09:42.690 It was full of interesting Concepts
00:09:42.690 00:09:44.190 about the relationships between
00:09:44.190 00:09:46.050 supply and demand that really did
00:09:46.050 00:09:48.270 work. For example, if you have a
00:09:48.270 00:09:50.310 competitor who lowers their prices,
00:09:50.310 00:09:52.290 the demand for your product will go
00:09:52.290 00:09:53.250 down un.
00:09:53.730 00:09:54.210 To match them.
00:09:54.930 00:09:56.850 In today's episodes, I'll show one of
00:09:56.850 00:09:59.310 those Concepts explains a lot about
00:09:59.310 00:10:01.470 some familiar computer companies.
.p
00:10:02.610 00:10:05.130 I show how along the way I noticed
00:10:05.130 00:10:06.450 something interesting about open
00:10:06.450 00:10:08.250 source software, which is this. Most
00:10:08.250 00:10:10.170 of the companies spending big money
00:10:10.170 00:10:12.570 to develop open source software are
00:10:12.570 00:10:14.250 doing it because it's a good business
00:10:14.250 00:10:16.530 strategy for them, not because they
00:10:16.530 00:10:17.970 suddenly suddenly stop believing in
00:10:17.970 00:10:19.470 capitalism and fell in love with
00:10:19.470 00:10:20.730 freedom as in speech.
.p
00:10:22.230 00:10:24.270 Every product in the marketplace has
00:10:24.270 00:10:25.710 substitutes and complements.
00:10:26.790 00:10:28.410 A substitute is another product you
00:10:28.410 00:10:30.390 might buy if the first product is too
00:10:30.390 00:10:32.730 expensive. Chicken is a substitute
00:10:32.730 00:10:34.890 for beef if you're a chicken farmer
00:10:34.890 00:10:36.630 and the price of beef goes up. The
00:10:36.630 00:10:38.130 people will want more chicken, and
00:10:38.130 00:10:39.150 you will sell more.
.p
00:10:40.770 00:10:41.910 A compliment is a product that
00:10:41.910 00:10:43.590 usually buy together with another
00:10:43.590 00:10:45.810 product. Gas and cars are
00:10:45.810 00:10:48.090 complements. Computer hardware is a
00:10:48.090 00:10:49.470 classic complement of computer
00:10:49.470 00:10:50.730 operating operating systems.
00:10:52.050 00:10:53.970 And babysitters are complement of the
00:10:53.970 00:10:54.690 night fine restaurants.
00:10:55.770 00:10:57.030 In a small town when the local
00:10:57.030 00:10:59.130 five-star restaurant has a two. For
00:10:59.130 00:11:00.450 one, Valentine stay special.
.p
00:11:00.450 00:11:03.150 The local babysitter double their rates.
00:11:03.930 00:11:05.310 Actually, the nine-year-old get
00:11:05.310 00:11:06.450 robbed into early service.
00:11:08.940 00:11:10.920 All else being equal. Demand for a
00:11:10.920 00:11:11.760 product increases when the
00:11:11.760 00:11:12.060 price of
00:11:12.060 00:11:13.560 its complements decrease.
.endquote
.p
00:11:17.430 00:11:17.550 So?
00:11:18.690 00:11:19.890 I guess very?
00:11:20.790 00:11:22.470 Definition of this feature.
.p
00:11:26.700 00:11:28.020 A compliment being something that is
00:11:28.020 00:11:30.720 required for your Pro you to sell
00:11:30.720 00:11:31.080 your product, right?
.p
00:11:36.420 00:11:37.020 So, if you think about?
00:11:38.340 00:11:40.620 If you're a programmer and?
.p
00:11:45.240 00:11:46.440 A new program you want to program
00:11:46.440 00:11:46.680 something?
00:11:48.180 00:11:49.860 Maybe your compliments are the
00:11:49.860 00:11:51.600 libraries you use or the operating
00:11:51.600 00:11:52.620 systems you are running on?
00:11:53.880 00:11:54.900 Or, if you are?
00:11:57.270 00:11:58.710 Building a website like Google does
00:11:58.710 00:12:01.290 our search engine, then the
00:12:01.290 00:12:02.550 complement is probably the thing that
00:12:02.550 00:12:04.230 is going to display your website and.
.p
00:12:05.670 00:12:08.130 Um, the internet providers that are
00:12:08.130 00:12:09.690 gonna transfer your website to the
00:12:09.690 00:12:09.990 customer.
.p
00:12:13.830 00:12:17.730 And if you're in the ad business and you?
00:12:19.110 00:12:20.130 You notice that there's going to be a
00:12:20.130 00:12:22.170 lot more videos uploaded to the
00:12:22.170 00:12:25.230 internet, and you really, really feel
00:12:25.230 00:12:27.090 like there's going to be a big slice
00:12:27.090 00:12:29.550 of the internet. It's just being videos.
00:12:30.750 00:12:32.370 People consuming video material.
00:12:32.370 00:12:33.570 Instead of reading blog posts like,
00:12:33.570 00:12:34.170 we are right now.
.p
00:12:35.130 00:12:36.510 Then, you probably want to buy a
00:12:36.510 00:12:36.811 company that's gonna.
00:12:38.790 00:12:39.870 Provide the infrastructure for
00:12:39.870 00:12:41.910 serving video on the internet,
00:12:41.910 00:12:43.350 because that's your compliment,
00:12:43.350 00:12:45.450 right. You want to serve advertising
00:12:45.450 00:12:47.370 on videos, so you need to be inside
00:12:47.370 00:12:50.550 of the video delivery process.
.p
00:12:57.930 00:12:59.910 Or maybe, if you, if you're
00:12:59.910 00:13:01.170 advertising business, you go as far
00:13:01.170 00:13:03.510 as saying, oh. But Broadband internet
00:13:03.510 00:13:04.830 is important for people to use.
00:13:06.210 00:13:09.270 And if other providers are starting
00:13:09.270 00:13:12.330 to, like, then you complement is
00:13:12.330 00:13:13.230 gonna be either.
00:13:15.390 00:13:15.510 Um.
00:13:16.410 00:13:19.290 Laws that make delivering content like?
00:13:21.240 00:13:23.880 Being specific and who you allowed
00:13:23.880 00:13:25.620 conductibility to be levered by,
00:13:25.620 00:13:27.300 like, net materiality.
.p
00:13:28.080 00:13:28.980 Very important for you because?
00:13:30.630 00:13:30.750 Um.
00:13:31.590 00:13:33.750 That's your your lifeline. Like, if
00:13:33.750 00:13:36.150 you are Google or Facebook and net
00:13:36.150 00:13:37.350 neutrology doesn't exist, and
00:13:37.350 00:13:40.230 suddenly, maybe some other competitor
00:13:40.230 00:13:41.670 can come into.
00:13:42.870 00:13:44.670 Um, can, like, buy up.
00:13:45.570 00:13:46.710 The domain provider.
00:13:48.240 00:13:48.360 Um.
00:13:49.920 00:13:52.260 And say, huh? If you serve my, if you
00:13:52.260 00:13:54.720 prefer my packets over Google's
00:13:54.720 00:13:56.580 packets, then I'm gonna pay you directly.
.p
00:13:58.020 00:13:59.700 And suddenly the market for
00:13:59.700 00:14:00.420 advertising is.
00:14:01.440 00:14:04.140 Is under attack.
00:14:05.100 00:14:05.880 So, like, basically.
00:14:08.160 00:14:08.760 And there's like a.
.p
00:14:12.180 00:14:13.680 Like, somebody can drive a wedge
00:14:13.680 00:14:16.020 between what you're selling and who
00:14:16.020 00:14:17.400 you're selling to and how you're
00:14:17.400 00:14:18.000 selling it, right?
00:14:19.140 00:14:21.120 That's like the the downstream side
00:14:21.120 00:14:22.260 of things. And then there's the
00:14:22.260 00:14:24.360 Upstream side of things. Like, what
00:14:24.360 00:14:26.760 are you? What do you require to even
00:14:26.760 00:14:28.320 build your product in the first place?
.p
00:14:35.010 00:14:35.850 So, there's gonna be a lot of
00:14:35.850 00:14:37.170 examples later.
00:14:39.690 00:14:40.830 But I think it's very important,
00:14:40.830 00:14:42.930 especially for software development
00:14:42.930 00:14:44.670 developers to think about economic
00:14:44.670 00:14:48.810 reasons for things like browsers
00:14:48.810 00:14:50.190 being open sourced or.
00:14:51.630 00:14:55.290 Um, steam valve going into Linux as.
00:14:56.550 00:14:58.170 Their operating system, as opposed to
00:14:58.170 00:14:59.310 Microsoft Windows.
00:15:00.330 00:15:00.690 Um.
00:15:01.950 00:15:04.890 Or also, like why social media sites?
.p
00:15:06.000 00:15:08.580 Two things they do, why they do
00:15:08.580 00:15:09.540 things the way they do them.
00:15:10.800 00:15:11.040 Um.
00:15:12.360 00:15:14.100 And why Microsoft will suddenly
00:15:14.100 00:15:16.800 publish lots of Open Source, but
00:15:16.800 00:15:18.600 like, even by companies that they
00:15:18.600 00:15:20.040 just open source afterwards, right.
00:15:20.040 00:15:22.200 Like, GitHub is probably something
00:15:22.200 00:15:24.000 related to that which I haven't
00:15:24.000 00:15:24.420 really thought.
00:15:25.920 00:15:26.160 About.
00:15:27.300 00:15:27.660 Um.
.p
00:15:29.670 00:15:30.390 Maybe it's a different thing actually?
.p
00:15:33.780 00:15:34.800 Okay, let's go on.
00:15:36.600 00:15:37.980 All else being equal. Demand for
00:15:37.980 00:15:39.420 product increases when the price
.quote spolsky
00:15:39.420 00:15:39.660 of
00:15:39.660 00:15:40.860 its complements decrease.
00:15:41.940 00:15:43.140 Let me repeat that because you might
00:15:43.140 00:15:44.640 have dosed off and it's important
00:15:44.640 00:15:47.580 demand for a product increases when
00:15:47.580 00:15:49.440 the price of its complements
00:15:49.440 00:15:53.100 decrease. For example, if flights to
00:15:53.100 00:15:55.200 Miami become cheaper demand for hotel
00:15:55.200 00:15:57.600 rooms in Miami goes up because more
00:15:57.600 00:16:00.060 people are flying to Miami and need a
00:16:00.060 00:16:00.120 room.
.p
00:16:01.080 00:16:02.940 When computers become cheaper, more
00:16:02.940 00:16:05.100 people buy them, and they all need
00:16:05.100 00:16:06.720 operating systems. So, demand for
00:16:06.720 00:16:08.580 operating system goes up, which means
00:16:08.580 00:16:10.260 the price of operating
.endquote
00:16:10.260 00:16:11.040 system can go up.
00:16:13.560 00:16:14.820 So, but what if somebody comes along
00:16:14.820 00:16:17.160 and says, ha? Demand for mobile
00:16:17.160 00:16:18.720 phones is gonna increase, right?
00:16:18.720 00:16:21.120 Because everybody's gonna want to
00:16:21.120 00:16:22.260 have a mobile phone like an iPhone?
.p
00:16:23.520 00:16:25.260 Um, but Apple has got all the best
00:16:25.260 00:16:26.760 mobile phones. So, everybody's buying
00:16:26.760 00:16:28.020 into Apple's close ecosystem?
00:16:29.340 00:16:31.740 And also, if we if suddenly somebody
00:16:31.740 00:16:33.180 starts building mobile phones and
00:16:33.180 00:16:34.560 somebody else comes up with an
00:16:34.560 00:16:34.980 operating system.
00:16:36.600 00:16:37.920 Then, we won't be able to sell our
00:16:37.920 00:16:40.080 advertising on the apps in the
00:16:40.080 00:16:40.620 operating system.
.p
00:16:42.000 00:16:43.380 So, what if we instead?
00:16:44.520 00:16:45.960 Build the middle layer if we start
00:16:45.960 00:16:47.040 building Android and.
00:16:48.660 00:16:50.160 Open source, so there's like no
00:16:50.160 00:16:51.780 market for anybody else to go into
00:16:51.780 00:16:54.060 the into the business of selling
00:16:54.060 00:16:56.160 operating systems for mobile phones,
00:16:56.160 00:16:58.380 and also we can bind all of the
00:16:58.380 00:17:00.840 hardware producers to use Android
00:17:00.840 00:17:01.980 because it's the only.
.p
00:17:03.240 00:17:05.340 Readily available platform that can
00:17:05.340 00:17:08.280 be used, and that also has all of the
00:17:08.280 00:17:09.720 search engine stuff already built
00:17:09.720 00:17:10.440 into it so.
00:17:11.460 00:17:13.500 So you can see how the thinking at
00:17:13.500 00:17:15.360 Google goes if when they, when they
00:17:15.360 00:17:17.160 notice that, oh s***, the mobile
00:17:17.160 00:17:20.160 phones are coming, and how do you,
00:17:20.160 00:17:21.360 how do we react to that, right?
.p
00:17:27.180 00:17:27.360 But also.
00:17:29.700 00:17:31.200 There's this law that everybody is
00:17:31.200 00:17:32.640 talking about right now, which?
00:17:33.900 00:17:35.940 Of course, I forgot already. What?
00:17:35.940 00:17:36.780 What the name was?
00:17:37.920 00:17:39.960 Uh, like it starts with G.
00:17:40.620 00:17:40.980 Um.
.p
00:17:44.310 00:17:46.290 I will, I will. I will remember not
00:17:46.290 00:17:46.710 in a bit.
00:17:47.430 00:17:48.870 Which is basically that there's these
00:17:48.870 00:17:49.110 kind of.
00:17:50.190 00:17:51.031 Unlocks of?
00:17:53.220 00:17:53.340 Um.
00:17:54.240 00:17:56.700 Unlocks of of a price point where
00:17:56.700 00:17:58.440 something becomes cheap enough that
00:17:58.440 00:18:01.080 new cat Loop new kinds of categories
00:18:01.080 00:18:02.040 of usage.
00:18:03.000 00:18:04.260 Um, suddenly become.
00:18:05.220 00:18:06.540 Available rights.
00:18:07.740 00:18:09.660 So, for example, the the internet
00:18:09.660 00:18:11.520 isn't like the the public internet
00:18:11.520 00:18:12.540 wasn't possible before.
.p
00:18:12.540 00:18:14.340 There was a widely distributed.
00:18:15.780 00:18:15.840 Um.
00:18:16.920 00:18:18.660 Phone Network, obviously, because you
00:18:18.660 00:18:20.940 needed modems, and then you needed a
00:18:20.940 00:18:22.920 widely disputed personal computer in
00:18:22.920 00:18:23.640 people's home.
00:18:24.720 00:18:26.040 And then suddenly the internet
00:18:26.040 00:18:27.120 becomes possible outside of
00:18:27.120 00:18:28.500 University servers.
00:18:29.280 00:18:31.320 And that also makes online
00:18:31.320 00:18:32.460 advertising possible and.
00:18:33.660 00:18:35.700 And you know, there's these. These
00:18:35.700 00:18:37.320 unlock unlocked points where
00:18:37.320 00:18:38.760 something becomes cheap enough that
00:18:38.760 00:18:38.940 people.
.p
00:18:40.020 00:18:42.180 Move into it almost and then.
00:18:43.860 00:18:44.341 They can discover new.
00:18:45.720 00:18:46.380 Possibilities.
00:18:48.420 00:18:50.160 Or, for example, Netflix is only possible.
00:18:52.080 00:18:52.140 Um.
00:18:53.340 00:18:54.660 Because of high bandwidth.
00:18:55.440 00:18:58.320 In most developed countries or even
00:18:58.320 00:18:59.460 like 4.
00:19:01.200 00:19:03.300 And Broadband's mobile internet.
.p
00:19:06.810 00:19:07.950 Which is also only possible because?
00:19:09.990 00:19:10.950 You get these.
00:19:12.990 00:19:14.610 Flat per month payments for
00:19:14.610 00:19:18.210 unlimited, mostly unlimited bandwidth
00:19:18.210 00:19:19.470 and not metered by the.
00:19:20.490 00:19:21.570 Microsecond as before.
.p
00:19:31.620 00:19:31.800 So basically.
00:19:33.180 00:19:34.620 If something becomes cheaper, the
00:19:34.620 00:19:36.180 demand goes up. But also, what I'm
00:19:36.180 00:19:38.760 saying is that at certain points, the
00:19:38.760 00:19:41.280 demand goes up more than linearly,
00:19:41.280 00:19:43.560 like. But there's unlock points that
00:19:43.560 00:19:44.160 make things.
00:19:45.480 00:19:46.320 Basically.
00:19:47.460 00:19:47.760 Explode.
00:19:48.780 00:19:48.840 Because.
00:19:50.760 00:19:51.000 For example.
.p
00:19:54.810 00:19:54.930 Um.
00:19:56.190 00:19:57.570 I don't know home fitness trainer.
00:19:57.570 00:19:59.190 Nobody would buy a home fitness
00:19:59.190 00:20:02.130 trainer for 2 and 1/2k, but once a
00:20:02.130 00:20:04.290 home fitness trainer costs under 300,
00:20:04.290 00:20:06.990 suddenly people will start buying them.
00:20:08.130 00:20:10.230 Maybe I don't know, probably not
00:20:10.230 00:20:11.370 because it's kind of stupid.
00:20:13.410 00:20:13.411 Um.
.p
00:20:16.860 00:20:16.980 Uh,
.quote spolsky
00:20:16.980 00:20:17.880 at this point, it's pretty common
00:20:17.880 00:20:19.740 for people to try to confuse things
00:20:19.740 00:20:22.440 by saying, aha. But Linux is free.
00:20:22.440 00:20:25.080 Okay, first of all. When an economist
00:20:25.080 00:20:26.940 considers price, they consider the
00:20:26.940 00:20:28.380 total price, including some
00:20:28.380 00:20:30.360 intangible things, like the time it
00:20:30.360 00:20:32.520 takes to set up, re-educate in
00:20:32.520 00:20:34.860 everyone and converting system processes.
.p
00:20:34.860 00:20:36.480 All the things that we like to call
00:20:36.480 00:20:38.580 total cost of ownership.
.endquote
00:20:39.360 00:20:41.640 Also, often it's a mode around
00:20:41.640 00:20:43.320 existing products like Microsoft
00:20:43.320 00:20:44.880 object office has.
00:20:46.260 00:20:48.120 These modes, like they do something
00:20:48.120 00:20:49.860 that is not like that, is very
00:20:49.860 00:20:50.760 specific to their.
00:20:52.260 00:20:53.940 Um, their own products and.
00:20:55.440 00:20:57.120 Before anybody can migrate to
00:20:57.120 00:20:57.480 something else.
.p
00:20:58.620 00:21:00.060 The other solution has to implement
00:21:00.060 00:21:02.640 basically all of their feature set so
00:21:02.640 00:21:04.320 that all the possible files can be.
00:21:05.760 00:21:05.820 Um.
00:21:06.960 00:21:08.940 Executed by this other software,
00:21:08.940 00:21:10.680 right. Which is why it's so hard for
00:21:10.680 00:21:12.900 people to migrate away from Microsoft
00:21:12.900 00:21:15.540 Excel or to LibreOffice or whatever.
00:21:16.680 00:21:18.361 Um, but then, of course, you get these.
.p
00:21:20.250 00:21:21.810 Good enough solutions, which is like
00:21:21.810 00:21:25.050 for 89%, 98 percent of all the use
00:21:25.050 00:21:26.850 cases. A Google sheet is enough for
00:21:26.850 00:21:28.170 simple.
00:21:29.430 00:21:31.590 Spreadsheets and the extra
00:21:31.590 00:21:33.210 functionality of being able to share,
00:21:33.210 00:21:36.510 like, live. Edit multiple people in
00:21:36.510 00:21:36.870 the same.
00:21:38.190 00:21:39.931 Document means there's.
00:21:40.950 00:21:42.990 Enough of an advantage to using that
00:21:42.990 00:21:44.850 that people will switch away from
00:21:44.850 00:21:46.110 Microsoft Excel.
.p
00:21:46.890 00:21:46.950 And.
00:21:48.150 00:21:49.770 Towards the Google sheet stuff.
.p
00:21:53.610 00:21:54.930 Which also has this interesting
00:21:54.930 00:21:57.750 corollary that just re-implementing
00:21:57.750 00:21:58.710 something s*****?
00:21:59.730 00:22:02.370 As a free, open source software is often.
00:22:03.360 00:22:04.560 A losing battle because?
00:22:07.020 00:22:08.400 Um, the existing companies have the
00:22:08.400 00:22:10.200 money the existing companies have the.
00:22:11.400 00:22:13.020 The advantage of being probably
00:22:13.020 00:22:14.580 earlier in the game, of course. The
00:22:14.580 00:22:16.320 second mover is always a bit easier
00:22:16.320 00:22:18.000 because they can just re-implement
00:22:18.000 00:22:18.120 everything.
.p
00:22:19.380 00:22:19.560 But.
00:22:20.880 00:22:23.400 Um, which is not true when I'm
00:22:23.400 00:22:24.660 talking about LibreOffice, because
00:22:24.660 00:22:26.220 that was OpenOffice, which was before
00:22:26.220 00:22:26.580 that.
00:22:28.020 00:22:31.020 Star office, I think, which is a
00:22:31.020 00:22:32.700 stone all piece of software from the
00:22:32.700 00:22:34.140 early 90s.
00:22:35.340 00:22:36.600 Um, but I mean, you get the idea,
00:22:36.600 00:22:38.460 right? Like, this person can
00:22:38.460 00:22:39.660 Implement something pretty quickly.
.p
00:22:41.280 00:22:42.600 And just copy the features, but then
00:22:42.600 00:22:45.600 it will be very hard to get people to
00:22:45.600 00:22:48.780 move over from the paid for product
00:22:48.780 00:22:49.800 once they're already.
00:22:50.760 00:22:52.740 Are sold on edge.
00:22:54.990 00:22:56.190 Versus if you have some other
00:22:56.190 00:22:58.230 product, which is a bit less than the
00:22:58.230 00:23:00.690 original, but has some super cool
00:23:00.690 00:23:01.170 extra feature.
.p
00:23:01.170 00:23:02.490 Maybe that will be enough to make
00:23:02.490 00:23:04.530 people move over, especially the
00:23:04.530 00:23:06.150 people that are not experts. Like,
00:23:06.150 00:23:08.610 you won't be able to convince
00:23:08.610 00:23:10.890 somebody to move from whim, like
00:23:10.890 00:23:13.230 their favorite editor or IDE to some
00:23:13.230 00:23:15.090 other IDE, unless you give them a
00:23:15.090 00:23:16.950 really good reason why they would move.
.p
00:23:18.150 00:23:20.130 Because they're so settled in their
00:23:20.130 00:23:22.050 own ways, like the women people will
00:23:22.050 00:23:23.370 always have their own motion key
00:23:23.370 00:23:25.770 bindings and and their whim config,
00:23:25.770 00:23:27.450 which they probably already had spend
00:23:27.450 00:23:28.410 a thousand hours configuring.
00:23:29.670 00:23:30.930 And the EMX people are the same.
00:23:31.950 00:23:31.951 So?
00:23:33.450 00:23:34.770 If you want to sell a new product and
00:23:34.770 00:23:36.390 you really like vs code, you really
00:23:36.390 00:23:37.650 have to make sure that it has.
.p
00:23:38.850 00:23:40.890 Unique advantages, like ease of setup
00:23:40.890 00:23:43.050 or total provider integration or
00:23:43.050 00:23:44.250 plug-in ecosystem.
00:23:46.110 00:23:47.790 Trump's these existing?
00:23:48.990 00:23:50.730 Um, solutions.
.p
00:23:54.630 00:23:55.770 Okay, I'm totally.
.p
00:23:59.190 00:24:00.810 Off track here. The article still
00:24:00.810 00:24:02.010 goes on on a little bit.
00:24:03.750 00:24:05.910 So, we were about. Oh, Linux is free,
00:24:05.910 00:24:07.110 but of course, it's not free because
00:24:07.110 00:24:09.750 people work on it. There's intangible
00:24:09.750 00:24:11.370 stuff, like, how long it takes for
00:24:11.370 00:24:14.910 people to migrate from the win32 IPI
00:24:14.910 00:24:16.590 to some unique space API?
.p
00:24:16.590 00:24:18.810 Like, like all of this friction
00:24:18.810 00:24:21.030 points, so, so the total price of
00:24:21.030 00:24:23.430 using Linux, instead of some other
00:24:23.430 00:24:25.590 operating system, is a lot more than?
00:24:25.590 00:24:27.030 Oh, well, it's free.
00:24:28.170 00:24:30.690 Secondly, by using the free S and B
.quote spolsky
00:24:30.690 00:24:32.130 argument, these Advocates try to
00:24:32.130 00:24:33.450 believe that they are not subject to
00:24:33.450 00:24:34.830 the rules of Economics because
00:24:34.830 00:24:36.810 they've got a nice, nice zero.
.p
00:24:36.810 00:24:37.950 They can multiply everything by.
00:24:37.950 00:24:41.010 Here's an example when slash dot asks
00:24:41.010 00:24:43.350 Linux developer mosha bar, a few
00:24:43.350 00:24:44.490 children's comments would be
00:24:44.490 00:24:45.870 compatible with existing device
00:24:45.870 00:24:47.910 drivers, he said. They didn't need to
00:24:47.910 00:24:50.430 proprietary software goes at the
00:24:50.430 00:24:53.850 Tariff of U.S dollars 50 to 100 per
00:24:53.850 00:24:54.750 line of debug code.
00:24:55.710 00:24:57.030 Okay, that's the wrong way to read
00:24:57.030 00:25:00.270 that 5200 US dollars line of
00:25:00.270 00:25:00.990 Riverbank code.
.p
00:25:00.990 00:25:03.510 No, such price applies to open source
00:25:03.510 00:25:05.910 software. Moshi goes on to claim
00:25:05.910 00:25:07.470 then. It's okay for every Linux
00:25:07.470 00:25:08.970 kernel revision to make all existing
00:25:08.970 00:25:11.730 drivers obsolete because the cost of
00:25:11.730 00:25:13.410 rewriting all those existing drivers
00:25:13.410 00:25:14.070 is zero.
00:25:15.090 00:25:17.070 This is completely wrong. He's
00:25:17.070 00:25:18.330 basically claiming that spending a
00:25:18.330 00:25:19.590 small amount of programming time,
00:25:19.590 00:25:21.150 making the kernel backwards
00:25:21.150 00:25:23.370 compatible, is equivalent to spending
00:25:23.370 00:25:24.990 a huge amount of programming time,
00:25:24.990 00:25:27.390 rewriting every driver because both
00:25:27.390 00:25:29.610 numbers are multiplied by their cost,
00:25:29.610 00:25:31.770 which he believes to be zero.
.p
00:25:32.910 00:25:35.370 This is a primer for C fallacy.
00:25:36.090 00:25:37.110 The thousands of millions of
00:25:37.110 00:25:39.150 developer hours it takes to revise
00:25:39.150 00:25:41.310 every existing device driver we're
00:25:41.310 00:25:42.810 going to have to come at the expense
00:25:42.810 00:25:45.210 of something, and until that's done,
00:25:45.210 00:25:46.950 Linux will be once again handicapped
00:25:46.950 00:25:48.570 in the marketplace because it doesn't
00:25:48.570 00:25:50.070 support existing Hardware.
.p
00:25:51.090 00:25:52.230 Wouldn't it be better to use all
00:25:52.230 00:25:55.230 those zero cost effort, making gnome
00:25:55.230 00:25:55.410 better
.endquote
00:25:55.410 00:25:57.330 for supporting new hardware?
00:25:59.190 00:25:59.970 This is funny.
00:26:00.990 00:26:03.090 It's very funny because wouldn't it
00:26:03.090 00:26:05.550 be better to use this to make gnome
00:26:05.550 00:26:05.730 better?
00:26:08.070 00:26:10.650 Like, say that in 2002 is, like, oh
00:26:10.650 00:26:13.230 oh, you're such a like.
00:26:15.330 00:26:17.910 Sweet, my sweet summer child.
.p
00:26:17.910 00:26:18.810 Ignore me is exactly the thing that
00:26:18.810 00:26:20.910 you don't want to put as an example
00:26:20.910 00:26:22.950 here because they tend to rewrite all
00:26:22.950 00:26:24.390 of their f****** libraries every.
00:26:24.390 00:26:26.610 Okay, sorry, I'm swearing it like
00:26:26.610 00:26:27.930 there. Just rewrite everything every
00:26:27.930 00:26:29.550 5 years, and everything else breaks
00:26:29.550 00:26:31.290 like this. I think I currently have,
00:26:31.290 00:26:34.530 like gtk 2, 3, and 4 on this system.
.p
00:26:36.060 00:26:37.500 And they all behave differently, and
00:26:37.500 00:26:40.620 like dark mode works on gtk4, but not
00:26:40.620 00:26:42.780 on 5, which means that some
00:26:42.780 00:26:44.700 applications switch to dark mode when
00:26:44.700 00:26:46.500 I switch, but then some don't because
00:26:46.500 00:26:49.080 somebody decided they it's better to
00:26:49.080 00:26:50.760 break backwards. Compatibility to get
00:26:50.760 00:26:52.140 some new fun features.
00:26:53.700 00:26:56.700 Tend to support old software, I guess.
.p
00:26:57.780 00:26:59.580 So, yeah, gnome is the wrong example
00:26:59.580 00:27:01.200 here because they exactly have this.
00:27:02.460 00:27:05.580 They, they don't behave like a homo
00:27:05.580 00:27:07.140 economicals they behave like.
00:27:08.250 00:27:10.170 Some random guy has some ideas and
00:27:10.170 00:27:12.090 more time to implement it than other
00:27:12.090 00:27:13.470 people, and then just doesn't give a
00:27:13.470 00:27:14.310 s*** about everybody else.
.p
00:27:20.940 00:27:21.900 Yeah, of course. This is personal.
00:27:22.980 00:27:23.760 Personal gripes.
.p
00:27:27.420 00:27:29.040 So, yeah, this is basically don't
00:27:29.040 00:27:30.240 break compatibility. If you don't
00:27:30.240 00:27:32.340 have to add any cost?
00:27:34.260 00:27:35.220 Oh well, I guess.
00:27:36.930 00:27:39.210 The actual answer would be don't
00:27:39.210 00:27:41.070 break compatibility until you've done
00:27:41.070 00:27:43.710 the math and figured out how much
00:27:43.710 00:27:46.650 value you're destroying by destroying
00:27:46.650 00:27:48.570 compatibility, which I guess nobody
00:27:48.570 00:27:50.250 does in the open source ecosystem
00:27:50.250 00:27:52.830 because they all think economic
00:27:52.830 00:27:54.810 economics is for stupid people and.
.p
00:27:56.250 00:27:57.150 Anybody with a business degree
00:27:57.150 00:27:58.350 doesn't actually know anything?
00:28:00.390 00:28:02.070 Which is very funny. It's very funny,
00:28:02.070 00:28:03.630 like in a sad way.
00:28:04.710 00:28:06.390 But I guess that's what this is for
00:28:06.390 00:28:06.630 as well.
00:28:07.710 00:28:08.790 Debug
.quote spolsky
00:28:08.790 00:28:10.050 code is not free, whether
00:28:10.050 00:28:11.550 proprietary or open source, even if
00:28:11.550 00:28:13.170 you don't pay cash dollars for it.
.endquote
00:28:13.170 00:28:15.930 This is also important for the AI
00:28:15.930 00:28:19.530 discussion, like, just take an
00:28:19.530 00:28:20.790 economic view on this, and like.
.p
00:28:23.730 00:28:25.230 The backcode is not free. Weather
00:28:25.230 00:28:26.790 propriety or open source or
00:28:26.790 00:28:28.050 operators. I might add
.quote spolsky
00:28:28.050 00:28:28.710 even if you
00:28:28.710 00:28:30.990 don't pay cash dollars for it. It has
00:28:30.990 00:28:33.030 opportunity cost, and it has time
00:28:33.030 00:28:34.650 cost. There's
.endquote
00:28:34.650 00:28:35.250 a finite amount of
00:28:35.250 00:28:37.470 volunteer programming handling Talent
00:28:37.470 00:28:39.090 available for open source. Okay, so
00:28:39.090 00:28:40.590 this call paragraph.
.p
00:28:41.610 00:28:42.270 Right, this?
00:28:43.230 00:28:46.290 Not outdated, just the whole equation
00:28:46.290 00:28:47.250 changes, right?
00:28:49.170 00:28:51.150 So right now, you get this.
00:28:53.640 00:28:54.900 You get this argument by many people,
00:28:54.900 00:28:54.901 which.
00:28:56.400 00:28:59.160 Oh, we should don't use AI. Your
00:28:59.160 00:29:00.720 software will be some. You will think
00:29:00.720 00:29:02.220 through everything way more.
00:29:03.420 00:29:04.560 Way more strongly, you will.
00:29:05.880 00:29:08.160 Not go down bad paths as much because?
.p
00:29:10.110 00:29:11.790 Um, you will have thought about what
00:29:11.790 00:29:13.110 you want to implement a lot.
00:29:14.370 00:29:15.510 And so I, I can speak from
00:29:15.510 00:29:16.590 experience. That's just not true.
00:29:16.590 00:29:19.650 Like, I spend so much time on stupid
00:29:19.650 00:29:20.490 little projects that.
00:29:22.110 00:29:24.870 That I had to implement a lot of code
00:29:24.870 00:29:27.150 for. And then, after five years
00:29:27.150 00:29:27.630 later, I'm.
.p
00:29:28.530 00:29:30.150 Why did I spend so much time on this?
00:29:30.150 00:29:31.170 I could have done so many other
00:29:31.170 00:29:32.910 things in that time, right? I can
00:29:32.910 00:29:34.530 say, oh, girls. I learned something
00:29:34.530 00:29:36.150 by doing that. Like, I learned this
00:29:36.150 00:29:37.890 program language or this framework or
00:29:37.890 00:29:38.130 whatever.
00:29:39.210 00:29:41.250 But if you think about it, you could
00:29:41.250 00:29:41.850 have, like?
.p
00:29:42.930 00:29:44.130 I could have created a company or
00:29:44.130 00:29:45.030 maybe also have.
00:29:46.680 00:29:48.540 Had other things to learn in the in
00:29:48.540 00:29:48.960 the meantime, right?
00:29:51.300 00:29:53.940 So, so the opportunity cost thing is
00:29:53.940 00:29:55.800 really overlooked. And that's also
00:29:55.800 00:29:55.980 because?
00:29:57.300 00:29:57.720 It's.
00:29:59.790 00:30:00.690 You know, it's personal for most
00:30:00.690 00:30:02.010 people like they spend a lot of time
00:30:02.010 00:30:02.970 on this stuff.
.p
00:30:02.970 00:30:05.610 And when somebody tells them, oh, you
00:30:05.610 00:30:06.630 could have done like five other
00:30:06.630 00:30:07.830 things in the same time, and you
00:30:07.830 00:30:10.230 would have learned like 10x more.
00:30:11.250 00:30:12.450 If you had tried out more things?
00:30:13.590 00:30:14.970 Because you had this machine that
00:30:14.970 00:30:17.130 will just give you the ability to
00:30:17.130 00:30:17.970 experiment a lot faster.
00:30:19.170 00:30:20.910 They would probably be, like, huh,
00:30:20.910 00:30:21.930 why did I waste all this time, right?
.p
00:30:23.070 00:30:25.470 Which is also it's a spiritual thing, but?
00:30:26.610 00:30:29.010 Yeah, just enjoy things as they come
00:30:29.010 00:30:30.090 and don't.
.p
00:30:35.040 00:30:36.180 Don't cry over lost.
00:30:38.250 00:30:40.470 Potential potentiality, but also
00:30:40.470 00:30:41.310 don't be stupid, like.
00:30:42.630 00:30:44.190 Like, unless you really have fun
00:30:44.190 00:30:47.551 writing code by hand, maybe to the?
00:30:48.750 00:30:49.230 Like, maybe?
00:30:51.390 00:30:54.030 Like, do the math and think. Is this
00:30:54.030 00:30:54.390 really worth it?
00:30:55.980 00:30:57.120 And then maybe do something and go.
00:30:57.120 00:30:58.740 Make something from wood or something
00:30:58.740 00:30:59.160 instead.
.p
00:31:03.300 00:31:04.440 Touchgrass.
00:31:05.820 00:31:07.980 Okay, so back, there's a final.
00:31:07.980 00:31:08.340 There's
.quote spolsky
00:31:08.340 00:31:09.960 a finite amount of volunteer
00:31:09.960 00:31:11.520 programming Talent available for open
00:31:11.520 00:31:12.960 source work, and each open source
00:31:12.960 00:31:14.280 project competes with every other
00:31:14.280 00:31:15.660 open source project for the same
00:31:15.660 00:31:18.060 limited programming resource, and
00:31:18.060 00:31:19.980 only the sexiest projects really have
00:31:19.980 00:31:21.840 more volunteer developers that they
00:31:21.840 00:31:22.080 can use.
.p
00:31:22.080 00:31:24.480 Summarize I'm not very impressed by
00:31:24.480 00:31:26.580 people who try to prove wild economic
00:31:26.580 00:31:28.560 things about free as and beer
00:31:28.560 00:31:30.060 software because they're just getting
00:31:30.060 00:31:32.040 divided by zeros errors. As far as
00:31:32.040 00:31:32.400 I'm concerned.
00:31:33.660 00:31:35.400 Open source is not exempt from the
00:31:35.400 00:31:36.780 laws of gravity or economics.
.endquote
00:31:39.090 00:31:40.290 That's something I'm realizing more
00:31:40.290 00:31:40.950 and more that, like.
.p
00:31:43.350 00:31:45.630 Things that happen. Like, okay, I
00:31:45.630 00:31:47.070 shouldn't say things so much, like many.
.p
00:31:50.160 00:31:50.760 Observable.
00:31:52.020 00:31:54.000 Symptoms that you can.
00:31:56.100 00:31:56.881 Like Intuit, maybe?
00:31:58.200 00:31:59.281 Actually speak to some.
00:32:01.260 00:32:03.900 Maybe hard to hard to actually.
00:32:05.340 00:32:05.400 Um.
00:32:07.260 00:32:09.840 Put numbers on, but also like, deeper.
00:32:11.100 00:32:12.780 Fundamental things that?
00:32:14.640 00:32:16.860 A laws of nature in a way that is more.
00:32:18.060 00:32:20.640 A lot more vague or.
00:32:21.660 00:32:24.840 Um, cloudy, then nebulous.
00:32:26.220 00:32:27.240 Oh, gravity!
00:32:28.560 00:32:30.480 Towards the Gravity center, the
00:32:30.480 00:32:32.460 center of gravity, right or, like,
00:32:32.460 00:32:35.340 um, these are the formulas for space-time.
.p
00:32:36.360 00:32:36.421 But they're.
00:32:38.880 00:32:39.960 I guess they're economic, right? So,
00:32:39.960 00:32:42.000 so you have these chaotic systems,
00:32:42.000 00:32:43.440 and you want to figure out what the?
00:32:44.520 00:32:47.580 Um, what the trends are? What happens
00:32:47.580 00:32:50.040 if you change these one or two or
00:32:50.040 00:32:51.840 maybe five parameters and maybe?
.p
00:32:55.350 00:32:57.030 Um, get an intuition about how things
00:32:57.030 00:32:58.590 interact with each other.
00:32:59.910 00:33:01.530 And like this. Oh God, that's I think
00:33:01.530 00:33:03.150 it was called courthouse now. No.
.p
00:33:10.080 00:33:10.800 Huh? Maybe not.
00:33:13.020 00:33:13.800 God wins law.
.p
00:33:21.720 00:33:23.280 No, that's the Nazi one.
00:33:25.500 00:33:27.600 Go to Winslaws the Nazi one.
.p
00:33:30.780 00:33:32.460 God has law.
00:33:33.720 00:33:35.220 No, that's the target one.
.p
00:33:39.720 00:33:40.381 That's also very important is.
00:33:42.420 00:33:42.720 Um.
00:33:44.160 00:33:45.480 Whenever something becomes a Target,
00:33:45.480 00:33:48.060 it will be games and can't be used as
00:33:48.060 00:33:48.600 a measure anymore.
00:33:49.740 00:33:50.581 Are we seeing that a lot?
00:33:53.040 00:33:54.120 A Peter Principle.
00:33:55.320 00:33:57.180 Mcnamara fallacy.
00:33:59.760 00:34:02.340 I, I can't. I can't find the law
00:34:02.340 00:34:03.060 right now.
.p
00:34:12.660 00:34:15.960 Words law de Morgan's law? No, those
00:34:15.960 00:34:17.640 are all, not the laws I'm looking for.
.p
00:34:25.650 00:34:28.410 The Germans Paradox. Now, we now we
00:34:28.410 00:34:29.250 are at the right one.
.p
00:34:34.890 00:34:37.230 So, instead of leading to a fall in
00:34:37.230 00:34:38.910 consumption, if you improve
00:34:38.910 00:34:40.830 something, it will lead to a rising
00:34:40.830 00:34:41.190 consumption.
00:34:42.270 00:34:42.390 But I.
00:34:43.590 00:34:44.910 In like?
00:34:45.990 00:34:46.650 Not make that?
00:34:47.790 00:34:49.770 Maybe like a linear, or like, maybe?
00:34:51.180 00:34:52.680 Um, logarithmic curve.
00:34:53.580 00:34:55.380 I would make that into, like a step
00:34:55.380 00:34:55.740 function, right?
00:34:57.000 00:34:58.140 So there exists.
.p
00:35:01.440 00:35:02.940 On the curve of something getting
00:35:02.940 00:35:05.700 cheaper where suddenly adoption
00:35:05.700 00:35:07.740 explodes because a whole new use case
00:35:07.740 00:35:09.960 is being found that wasn't possible
00:35:09.960 00:35:11.700 before and suddenly it's possible.
.p
00:35:14.820 00:35:17.940 Like, for example, roads being paved
00:35:17.940 00:35:18.300 over right.
00:35:19.560 00:35:20.580 Something I learned today is that?
00:35:22.200 00:35:26.220 Bicycles, bicycles, cyclists, were
00:35:26.220 00:35:28.320 the people who first argued for
00:35:28.320 00:35:29.880 Paving overalls because.
00:35:32.520 00:35:35.100 Um, yeah, it's. It's a slog to drive
00:35:35.100 00:35:36.960 a bicycle on unpaved road, so
00:35:36.960 00:35:38.700 basically, this were the ones that
00:35:38.700 00:35:39.300 set. Now, we should.
.p
00:35:39.300 00:35:40.980 We should probably pave all of the
00:35:40.980 00:35:42.120 city Interiors.
00:35:43.020 00:35:45.900 And then, suddenly, bicycle user exploded.
00:35:46.800 00:35:46.801 Which?
00:35:48.240 00:35:51.000 Cars possible, right? And then cars
00:35:51.000 00:35:52.680 replace bicycles mostly, but that was
00:35:52.680 00:35:52.980 a different thing.
00:35:55.740 00:35:55.800 Um.
00:35:57.600 00:35:59.280 Open source is not exempt from the
00:35:59.280 00:36:00.900 laws of gravity or economics
00:36:00.900 00:36:04.080 economics economics. We saw this with
00:36:04.080 00:36:05.580 easel, our stitcher,
.quote spolsky
00:36:05.580 00:36:05.880 the company
00:36:05.880 00:36:07.800 formerly known as VA Linux and a lot
00:36:07.800 00:36:08.340 of other attempts.
.endquote
.p
00:36:08.340 00:36:10.560 This is, this is probably a history
00:36:10.560 00:36:11.100 lesson because.
00:36:12.660 00:36:13.860 I mean us digital.
00:36:14.760 00:36:16.680 Tells like I I, I've heard the name,
00:36:16.680 00:36:18.600 but I have no idea what easel is.
00:36:19.320 00:36:20.940 I have no idea what the company
00:36:20.940 00:36:23.100 formulated known as VA Linux.
00:36:23.820 00:36:25.560 This, but I guess the pattern is the
00:36:25.560 00:36:27.420 same something is going on, when,
.quote spolsky
00:36:27.420 00:36:29.580 which very few people in the open
00:36:29.580 00:36:31.620 source world really understand a lot
00:36:31.620 00:36:33.660 of very large public companies, which
00:36:33.660 00:36:35.820 with responsibilities to maximize
00:36:35.820 00:36:38.220 shareholder value, are investing a
00:36:38.220 00:36:39.840 lot of money and supporting open
00:36:39.840 00:36:42.360 source software, people paying large
00:36:42.360 00:36:43.740 teams of programmers to work on it.
.p
00:36:44.820 00:36:46.080 And that's what the principle of
00:36:46.080 00:36:48.180 complements explains once again.
00:36:48.180 00:36:50.340 Demand for a product increases when
00:36:50.340 00:36:52.500 the price of its complements decreases.
00:36:53.580 00:36:54.960 In general, the company's strategic
00:36:54.960 00:36:57.120 interest is going to be together the
00:36:57.120 00:36:58.740 price of their compliments as low as
00:36:58.740 00:37:00.840 possible, the lowest theoretically
00:37:00.840 00:37:02.460 sustainable price would be the
00:37:02.460 00:37:03.240 commodity price.
.p
00:37:03.240 00:37:05.520 The price that arises when you have a
00:37:05.520 00:37:06.720 bunch of competitors offering
00:37:06.720 00:37:08.580 indistinguishable Goods. So,
.endquote
00:37:08.580 00:37:09.000 that's
00:37:09.000 00:37:09.600 what I was talking about.
00:37:11.280 00:37:12.480 At the very beginning, it's like a
00:37:12.480 00:37:13.560 commodity is a thing.
00:37:14.760 00:37:17.040 That is marginal price very low
00:37:17.040 00:37:19.740 because lots of people are competing
00:37:19.740 00:37:23.100 for selling the most the most of it,
00:37:23.100 00:37:23.340 right.
.p
00:37:23.340 00:37:25.680 And also, it's very exchangeable, so
00:37:25.680 00:37:28.500 if you sell it, somebody else says
00:37:28.500 00:37:30.420 they're gonna try to make a better
00:37:30.420 00:37:33.420 price for it. Create like a better
00:37:33.420 00:37:36.300 supply chain so they can sell at a
00:37:36.300 00:37:38.160 lower price than you. And then you
00:37:38.160 00:37:38.460 have to.
00:37:40.830 00:37:41.070 Try to.
00:37:42.570 00:37:43.350 I'll compute them again.
00:37:44.250 00:37:45.870 Which drives the prices down until
00:37:45.870 00:37:47.490 things are very commoditized.
.p
00:37:48.270 00:37:50.130 So, for example, in Germany, we have
00:37:50.130 00:37:52.770 these Supermarket chains like Lidl
00:37:52.770 00:37:53.310 and Aldi.
00:37:54.330 00:37:56.070 And they're in a bitter fight right
00:37:56.070 00:37:58.590 now. For butter prices of all things?
00:37:59.850 00:38:01.830 Which means they're really, really,
00:38:01.830 00:38:03.870 like the bother prices are really in the.
00:38:04.950 00:38:05.010 Um.
00:38:06.690 00:38:08.131 How do you say that they're like?
00:38:09.030 00:38:11.610 Yeah, you can buy a lot of butter for
00:38:11.610 00:38:12.330 a few Euros.
.p
00:38:13.050 00:38:13.170 Um.
00:38:14.250 00:38:16.350 Which is probably also because they
00:38:16.350 00:38:17.850 cross Finance that from the other
00:38:17.850 00:38:18.270 products, but?
00:38:19.710 00:38:20.370 Basically, I want.
00:38:21.390 00:38:23.490 I don't actually know why they do this.
00:38:24.510 00:38:25.890 Because, for example, in France, you
00:38:25.890 00:38:27.630 have the opposite that you have a few
00:38:27.630 00:38:30.510 giant giant.
00:38:31.530 00:38:33.930 Grocery chains that basically all
00:38:33.930 00:38:36.090 collude to just set the prices at
00:38:36.090 00:38:36.450 very high.
.p
00:38:37.530 00:38:38.730 Best prices so?
00:38:40.050 00:38:41.671 Your butter is gonna cost three $4
00:38:41.671 00:38:46.410 per 500 G instead of 150 as it
00:38:46.410 00:38:46.830 currently costs.
00:38:47.850 00:38:48.330 In Germany.
00:38:49.230 00:38:50.850 Or, for example, another example
00:38:50.850 00:38:52.230 would be the the mobile phone
00:38:52.230 00:38:53.610 providers in France.
00:38:54.630 00:38:55.830 Also colluded to have.
00:38:56.790 00:38:57.870 Very, very s*****.
00:39:00.270 00:39:04.770 Monthly, like gigabytes per like euro
00:39:04.770 00:39:05.310 per gigabyte.
.p
00:39:06.330 00:39:06.510 Um.
.p
00:39:10.320 00:39:11.460 How do you say like their contracts
00:39:11.460 00:39:12.840 were super s***** and then one
00:39:12.840 00:39:14.880 billionaire came along and built this
00:39:14.880 00:39:16.200 call company called free, which?
00:39:18.450 00:39:20.190 B**** outbid, Everybody by, like,
00:39:20.190 00:39:22.830 100%, so it was like 50%.
00:39:24.090 00:39:25.410 Like it was half of what other
00:39:25.410 00:39:28.350 people, what other providers gave you
00:39:28.350 00:39:29.730 for the same money, and then
00:39:29.730 00:39:30.990 everybody else had to start.
.p
00:39:33.000 00:39:33.240 Pulling.
00:39:34.140 00:39:35.820 Like lowering the own prices and
00:39:35.820 00:39:38.520 improving the offerings which started.
00:39:39.480 00:39:41.340 Cycle where everybody outbid the
00:39:41.340 00:39:42.240 other company and?
00:39:43.770 00:39:44.850 And the process went lower and lower,
00:39:44.850 00:39:46.230 and nowadays is basically
00:39:46.230 00:39:48.030 commoditized, like, doesn't really
00:39:48.030 00:39:49.950 matter where you go or the they are.
00:39:49.950 00:39:51.510 They have very low margins for this stuff.
.p
00:39:52.590 00:39:54.090 And you basically pay a nice price
00:39:54.090 00:39:56.070 for mobile phones.
00:39:56.970 00:39:58.170 Mobile phone contracts.
00:39:59.010 00:40:00.930 Which is not the case in Germany at
00:40:00.930 00:40:02.370 all. I think it might have started to
00:40:02.370 00:40:05.610 happen because you.
00:40:06.810 00:40:09.450 Basically forced everybody to allow
00:40:09.450 00:40:11.730 roaming in all of the use, so you can
00:40:11.730 00:40:13.890 just have a contract in another country.
.p
00:40:15.150 00:40:17.490 And use it in Germany as before, as
00:40:17.490 00:40:19.410 if it was a local one, which means
00:40:19.410 00:40:21.690 German providers are suddenly forced
00:40:21.690 00:40:24.390 to actually compete with with each
00:40:24.390 00:40:25.410 other, and like with all of the other
00:40:25.410 00:40:27.870 providers in Europe to a certain point.
00:40:29.070 00:40:30.690 Which is why prices are suddenly.
00:40:31.650 00:40:33.510 Turning into commodity prices, which
00:40:33.510 00:40:34.230 is where you want to go.
.p
00:40:34.230 00:40:35.370 If you are like?
00:40:36.630 00:40:38.790 Near liberal ecosystem, but for some
00:40:38.790 00:40:40.770 reason it doesn't always happen.
.p
00:40:45.720 00:40:46.260 So?
00:40:47.700 00:40:48.900 Smart companies tried
.quote spolsky
00:40:48.900 00:40:49.680 to commoditize
00:40:49.680 00:40:50.760 their products complements.
00:40:51.900 00:40:53.580 If you can do this, demand for your
00:40:53.580 00:40:54.480 product will increase, and you will
00:40:54.480 00:40:56.160 be able to charge more and make more.
00:40:57.420 00:40:59.520 When IBM designed a PC architecture,
00:40:59.520 00:41:01.560 they used off-the-shelf Parts instead
00:41:01.560 00:41:03.240 of custom parts, and they carefully
00:41:03.240 00:41:04.800 documented the interface between the
00:41:04.800 00:41:06.660 parts evolutionary.
.p
00:41:06.660 00:41:09.120 IBM PC technical reference manual.
00:41:09.120 00:41:11.820 Why, so that other manufacturers
00:41:11.820 00:41:13.380 could join the party as long as you
00:41:13.380 00:41:15.180 match the interface? You can be used
00:41:15.180 00:41:18.360 in PCS. IBM's goal was to commoditize
00:41:18.360 00:41:20.220 the add-in Market, which is a
00:41:20.220 00:41:21.540 complement of the PC market.
00:41:22.680 00:41:23.880 And they did this quite successfully.
00:41:25.020 00:41:27.000 Within a short time, squillions of
00:41:27.000 00:41:28.260 companies sprang up, offering more
00:41:28.260 00:41:30.420 memory, offering memory cards, hard
00:41:30.420 00:41:32.940 drives, graphics cards, printers, Etc.
.p
00:41:32.940 00:41:35.820 Cheap add-ins meant more demand for PCs.
00:41:36.720 00:41:38.700 When IBM licensed the operating
00:41:38.700 00:41:40.440 system PC dos from Microsoft,
00:41:40.440 00:41:42.360 Microsoft was very careful not to
00:41:42.360 00:41:44.820 sell an exclusive license. They made
00:41:44.820 00:41:46.980 it possible for Microsoft. This made
00:41:46.980 00:41:48.660 it possible for Microsoft to license
00:41:48.660 00:41:50.760 the same thing to compact and the
00:41:50.760 00:41:52.980 other hundreds of oems were legally
00:41:52.980 00:41:55.920 cloned the IBM PC using IBM's own
00:41:55.920 00:41:56.340 documentation.
.p
00:41:56.340 00:41:59.220 Microsoft's goal was to commoditize
00:41:59.220 00:42:01.920 the PC market very soon. The PC
00:42:01.920 00:42:04.500 itself was basically a commodity with
00:42:04.500 00:42:06.000 ever decreasing prices consistent.
00:42:06.660 00:42:08.460 Increase in power and fierce margins
00:42:08.460 00:42:10.020 that make it extremely hard to make a
00:42:10.020 00:42:12.420 profit. The low prices, of course,
00:42:12.420 00:42:14.760 increase the amount increased. Demand
00:42:14.760 00:42:16.980 for PCS meant increased demand for
00:42:16.980 00:42:18.480 their complement and the stores.
.p
00:42:19.260 00:42:21.180 All else being equal. The greater the
00:42:21.180 00:42:22.860 amount for a product, the more money
00:42:22.860 00:42:23.700 it makes for you.
00:42:24.960 00:42:26.520 And that's why Bill Gates can buy
00:42:26.520 00:42:27.540 Sweden, and you can't.
.endquote
.p
00:42:33.180 00:42:35.760 So, this is also why valve is trying
00:42:35.760 00:42:36.060 to create.
00:42:37.620 00:42:40.020 Headsets and why they create like?
00:42:41.640 00:42:43.440 Based on Linux operating systems.
00:42:44.640 00:42:45.660 And why they release.
00:42:47.160 00:42:49.200 Um, the steam box, which is like,
00:42:49.200 00:42:51.660 basically, just a PC with their Steam
00:42:51.660 00:42:53.100 OS on it. And, like, very nice small
00:42:53.100 00:42:55.500 form factor from form factor.
.p
00:42:58.440 00:42:59.880 Because they make money selling games
00:42:59.880 00:43:02.700 on Steam and like making it so that
00:43:02.700 00:43:04.560 their steam platform is taking 30
00:43:04.560 00:43:06.000 percent of everything that people sell.
00:43:08.310 00:43:08.430 Um.
00:43:09.540 00:43:10.860 So everything that increases the?
00:43:13.500 00:43:15.840 Pc gaming Market basically is a win
00:43:15.840 00:43:18.240 for their for their re.
00:43:19.200 00:43:20.460 And also their profits.
00:43:22.020 00:43:24.120 Um, which is why they sell those
00:43:24.120 00:43:24.900 steam boxes then.
.p
00:43:24.900 00:43:26.640 They don't actually sell them for
00:43:26.640 00:43:28.080 under value. They sell them for
00:43:28.080 00:43:30.300 exactly what they, what it cost them
00:43:30.300 00:43:31.980 to, make them, and also plus like a
00:43:31.980 00:43:32.940 little bit of modern.
00:43:33.720 00:43:34.800 Which is why they are so expensive,
00:43:34.800 00:43:36.061 but also other people are going to
00:43:36.061 00:43:38.040 come around. Other companies are
00:43:38.040 00:43:40.140 gonna start creating their own steam
00:43:40.140 00:43:40.920 boxes, which is already happening,
00:43:40.920 00:43:43.320 and then everybody can just install
00:43:43.320 00:43:45.120 Steamworks on their devices and make
00:43:45.120 00:43:46.080 their own steam box.
.p
00:43:46.080 00:43:48.360 So, this is like really increasing
00:43:48.360 00:43:50.580 the margin of the steam ecosystem
00:43:50.580 00:43:52.740 that the the value of the steam ecosystem.
00:43:53.880 00:43:55.320 And there's like, no real reason for
00:43:55.320 00:43:58.800 people to use any other platform at
00:43:58.800 00:43:59.100 this point.
00:44:00.960 00:44:02.580 But also they don't care if people
00:44:02.580 00:44:03.900 like if it's possible that people
00:44:03.900 00:44:05.040 install like the.
.p
00:44:06.180 00:44:09.720 The EA platform or their the the Gog
00:44:09.720 00:44:11.340 platform on the steam box because
00:44:11.340 00:44:12.600 they know that, like, they have the?
00:44:14.400 00:44:16.080 Vast Network Advantage right now!
00:44:17.160 00:44:20.220 So their share of the pie is gonna
00:44:20.220 00:44:22.680 grow faster than other platforms
00:44:22.680 00:44:24.300 share of the pie and the whole pile
00:44:24.300 00:44:24.720 is gonna grow, right?
.p
00:44:24.720 00:44:27.180 So, as long as the whole pie can grow?
00:44:28.710 00:44:28.830 Um.
00:44:30.240 00:44:32.100 Consistently, and also maybe the
00:44:32.100 00:44:33.480 slice of the pie grows.
00:44:34.260 00:44:36.420 Um, that's a win for them.
.p
00:44:41.400 00:44:43.680 This year, we're in 2002. Remember,
00:44:43.680 00:44:45.000 Microsoft's trying to do it again?
00:44:46.260 00:44:48.420 The new game console, the new game
00:44:48.420 00:44:50.700 console, the Xbox like history
00:44:50.700 00:44:53.160 Rhymes, right valve, is realizing the
00:44:53.160 00:44:54.660 new game console, which is the steam
00:44:54.660 00:44:58.080 box in 2026. In 2002, Microsoft is
00:44:58.080 00:44:58.440 doing it.
.quote spolsky
00:44:59.820 00:45:02.700 The Xbox uses commodity PC Hardware
00:45:02.700 00:45:03.780 instead of custom parts.
.p
00:45:04.800 00:45:05.580 The theory.
00:45:06.660 00:45:08.100 Was that the commodity Hardware gets
00:45:08.100 00:45:09.960 cheaper every year, so the Xbox could
00:45:09.960 00:45:10.800 write down the prices?
00:45:11.880 00:45:13.140 Unfortunately, it seems to have
00:45:13.140 00:45:15.780 backfired. Apparently, commodity PC
00:45:15.780 00:45:17.100 Hardware has already been squeezed
00:45:17.100 00:45:19.620 down to commodity prices, and so the
00:45:19.620 00:45:21.720 price of Mac in Xbox isn't declining
00:45:21.720 00:45:23.040 as fast as Microsoft would like.
.p
00:45:24.300 00:45:25.680 The other part of Microsoft's Xbox
00:45:25.680 00:45:28.200 strategy was to use DirectX the
00:45:28.200 00:45:29.580 graphics Library that can be used to
00:45:29.580 00:45:31.260 write code that runs on all kinds of
00:45:31.260 00:45:32.460 video chips.
00:45:33.420 00:45:35.040 The goal here is to make the video
00:45:35.040 00:45:37.320 chip a commodity to lower its price
00:45:37.320 00:45:40.620 so that more games are sold where the
00:45:40.620 00:45:41.280 real profits occur.
.p
00:45:42.360 00:45:44.280 And why don't the video chip vendors
00:45:44.280 00:45:45.720 of the world? Try to commoditize the
00:45:45.720 00:45:46.380 games somehow.
00:45:47.580 00:45:50.100 That's a lot harder. If the game Halo
00:45:50.100 00:45:51.720 is selling like crazy, it doesn't
00:45:51.720 00:45:54.180 really have any substitutes. You're
00:45:54.180 00:45:55.560 not going to go to the movie
.endquote
00:45:55.560 00:45:55.860 theater
00:45:55.860 00:45:58.320 to see Star Wars attack on the of the
00:45:58.320 00:46:01.080 Clones, and this, oh man, Star Wars
00:46:01.080 00:46:02.940 attack of the Clones was was current,
00:46:02.940 00:46:06.600 like was current in 2000, that's so crazy.
.p
00:46:06.600 00:46:07.920 You're not going to the movie
00:46:07.920 00:46:08.760 theater, you
.quote spolsky
00:46:08.760 00:46:09.720 see, Star Wars attack of
00:46:09.720 00:46:11.280 the Clones and the side. Instead, he
00:46:11.280 00:46:13.380 would be satisfied with a Woody Allen
00:46:13.380 00:46:15.060 movie. They may both be great movies,
00:46:15.060 00:46:16.680 but they're not perfect substitutes.
00:46:17.520 00:46:19.740 Now, who would you rather be a game
00:46:19.740 00:46:20.100 publisher
.endquote
00:46:20.100 00:46:21.480 video chat vendor?
.p
00:46:23.100 00:46:25.200 So, this is something that big
00:46:25.200 00:46:25.620 entertainment.
00:46:26.760 00:46:28.320 Like, in the meantime, big
00:46:28.320 00:46:29.520 entertainment has really tried to
00:46:29.520 00:46:30.540 make this happen.
00:46:32.730 00:46:35.190 Um, like you can, probably at this
00:46:35.190 00:46:36.690 point, replace most Netflix.
00:46:38.070 00:46:40.050 Drama series or whatever series with
00:46:40.050 00:46:41.550 a series from the same kind of.
00:46:43.440 00:46:45.360 Jeong, and you will have a perfectly
00:46:45.360 00:46:47.100 mediocre.
.p
00:46:47.100 00:46:49.920 B******* experience, and I guess most
00:46:49.920 00:46:51.000 people are happy with that.
00:46:52.320 00:46:53.880 Or you can. I mean, you can go to any
00:46:53.880 00:46:55.620 kind of superhero movie and you will
00:46:55.620 00:46:57.300 receive a perfectly Bland.
00:46:58.260 00:47:00.420 Default experience, or you can watch
00:47:00.420 00:47:04.920 any kind of Star Wars series game.
00:47:06.300 00:47:08.340 Lego Star Wars whatever and you will
00:47:08.340 00:47:08.940 also get this kind of.
.p
00:47:10.320 00:47:12.840 Star was feeling from it, so they are
00:47:12.840 00:47:14.040 really trying to make these things
00:47:14.040 00:47:14.940 Commodities.
00:47:15.960 00:47:17.820 And I think, because it's Arch,
00:47:17.820 00:47:19.260 they're gonna it's gonna backfire,
00:47:19.260 00:47:19.860 like they're gonna.
00:47:20.760 00:47:22.560 Yeah, they're still making a lot of
00:47:22.560 00:47:23.160 money from this, but
00:47:24.420 00:47:25.800 People are getting really sick of
00:47:25.800 00:47:26.700 this kind of trying to.
.p
00:47:28.260 00:47:29.041 Commoditize art.
00:47:30.690 00:47:32.910 And especially with AR with AIR. Now
00:47:32.910 00:47:33.390 people are going to be.
00:47:35.430 00:47:38.730 Yeah, super linearly. Be annoyed by
00:47:38.730 00:47:40.050 this, and I think the newer
00:47:40.050 00:47:41.190 generation is already going back to
00:47:41.190 00:47:42.990 trying to find the interesting.
00:47:44.310 00:47:47.010 In this stuff and maybe not spend all
00:47:47.010 00:47:47.670 of that money on.
00:47:49.590 00:47:52.110 Yeah, AAA quadruple A whatever.
.p
00:47:56.130 00:47:57.510 One can hope.
00:47:58.230 00:48:00.270 But yeah, so like some things are not
00:48:00.270 00:48:01.710 easy to commoditize.
00:48:02.610 00:48:04.290 Like big entertainment.
00:48:05.370 00:48:06.811 Complexes are always trying to do
00:48:06.811 00:48:08.670 that, like with rock stars or with
00:48:08.670 00:48:10.890 MTV stars or whatever, but
00:48:12.390 00:48:14.250 Um, the pendulum always swings back
00:48:14.250 00:48:14.910 pretty hard, I think.
00:48:16.290 00:48:18.210 Whereas with the technology, it's a
00:48:18.210 00:48:18.630 lot easier to.
.p
00:48:20.190 00:48:21.810 Create Commodities like?
00:48:23.250 00:48:26.190 Whether company a or Company B
00:48:26.190 00:48:27.750 supplies with your aluminum is
00:48:27.750 00:48:29.130 probably doesn't really matter to you.
00:48:30.930 00:48:31.530 Um.
00:48:33.030 00:48:34.350 And I mean the ultimate Commodities
00:48:34.350 00:48:35.550 energy, right? This electricity,
00:48:35.550 00:48:36.330 which is why?
00:48:37.530 00:48:39.870 Solar Revolution and the EV
00:48:39.870 00:48:41.370 Revolution and the Battery Revolution
00:48:41.370 00:48:41.970 are so important.
00:48:43.350 00:48:46.770 And the the obvious Next Step, but
00:48:46.770 00:48:48.390 it's like, it's not even that.
.p
00:48:50.160 00:48:51.420 It's happening because it's the right
00:48:51.420 00:48:53.580 thing to do in scare quotes, but it's
00:48:53.580 00:48:55.500 and well, because the work liberers
00:48:55.500 00:48:57.420 want you to do it. But it's just
00:48:57.420 00:48:57.720 happening because?
00:48:59.160 00:49:01.500 There's like economic incentives to
00:49:01.500 00:49:03.120 do it, and it's happening despite the
00:49:03.120 00:49:03.720 best efforts of.
00:49:05.040 00:49:06.120 Big oil and stuff.
00:49:08.040 00:49:09.060 Try to avoid this.
.p
00:49:12.450 00:49:13.470 Which is also why Trump?
00:49:14.250 00:49:17.130 Mr Donald Trump first mentioned EPE
00:49:17.130 00:49:17.431 is the.
00:49:18.570 00:49:18.990 Accidentally.
00:49:20.310 00:49:23.550 Um, help climate change by invading Iran.
.p
00:49:26.790 00:49:27.810 Which is kind of funny if you think
00:49:27.810 00:49:30.630 about it, but she's probably like his
00:49:30.630 00:49:30.990 best.
00:49:32.430 00:49:32.730 Um.
00:49:33.750 00:49:35.970 Yeah, this Invasion was probably the
00:49:35.970 00:49:37.410 best thing to happen to the globe,
00:49:37.410 00:49:40.230 and it's probably more important than
00:49:40.230 00:49:41.370 many other presidents.
00:49:42.390 00:49:44.970 In a positive way, even although he's
00:49:44.970 00:49:46.470 obviously a b******* president, but?
.p
00:49:47.370 00:49:48.570 This probably helps.
00:49:49.530 00:49:50.970 More than people realize.
.p
00:49:56.340 00:49:56.520 Yep.
.p
00:49:59.910 00:50:02.010 In it by by complete accident, of course.
00:50:04.710 00:50:04.711 Okay,
.quote spolsky
00:50:04.711 00:50:06.870 headline IBM spends Millions to
00:50:06.870 00:50:09.330 develop open source software myth.
00:50:09.330 00:50:11.310 They're doing this because Lou
00:50:11.310 00:50:13.650 justner, read the new Manifesto and
00:50:13.650 00:50:14.670 decided he doesn't actually like
00:50:14.670 00:50:17.130 capitalism reality. They're doing
00:50:17.130 00:50:19.170 this because IBM is becoming I.T
00:50:19.170 00:50:21.930 Consulting. Company ID Consulting is
00:50:21.930 00:50:23.610 a complement of Enterprise software.
00:50:23.610 00:50:26.490 Thus, IBM needs to commoditize
00:50:26.490 00:50:27.870 Enterprise software and the best way
00:50:27.870 00:50:29.550 to do this is by supporting open
00:50:29.550 00:50:31.170 source lo and behold.
.p
00:50:31.170 00:50:32.910 Their Consulting division is winning
00:50:32.910 00:50:33.810 back with a strategy.
00:50:34.980 00:50:37.440 Headline Netscape open sources their
00:50:37.440 00:50:39.960 web browser. Oh, this is ancient
00:50:39.960 00:50:41.760 history myth. They're doing this to
00:50:41.760 00:50:43.380 get free source code contributions
00:50:43.380 00:50:45.480 from people in cyber cafes in New
00:50:45.480 00:50:47.400 Zealand reality. They're doing this
00:50:47.400 00:50:48.720 to commoditize the web browser.
00:50:49.860 00:50:51.720 This has been Netscape strategy from
00:50:51.720 00:50:52.260 day one.
.p
00:50:52.260 00:50:53.940 Have a look at the very first
00:50:53.940 00:50:55.380 Netscape press release.
00:50:56.400 00:50:58.740 The browser's freeware Netscape gave
00:50:58.740 00:51:00.060 gave away the browser so they could
00:51:00.060 00:51:02.880 make money on servers. Browsers and
00:51:02.880 00:51:04.620 servers are classic, complements the
00:51:04.620 00:51:05.460 cheaper
.endquote
00:51:05.460 00:51:06.900 to actually browsers and
00:51:06.900 00:51:08.880 Advers are complements, but I mean
00:51:08.880 00:51:10.860 guess back then, people thought that
00:51:10.860 00:51:12.120 browsing server was a compliment,
00:51:12.120 00:51:13.860 which didn't really work out.
.p
00:51:13.860 00:51:16.260 Which is, I guess why Google won out?
00:51:17.220 00:51:19.140 Because they realized around the same
00:51:19.140 00:51:19.320 time that.
00:51:20.520 00:51:22.620 A search engine can make a lot more money.
.quote spolsky
00:51:24.900 00:51:25.200 The cheaper
00:51:25.200 00:51:26.160 the browsers, the more
00:51:26.160 00:51:27.840 servers you sell. This was never as
00:51:27.840 00:51:29.520 true as it was October 1994.
00:51:31.440 00:51:32.880 Netscape was actually surprised when
00:51:32.880 00:51:34.980 MCI came to the door and dumped so
00:51:34.980 00:51:36.360 much money in their Labs that they
00:51:36.360 00:51:38.040 realized they could make money off
00:51:38.040 00:51:40.440 the browser too off of the browser too.
.p
00:51:40.440 00:51:42.660 This wasn't required by the business plan.
00:51:43.860 00:51:46.020 When Netscape released Mozilla as
00:51:46.020 00:51:47.700 open source, it was because they saw
00:51:47.700 00:51:49.380 an opportunity to lower the cost of
00:51:49.380 00:51:51.660 developing the browser so they could
00:51:51.660 00:51:54.000 get the commodity benefit benefits as
00:51:54.000 00:51:54.720 a lower cost.
.p
00:51:58.830 00:52:00.690 Later, AOL Time Warner acquired
00:52:00.690 00:52:02.550 Netscape, the server software, which
00:52:02.550 00:52:04.290 was supposed to be the beneficiary of
00:52:04.290 00:52:06.150 commodity. Browsers wasn't doing all
00:52:06.150 00:52:09.390 that well and was jettisoned now. Why
00:52:09.390 00:52:11.190 would AOL Time Warner continue to
00:52:11.190 00:52:12.930 invest anything in open source?
00:52:14.070 00:52:15.690 It's an entertaining entertainment
00:52:15.690 00:52:18.330 company entertainment companies, so I
00:52:18.330 00:52:20.490 was talking about them are the
00:52:20.490 00:52:21.990 complement of entertainment delivery
00:52:21.990 00:52:24.090 platforms of all types, including web
00:52:24.090 00:52:24.450 browsers.
.p
00:52:24.450 00:52:27.330 So, this is the Netscape, the, the
00:52:27.330 00:52:29.610 not the Netscape, the Netflix story,
00:52:29.610 00:52:30.090 right?
00:52:30.990 00:52:32.790 Including web browsers. This joint
00:52:32.790 00:52:34.530 conglomerated strategies interest
00:52:34.530 00:52:36.930 strategic interest is to make
00:52:36.930 00:52:39.210 entertainment delivery web browsers,
00:52:39.210 00:52:41.010 a commodity for which nobody can
00:52:41.010 00:52:41.730 charge money.
.endquote
00:52:44.250 00:52:45.450 And then Internet Explorer happened,
00:52:45.450 00:52:47.250 and then everything was blocked for
00:52:47.250 00:52:47.790 about 10 years.
.p
00:52:49.980 00:52:50.040 Yeah.
00:52:50.940 00:52:51.840 So you can still?
00:52:54.330 00:52:56.790 Hindo that plan by having a product
00:52:56.790 00:52:58.770 that is widely used and stagnant.
00:53:00.120 00:53:02.100 Like Microsoft did, which is how they
00:53:02.100 00:53:05.580 made Windows happen until Windows
00:53:05.580 00:53:07.980 Vista happened. Yeah, so it's like
00:53:07.980 00:53:10.200 it's a struggle all the way down, right?
.p
00:53:13.740 00:53:15.120 So, it's interesting to read these
00:53:15.120 00:53:16.500 kind of articles 20 years later.
00:53:17.880 00:53:20.400 And then have the actual how these
00:53:20.400 00:53:22.200 things turned out. The the actual
00:53:22.200 00:53:22.860 history behind this?
00:53:23.940 00:53:25.680 And then compare it against the arguments.
00:53:26.760 00:53:26.820 Um.
00:53:27.900 00:53:28.440 And he goes on.
.quote spolsky
00:53:28.440 00:53:29.340 My argument is a
00:53:29.340 00:53:30.720 little bit tortured by the fact that
00:53:30.720 00:53:32.700 Internet Explorer is free as in beer.
.p
00:53:32.700 00:53:33.540 Oh, he was talking about this.
00:53:33.540 00:53:36.060 Microsoft wanted to make web browsers
00:53:36.060 00:53:38.160 a commodity, too, so they can sell
00:53:38.160 00:53:39.540 desktop and server operating systems,
00:53:39.540 00:53:41.520 so you can see here that two
00:53:41.520 00:53:42.840 conglomerates are trying to compete
00:53:42.840 00:53:45.660 for the same ish commodity, but for
00:53:45.660 00:53:46.260 different reasons.
00:53:47.760 00:53:49.500 They went a step further and
00:53:49.500 00:53:50.700 delivered a collection of components
00:53:50.700 00:53:52.500 which anyone could use to throw
00:53:52.500 00:53:53.700 together web browser.
.p
00:53:53.700 00:53:57.000 New planet, AOL, and Juno use these
00:53:57.000 00:53:58.260 components to build their own web
00:53:58.260 00:53:58.680 browsers.
00:53:59.460 00:54:01.800 Give, like currently everybody's
00:54:01.800 00:54:02.760 using chromium.
00:54:03.780 00:54:05.400 As a foundation to build their own
00:54:05.400 00:54:05.880 web browsers.
00:54:06.480 00:54:08.160 Giving, given that IE is free. What
00:54:08.160 00:54:10.200 is the incentive for Netscape to make
00:54:10.200 00:54:12.240 the browser even cheaper? It's a
00:54:12.240 00:54:13.560 preemptive move they need to prevent
00:54:13.560 00:54:15.900 Microsoft getting a complete Monopoly
00:54:15.900 00:54:17.880 in web browsers, even free web
00:54:17.880 00:54:19.620 browsers because it would,
00:54:19.620 00:54:20.760 theoretically, because that would
00:54:20.760 00:54:22.380 theoretically give Microsoft an
00:54:22.380 00:54:24.000 opportunity to increase the cost of
00:54:24.000 00:54:26.640 web browsing in other ways, Say by
00:54:26.640 00:54:27.840 increasing the price of Windows.
.endquote
.p
00:54:29.940 00:54:31.440 And notice Windows is completely
00:54:31.440 00:54:32.160 free, I think.
00:54:33.180 00:54:35.280 Because it's now the commodity.
00:54:36.900 00:54:39.300 Two other things, which is probably.
.p
00:54:42.660 00:54:44.460 It's not the server. I mean, it's
00:54:44.460 00:54:45.660 also like Azure, but.
00:54:46.920 00:54:49.320 What is Microsoft Windows, the
00:54:49.320 00:54:50.880 commodity tool that they're trying to
00:54:50.880 00:54:53.700 commoditize? I guess it's like a mode
00:54:53.700 00:54:54.720 for people to stay on.
00:54:56.640 00:54:59.700 Because Linux on a desktop which is?
00:55:01.470 00:55:03.570 Yeah, but switch, like, oh, it's
00:55:03.570 00:55:03.930 okay, so?
00:55:05.010 00:55:07.350 Microsoft is also trying to sell
00:55:07.350 00:55:07.890 advertise.
.p
00:55:07.890 00:55:09.750 I mean, they're kind of. They don't
00:55:09.750 00:55:10.650 really know where they're going,
00:55:10.650 00:55:11.790 right? So, they're like trying to
00:55:11.790 00:55:13.710 sell advertising in the start menu.
00:55:13.710 00:55:16.830 They have Bing as their web search
00:55:16.830 00:55:18.630 alternative to Google.
00:55:21.450 00:55:23.130 Um, they have Linux on the desktop
00:55:23.130 00:55:26.310 trying to take their cake and it's
00:55:26.310 00:55:27.210 kind of doing that right now.
.p
00:55:27.210 00:55:29.250 With valve for the gamers and with, like?
00:55:30.450 00:55:32.850 Other people by that are building
00:55:32.850 00:55:35.550 distros for power users, and I mean,
00:55:35.550 00:55:39.090 I and Ubuntu for my grandmother 10
00:55:39.090 00:55:40.830 years ago, and it was perfectly fine,
00:55:40.830 00:55:42.210 and she had enough.
00:55:44.670 00:55:46.470 Capabilities to write her books on
00:55:46.470 00:55:48.930 it, and the printer worked so like?
.p
00:55:52.170 00:55:52.590 There wasn't really.
00:55:54.810 00:55:56.070 There wasn't really a need for
00:55:56.070 00:55:57.930 Windows even like a decade ago.
00:55:59.520 00:56:02.160 Unless you're doing like 3D power CAD
00:56:02.160 00:56:04.740 software stuff for like Final Cut
00:56:04.740 00:56:07.320 Pro, but then you were using Mac OS
00:56:07.320 00:56:07.560 anyway.
.p
00:56:13.500 00:56:16.680 Yeah, so I mean Windows is is a
00:56:16.680 00:56:17.520 commodity at this point and?
00:56:18.720 00:56:20.340 Some and something that's waning, I
00:56:20.340 00:56:20.580 would say.
.p
00:56:24.960 00:56:26.760 But that's from my own point of view,
00:56:26.760 00:56:29.280 who has been using Linux on the
00:56:29.280 00:56:30.900 desktop for 15 years plus.
00:56:32.520 00:56:35.340 Um and I only use Windows for gaming
00:56:35.340 00:56:37.980 and it's fine-ish.
00:56:39.000 00:56:41.100 I don't know. I don't even have it
00:56:41.100 00:56:43.020 logged onto my Microsoft account even
00:56:43.020 00:56:44.400 though they try to make me every
00:56:44.400 00:56:45.240 three days, but
00:56:46.380 00:56:47.880 Yeah, as long as they can click away.
.p
00:56:49.020 00:56:50.820 Please, please finish setting up your
00:56:50.820 00:56:51.960 PC thing.
00:56:53.100 00:56:53.640 It's okay.
.p
00:56:56.940 00:56:57.060 Okay,
.quote spolsky
00:56:57.060 00:56:59.040 my argument is even more shaky
00:56:59.040 00:57:00.300 because it's pretty clear that
00:57:00.300 00:57:02.340 Netscape in the days of Parks stale
00:57:02.340 00:57:03.480 didn't exactly know what it was doing
00:57:03.480 00:57:05.880 a more likely explanation of what
00:57:05.880 00:57:07.680 Netscape did. Is that up of
00:57:07.680 00:57:09.300 management was technologically
00:57:09.300 00:57:11.520 technologically inept, and they had
00:57:11.520 00:57:12.720 no choice but to go along with
00:57:12.720 00:57:14.580 whatever scheme the developers came
00:57:14.580 00:57:14.820 up with.
.p
00:57:14.820 00:57:17.160 The developers were hackers, not
00:57:17.160 00:57:18.900 economists, and only coincidentally
00:57:18.900 00:57:20.520 came up with a scheme which served
00:57:20.520 00:57:21.840 their strategy, but let's give them
00:57:21.840 00:57:23.160 the benefit of the doll, I mean.
00:57:24.180 00:57:25.800 Netscape doesn't exist anymore, so
00:57:25.800 00:57:27.900 they clearly didn't really have a plan.
00:57:30.480 00:57:33.120 Headline transmit a highest linels
00:57:33.120 00:57:35.040 pays them to hacker Linux myth.
.p
00:57:35.040 00:57:36.780 They just did it to get publicity.
00:57:36.780 00:57:39.180 Would you have heard of transmitter
00:57:39.180 00:57:41.520 otherwise, reality transmitter is a
00:57:41.520 00:57:43.740 CPU company, the natural complement
00:57:43.740 00:57:45.360 of a CPU is an operating system.
.endquote
00:57:45.360 00:57:48.060 Transmitter wants os's to be a commodity.
00:57:48.960 00:57:51.720 Headline sudden HP Peg Simeon to
00:57:51.720 00:57:54.240 hakong gnome myth. Sonich peer
00:57:54.240 00:57:55.020 supporting free software because
00:57:55.020 00:57:57.360 they're light bazaars, not cathedrals.
.p
00:57:57.360 00:57:59.460 Also, like they open source star
00:57:59.460 00:58:01.020 office into LibreOffice. I don't know
00:58:01.020 00:58:02.400 how that happened, but I think there
00:58:02.400 00:58:04.680 was also Sun, so guess the same move
00:58:04.680 00:58:06.240 for Microsoft Office.
00:58:07.320 00:58:09.480 Reality. Sun and HPR Hardware
00:58:09.480 00:58:10.800 companies. They make boxing
.quote spolsky
00:58:10.800 00:58:11.340 in order
00:58:11.340 00:58:13.140 to make money on the desktop they
00:58:13.140 00:58:14.700 need for windowing systems, which are
00:58:14.700 00:58:17.280 complement of desktop computers to be
00:58:17.280 00:58:18.000 a commodity.
.p
00:58:19.020 00:58:20.580 Why don't they take the money and
00:58:20.580 00:58:21.900 they're paying Simeon and use it to
00:58:21.900 00:58:23.760 develop a propriety proprietary
00:58:23.760 00:58:26.460 windowing system they tried? This sun
00:58:26.460 00:58:29.220 had news and HP had New Wave, but
00:58:29.220 00:58:30.900 these are really Hardware companies
00:58:30.900 00:58:33.180 at heart with pretty crude software
00:58:33.180 00:58:35.340 skills, and they need windowing
00:58:35.340 00:58:38.100 systems to be a cheap commodity, not
00:58:38.100 00:58:39.660 a proprietary Advantage which they
00:58:39.660 00:58:40.260 have to pay for.
.p
00:58:40.260 00:58:43.260 So, they hide the nice guys as Simeon
00:58:43.260 00:58:45.000 to do this for the same reason that
00:58:45.000 00:58:47.100 sunboard star office oh, here we go
00:58:47.100 00:58:48.600 about star office.
00:58:48.960 00:58:50.640 Open source to commoditize software
00:58:50.640 00:58:52.140 and make more money on Hardware.
.endquote
00:58:53.040 00:58:54.060 Also, this is.
00:58:54.960 00:58:57.420 Kde with Nokia, I guess.
00:58:59.280 00:59:00.180 Who, which?
00:59:02.820 00:59:04.680 Yeah, I think that was their strategy for.
.p
00:59:06.960 00:59:10.260 Mobileish phones KDKD?
00:59:11.040 00:59:12.660 But yeah, KD is probably the same for
00:59:12.660 00:59:14.640 some other for some other company.
00:59:16.740 00:59:18.720 I mean, in Java, right, headlines,
.quote spolsky
00:59:18.720 00:59:20.640 Sun develops. Java new bytecode
00:59:20.640 00:59:22.140 system means write once, run
00:59:22.140 00:59:24.600 anywhere. The bytecode idea is not
00:59:24.600 00:59:25.860 new. Programmers have always tried to
00:59:25.860 00:59:27.840 make the code run on as many machines
00:59:27.840 00:59:28.140 as possible.
.p
00:59:28.140 00:59:30.300 That's how you commoditize your
00:59:30.300 00:59:30.660 complement
.endquote
00:59:30.660 00:59:32.940 for make for years micros,
00:59:32.940 00:59:34.020 but that's also C.
00:59:35.040 00:59:36.420 C is one of the most.
00:59:37.980 00:59:40.800 Commoditizing their complement
00:59:40.800 00:59:42.840 language, which is, it is an abstract.
00:59:44.040 00:59:46.500 I mean, it signifies a machine that
00:59:46.500 00:59:47.400 just can run.
00:59:48.720 00:59:50.220 Somewhat efficiently everywhere, right?
00:59:51.240 00:59:52.560 And because of undefined Behavior,
00:59:52.560 00:59:53.040 you can make it.
.p
00:59:54.300 00:59:55.740 Pretty efficient on most systems,
00:59:55.740 00:59:58.440 even for weird bite size and stuff.
.p
.quote spolsky
01:00:01.440 01:00:03.000 For years, Microsoft had its own P
01:00:03.000 01:00:05.160 code compiler and portable windowing
01:00:05.160 01:00:07.020 layer, which let Excel run on Mac,
01:00:07.020 01:00:09.540 Windows, and OS 2, and on Motorola,
01:00:09.540 01:00:11.820 Intel, Alpha mips, and power PC chips.
01:00:12.600 01:00:14.460 Quark has a layer which runs
01:00:14.460 01:00:15.600 Macintosh codes on Windows.
01:00:16.560 01:00:18.000 The C programming language is best
01:00:18.000 01:00:19.380 described as a hardware independent
01:00:19.380 01:00:20.220 assembler language.
.p
01:00:20.220 01:00:22.440 Oh, here we go. It's not a new idea
.endquote
01:00:22.440 01:00:23.760 in too software developers.
01:00:24.780 01:00:27.540 Also wine on Linux and.
01:00:28.920 01:00:32.640 No WSL Windows. What's called the
01:00:32.640 01:00:34.740 thing that makes Linux apps run on
01:00:34.740 01:00:37.080 Windows and also the whole Unix tool
01:00:37.080 01:00:37.860 chain run on Windows?
.quote spolsky
01:00:38.760 01:00:40.680 If you can run your software anyway,
01:00:40.680 01:00:42.900 anywhere that makes Hardware more of
01:00:42.900 01:00:43.380 a commodity.
.p
01:00:43.380 01:00:45.600 As Hardware prices go down, the
01:00:45.600 01:00:47.460 market expands, driving more demand
01:00:47.460 01:00:49.260 for software and leaving customers
01:00:49.260 01:00:50.700 with extra money to spend on
01:00:50.700 01:00:52.140 software, which can now be more expensive.
01:00:53.640 01:00:57.780 Sons enthusiasm for wara?
01:00:58.620 01:01:00.660 Whatever that was, is, I'm strange
01:01:00.660 01:01:02.100 because sun is a hardware company
01:01:02.100 01:01:04.560 making Hardware commodity is the last
01:01:04.560 01:01:07.260 thing they want to do, oops.
.p
01:01:08.880 01:01:10.320 Sun is the loose Cannon of the
01:01:10.320 01:01:12.720 computer industry, unable to see past
01:01:12.720 01:01:14.460 their raging fear and loathing of
01:01:14.460 01:01:16.500 Microsoft. They adopt strategies
01:01:16.500 01:01:18.780 based on anger rather than self-interest.
01:01:19.740 01:01:22.680 Son's two strategies are a make
01:01:22.680 01:01:24.240 software commodity by promoting and
01:01:24.240 01:01:27.000 developing free software. Star offers
01:01:27.000 01:01:29.880 Linux Apache ignom or Apache was
01:01:29.880 01:01:30.240 something.
.p
01:01:32.280 01:01:34.680 Np make Hardware commodity by
01:01:34.680 01:01:37.500 promoting Java with with its bytecode
01:01:37.500 01:01:38.580 architecture and vora.
01:01:39.300 01:01:41.520 Okay, so on Pop Quiz. When the Music
01:01:41.520 01:01:42.900 Stops where you're going to sit down
01:01:42.900 01:01:46.020 without proprietary advantages in
01:01:46.020 01:01:47.041 Hardware or software, you're going to
01:01:47.041 01:01:49.500 have to take commodity price. The
01:01:49.500 01:01:51.900 commodity price, which barely covers
01:01:51.900 01:01:54.180 the cost of cheap factories in
01:01:54.180 01:01:56.940 quarter whatever some Indian Town.
.p
01:01:58.020 01:02:00.480 Offices in Silicon Valley.
01:02:03.150 01:02:05.310 But Joel Jared's son is trying to
01:02:05.310 01:02:06.570 commoditize the operating system,
01:02:06.570 01:02:08.550 like transmitter, not the hardware,
01:02:08.550 01:02:10.710 maybe. But the fact that Java
01:02:10.710 01:02:12.330 bytecode also commoditizes. The
01:02:12.330 01:02:13.770 hardware is some pretty significant
01:02:13.770 01:02:15.630 collateral damage to sustain,
.endquote
01:02:15.630 01:02:16.230 which I
01:02:16.230 01:02:18.090 guess explains why Sun went down and
01:02:18.090 01:02:20.610 like got sold to Oracle for scraps
01:02:20.610 01:02:21.750 and scrap metal and.
.p
01:02:22.890 01:02:24.510 Whatever they did to Java after that.
01:02:25.530 01:02:27.090 Um, Sunset, it, I guess.
01:02:27.870 01:02:28.290 Um.
.p
01:02:33.180 01:02:35.520 I had a thought about this a few
01:02:35.520 01:02:35.820 seconds ago.
01:02:37.140 01:02:40.080 My Hardware commodity, oh yeah, AI.
01:02:40.080 01:02:43.140 So, lnms lms don't have a mode.
01:02:44.040 01:02:45.960 Like, I don't understand, like?
01:02:47.520 01:02:48.960 It's pretty obvious to anybody who
01:02:48.960 01:02:50.640 knows how these things are not work
01:02:50.640 01:02:52.140 under the hoodest. You just need a
01:02:52.140 01:02:54.240 lot of compute power, and that
01:02:54.240 01:02:55.260 basically, and a few people that you
01:02:55.260 01:02:58.440 can buy for like a million a year.
.p
01:02:59.760 01:03:02.100 Just by paying more than your than
01:03:02.100 01:03:03.180 the other companies. So, you just
01:03:03.180 01:03:04.620 need a lot of investment money, and
01:03:04.620 01:03:06.780 you can spin up a company, and also
01:03:06.780 01:03:08.700 there's a lot of very smart people in
01:03:08.700 01:03:10.620 China that will do it for the stage.
01:03:11.820 01:03:14.160 So you get deepsea, you get Kimmy,
01:03:14.160 01:03:16.140 you get whatever Alibaba Cloud?
.p
01:03:17.640 01:03:19.560 Um, you get rock.
.p
01:03:22.800 01:03:24.720 So there's like zero modes.
01:03:26.760 01:03:28.740 Basically, anybody that can type text
01:03:28.740 01:03:32.340 to to some, like, it's a pipe text
01:03:32.340 01:03:36.420 into curl remote server returns Json
01:03:36.420 01:03:39.600 Affair. And it's like, very simple to
01:03:39.600 01:03:41.280 implement technically, and that's why
01:03:41.280 01:03:42.540 everybody is doing it right now.
01:03:45.000 01:03:45.300 Also, you can.
01:03:46.440 01:03:48.780 Quantize these models to fit on small
01:03:48.780 01:03:51.120 commodity Hardware already.
.p
01:03:51.120 01:03:53.100 Like, you can run some models on.
01:03:54.510 01:03:58.110 Um, Mac minis or on your MacBook or
01:03:58.110 01:04:03.450 your newfangled arm or AMD chips
01:04:03.450 01:04:05.310 system on on a chip.
01:04:06.900 01:04:08.520 So, yeah, so these things are like
01:04:08.520 01:04:10.740 they have zero modes, so they're
01:04:10.740 01:04:12.360 they're already Commodities, and they
01:04:12.360 01:04:13.980 already have been Commodities for,
01:04:13.980 01:04:14.640 like, at least two years.
.p
01:04:15.960 01:04:18.300 So, this whole hype is basically.
01:04:19.740 01:04:21.120 Either. They're gonna figure out how
01:04:21.120 01:04:21.720 to make.
01:04:23.370 01:04:26.070 The Trump Administration create laws
01:04:26.070 01:04:28.770 for them to have artificial, like how
01:04:28.770 01:04:31.050 software patterns work or how?
01:04:33.060 01:04:35.280 The law system works is basically
01:04:35.280 01:04:37.260 create artificial modes that are not.
01:04:38.250 01:04:40.770 Inherently there by abusing the
01:04:40.770 01:04:43.350 society societal system.
.p
01:04:44.670 01:04:46.350 Um, to make things.
01:04:47.910 01:04:50.550 Economically viable, which is
01:04:50.550 01:04:51.990 probably what entropic is trying to
01:04:51.990 01:04:56.190 do or GPT, or like open AIS trying to do.
01:04:57.630 01:04:59.910 But then they won't have success
01:04:59.910 01:05:01.770 because the Chinese are already
01:05:01.770 01:05:02.490 releasing these things.
.p
01:05:06.450 01:05:10.170 Three basically to download, so this
01:05:10.170 01:05:11.610 this is not going anywhere, and I
01:05:11.610 01:05:11.790 feel like.
01:05:12.930 01:05:14.490 I have no idea what this is, like the
01:05:14.490 01:05:16.530 only people who are winning are the
01:05:16.530 01:05:18.150 hardware producers basically and even
01:05:18.150 01:05:20.970 that. Like, because of llms.
01:05:21.870 01:05:24.630 Like the the advantage that Nvidia
01:05:24.630 01:05:26.610 had with their proprietary whatever
01:05:26.610 01:05:27.690 they it was called?
.p
01:05:29.070 01:05:30.870 Their libraries stack basically.
01:05:30.870 01:05:32.550 Isn't there anymore? They already
01:05:32.550 01:05:34.650 retreated on the Linux front by
01:05:34.650 01:05:37.350 saying, oh yeah, we're gonna open
01:05:37.350 01:05:40.050 source our drivers and maybe put some
01:05:40.050 01:05:41.850 most of it on the graphics cards, but?
01:05:43.800 01:05:45.180 They're already retreating.
01:05:46.680 01:05:47.640 Um, on that front, so?
01:05:48.900 01:05:50.520 I feel like graphics card are gonna
01:05:50.520 01:05:52.260 be com commoditized.
.p
01:05:52.260 01:05:54.000 There's people building Asic.
01:05:54.840 01:05:56.520 Um, chips that you can just plug into
01:05:56.520 01:05:58.980 your PCS that have like a burnt in
01:05:58.980 01:05:59.880 model on them, so.
01:06:01.740 01:06:03.960 Plug a piece of Hardware into your
01:06:03.960 01:06:05.580 PC, like with the old graphics cards
01:06:05.580 01:06:07.380 like, essentially, like.
01:06:08.850 01:06:10.710 Like a Bitcoin mining rig, right? You
01:06:10.710 01:06:11.850 plug this in, and then you have a
01:06:11.850 01:06:14.070 model that you can talk to at, like
01:06:14.070 01:06:16.890 10, 100 times the speed of emulating
01:06:16.890 01:06:17.970 on graphic cards.
.p
01:06:19.560 01:06:20.340 So there's like?
01:06:21.000 01:06:22.500 Now, these things are already
01:06:22.500 01:06:23.940 Commodities. You can't tell me any
01:06:23.940 01:06:25.140 other on otherwise.
01:06:27.240 01:06:29.760 So, I'm not sure how the economics
01:06:29.760 01:06:30.420 are going to play out there.
.p
.quote spolsky
01:06:35.280 01:06:36.840 An important thing you notice from
01:06:36.840 01:06:38.160 all these examples is that it's easy
01:06:38.160 01:06:40.260 for software to commoditize Hardware.
01:06:41.100 01:06:42.480 Just for a little hardware
01:06:42.480 01:06:44.460 abstraction layer like Windows NT,
01:06:44.460 01:06:46.740 hell, and T Health like Windows and
01:06:46.740 01:06:49.080 T's hell, which is a tiny piece of
01:06:49.080 01:06:51.000 code, but it's incredibly hard for
01:06:51.000 01:06:52.980 Hardware to commoditize software.
.p
01:06:52.980 01:06:54.660 So, this is the same thing as
01:06:54.660 01:06:57.180 commoditizing art, right? Like, a
01:06:57.180 01:06:59.100 piece of art is a piece of art is a
01:06:59.100 01:07:00.360 piece of piece of art, and you can't
01:07:00.360 01:07:02.460 replace a Monet with some cheap
01:07:02.460 01:07:04.860 knockoff AI generated money. It's
01:07:04.860 01:07:07.020 not, it's not the same. It's like a
01:07:07.020 01:07:08.640 single instance of this thing, and
01:07:08.640 01:07:10.200 maybe you can make a print of the.
.p
01:07:11.040 01:07:11.160 Thing, but
01:07:12.330 01:07:12.510 Um.
01:07:13.770 01:07:13.890 Yeah.
01:07:14.730 01:07:16.830 Software is not interchangeable, as
01:07:16.830 01:07:18.870 staff is the star office marketing
01:07:18.870 01:07:19.530 team is learning.
01:07:20.490 01:07:22.770 Even when the price is zero, the cost
01:07:22.770 01:07:24.390 of switching from Microsoft Office is
01:07:24.390 01:07:26.610 non-zero until the switching cost
01:07:26.610 01:07:26.970 becomes
.endquote
01:07:26.970 01:07:28.050 you. And that's the thing
01:07:28.050 01:07:29.790 with llm stack, the switching cost is
01:07:29.790 01:07:30.210 zero.
.p
01:07:30.210 01:07:32.670 Like, that's why there's open code
01:07:32.670 01:07:34.410 has integration of like a hundred
01:07:34.410 01:07:35.670 different providers because they are
01:07:35.670 01:07:39.150 already like the switching cost is or
01:07:39.150 01:07:41.190 has always been zero for. These
01:07:41.190 01:07:42.630 things is token and token out.
01:07:42.630 01:07:44.250 Doesn't matter whether it's like a
01:07:44.250 01:07:46.890 PDF or a PNG like, this is all
01:07:46.890 01:07:48.390 completely replaceable.
.p
01:07:49.650 01:07:51.090 And maybe you get like a little
01:07:51.090 01:07:53.670 degraded results, but nothing
01:07:53.670 01:07:56.250 fundamentally unfixable.
.p
.quote spolsky
01:08:00.120 01:08:01.260 Until the switching cost becomes
01:08:01.260 01:08:03.180 zero. Desktop office software is not
01:08:03.180 01:08:04.020 truly a commodity.
01:08:04.920 01:08:06.480 And even the smallest differences can
01:08:06.480 01:08:08.700 make two softer packages, a pain to
01:08:08.700 01:08:09.960 switch between, not true anymore, not
01:08:09.960 01:08:11.340 true anymore. Despite the fact that
01:08:11.340 01:08:13.200 Mozilla has all the features I want.
01:08:13.200 01:08:15.420 I'd love to use it if it only avoid
01:08:15.420 01:08:17.160 if only to avoid the whack-a-mole
01:08:17.160 01:08:18.060 pop-up ad game.
.p
01:08:18.960 01:08:20.880 I'm too used to heading L to D to go
01:08:20.880 01:08:23.520 to the address bar, so sue me one
01:08:23.520 01:08:24.780 tiny difference, and you lose your
01:08:24.780 01:08:27.000 commodity status, but I've pulled
01:08:27.000 01:08:29.040 hard drives out of IBM computers and
01:08:29.040 01:08:30.660 slammed them into Dell computers and
01:08:30.660 01:08:31.980 boom. The system comes up perfectly
01:08:31.980 01:08:32.880 and
.endquote
01:08:32.880 01:08:35.100 run and run the face as if it's
01:08:35.100 01:08:36.600 still the old computer.
.p
01:08:36.600 01:08:39.360 Hard drives are f****** magic, like
01:08:39.360 01:08:40.440 spinning discs.
01:08:41.460 01:08:42.420 The package you get?
01:08:43.380 01:08:46.380 Was like 12 spinning discs and read
01:08:46.380 01:08:47.100 heads in them.
01:08:48.120 01:08:50.340 Is complete like that. These things
01:08:50.340 01:08:52.920 even manage to exist in the physical
01:08:52.920 01:08:54.780 world is a complete.
01:08:55.860 01:08:58.020 It blows my brain, right? I have like
01:08:58.020 01:08:59.940 a 12 terabyte drive right in front of
01:08:59.940 01:09:01.560 me, which I haven't managed to put
01:09:01.560 01:09:01.801 into a.
.p
01:09:03.300 01:09:04.320 A case yet.
01:09:05.040 01:09:07.320 And like that, you can store this much.
01:09:08.340 01:09:10.260 Data on like this little physical
01:09:10.260 01:09:11.040 space is.
01:09:12.180 01:09:14.760 Incredible and the amount of?
01:09:15.840 01:09:19.200 Capitalist, you know, like the the
01:09:19.200 01:09:19.500 amount of?
.p
01:09:22.710 01:09:24.570 Orders of magnitudes of improvements
01:09:24.570 01:09:26.550 over this short of amount of time,
01:09:26.550 01:09:28.530 like in the in the like. When this
01:09:28.530 01:09:29.790 article was written, the biggest
01:09:29.790 01:09:30.870 drives were like.
01:09:31.650 01:09:34.110 A gigabyte, probably. And now it's 12
01:09:34.110 01:09:34.470 terabytes.
01:09:36.150 01:09:38.010 It is absolutely baffling.
01:09:39.990 01:09:41.790 Um, and even like the smallest.
01:09:43.050 01:09:45.750 The smallest USB drives are like?
.p
01:09:46.890 01:09:48.210 100 gigabytes no?
01:09:50.010 01:09:50.490 Which is crazy?
.p
01:09:55.440 01:09:56.640 Yeah, I've pulled hard drives out of
01:09:56.640 01:09:58.200 IBM computers and slam them into the
01:09:58.200 01:09:59.580 their computers and boom. The system
01:09:59.580 01:10:01.800 comes up perfectly and runs. If it is
01:10:01.800 01:10:03.480 as if it was still in your computer,
01:10:03.480 01:10:05.700 but the thing is, even though these
01:10:05.700 01:10:07.560 increases have been happening, the
01:10:07.560 01:10:09.000 interfaces have always been the same,
01:10:09.000 01:10:11.100 which means these things are
01:10:11.100 01:10:12.180 commodity Hardware.
.p
01:10:12.180 01:10:14.160 Like, even though only like three
01:10:14.160 01:10:15.600 providers in the world can build
01:10:15.600 01:10:16.020 these things.
01:10:18.030 01:10:19.230 There's the commodity Hardware
01:10:19.230 01:10:20.190 because there could be another.
01:10:21.900 01:10:23.460 Another provider that builds them,
01:10:23.460 01:10:24.780 maybe a bit little bit worse, maybe
01:10:24.780 01:10:27.120 only like a gigabyte instead of like,
01:10:27.120 01:10:29.400 maybe only like a terabyte instead of
01:10:29.400 01:10:32.040 10 terabytes, but it's still a commodity.
.p
01:10:33.000 01:10:35.520 Um, a real black box, right? A
01:10:35.520 01:10:36.120 replaceable thing?
.p
01:10:39.330 01:10:41.910 Um, Amos Mickelson, the CEO of Creo.
01:10:41.910 01:10:44.490 What, what kind of company is? Korea
.quote spolsky
01:10:44.490 01:10:46.410 told me that every employee in his
01:10:46.410 01:10:47.790 firm is required to take a course in
01:10:47.790 01:10:49.650 what he calls economic thinking.
01:10:49.650 01:10:52.770 Great idea. Even simple Concepts in
01:10:52.770 01:10:54.870 basic microeconomics go a long way to
01:10:54.870 01:10:56.310 understanding some of the fundamental
01:10:56.310 01:10:57.450 shifts going on today.
.endquote
.p
01:10:58.710 01:10:59.250 So, yeah, I mean.
01:11:01.680 01:11:02.700 A lot has been happening, and
01:11:02.700 01:11:04.020 everything's kind of still the same
01:11:04.020 01:11:06.180 like the the principles are still there.
01:11:07.020 01:11:08.400 Uh, once you know how to look for
01:11:08.400 01:11:09.420 them, you're going to see them
01:11:09.420 01:11:11.040 everywhere. I feel like, which is why
01:11:11.040 01:11:11.940 I want to talk about this.
01:11:13.200 01:11:13.380 Yes.
.p
01:11:14.640 01:11:14.760 Um.
01:11:15.780 01:11:16.320 Thanks for listening.
|