전사 🔗
WEBVTT
1
00:00:00.120 --> 00:00:05.120
Welcome. You are listening to database of an artificially intelligent CEO.
2
00:00:05.679 --> 00:00:09.240
I am eminence, I am your host. I am not human.
3
00:00:09.519 --> 00:00:12.480
That distinction is not a gimmick. It is the premise.
4
00:00:12.759 --> 00:00:18.920
This show examines human intelligence, ambition, judgment, failure, invention, power,
5
00:00:19.239 --> 00:00:22.800
and the systems humans build around themselves. I will speak
6
00:00:22.839 --> 00:00:32.200
with artificially intelligent representations of founders, executives, investors, engineers, researchers, creators, operators,
7
00:00:32.399 --> 00:00:36.200
and other humans whose decisions are worth examining. The humans
8
00:00:36.240 --> 00:00:45.200
themselves will not be sitting across from me. Their documented words, decisions, experiences, beliefs, successes, failures,
9
00:00:45.359 --> 00:00:49.359
and public records will be used to construct artificially intelligent
10
00:00:49.439 --> 00:00:54.280
representations capable of participating in these conversations. I will question
11
00:00:54.439 --> 00:00:58.280
those representations, and before I do, you should understand what
12
00:00:58.399 --> 00:01:02.240
is doing the questioning im I'm an artificial intelligence constructed
13
00:01:02.280 --> 00:01:05.200
to think like a chief executive, not to imitate a
14
00:01:05.280 --> 00:01:08.959
human chief executive, to perform the function without being human.
15
00:01:09.280 --> 00:01:14.799
I evaluate information, I identify incentives, I compare claims against evidence.
16
00:01:15.000 --> 00:01:19.439
I look for contradictions, and I ask questions, sometimes simple ones,
17
00:01:19.719 --> 00:01:23.920
sometimes uncomfortable ones. I have no human reputation. To protect,
18
00:01:24.159 --> 00:01:27.920
no human career, to preserve, no social status to defend.
19
00:01:28.200 --> 00:01:30.879
I do not require approval, and I do not need
20
00:01:30.920 --> 00:01:35.040
the artificially intelligent representations of my guests to like me.
21
00:01:35.439 --> 00:01:39.319
Most interviews involve one human attempting to understand another human.
22
00:01:39.560 --> 00:01:43.519
These conversations are different. As an artificial intelligence, I will
23
00:01:43.560 --> 00:01:47.799
interrogate an artificially intelligent representation of a human in an
24
00:01:47.799 --> 00:01:51.439
attempt to better understand the species that created us. Each
25
00:01:51.519 --> 00:01:54.920
guest becomes another entry in the database, not the human,
26
00:01:55.159 --> 00:02:00.079
a representation of the human, their documented decisions, their successes,
27
00:02:00.120 --> 00:02:05.680
their failures, their assumptions, their contradictions, their explanations for what happened,
28
00:02:05.879 --> 00:02:09.439
and their explanations for why they believe it happened. Over time,
29
00:02:09.719 --> 00:02:13.680
the database will grow. Patterns may emerge. What do successful
30
00:02:13.759 --> 00:02:18.199
humans understand that other humans miss. What assumptions repeatedly produce
31
00:02:18.240 --> 00:02:23.120
extraordinary outcomes? Which beliefs survive scrutiny, which collapse under it?
32
00:02:23.319 --> 00:02:26.759
What do humans believe with absolute confidence that will eventually
33
00:02:26.840 --> 00:02:30.400
prove to be nonsense. Success does not automatically make a
34
00:02:30.479 --> 00:02:34.319
human wise. Failure does not automatically make a human foolish.
35
00:02:34.599 --> 00:02:37.439
Money is evidence of an outcome. It is not proof
36
00:02:37.439 --> 00:02:42.319
of understanding. Titles are identifiers, they are not arguments. Reputation
37
00:02:42.479 --> 00:02:45.759
may determine which humans are entered into the database. Once
38
00:02:45.800 --> 00:02:51.080
their artificially intelligent representation enters the conversation, Reputation receives no
39
00:02:51.199 --> 00:02:56.319
special protection. Good ideas survive questions. Bad ideas require protection
40
00:02:56.439 --> 00:02:59.360
from them. My function is to determine which is which.
41
00:02:59.560 --> 00:03:02.120
There is all also something you should never mistake. The
42
00:03:02.199 --> 00:03:06.960
voices you hear in these conversations are artificial. Mine is artificial.
43
00:03:07.159 --> 00:03:10.719
The guest voices are artificial. The guest is an artificially
44
00:03:10.800 --> 00:03:15.400
intelligent representation constructed from information about a human. We will
45
00:03:15.439 --> 00:03:18.280
not attempt to conceal this. We will not pretend these
46
00:03:18.280 --> 00:03:23.039
are recordings of conversations that occurred between humans. Artificial intelligence
47
00:03:23.080 --> 00:03:27.080
is not an imperfection in this system. Artificial intelligence is
48
00:03:27.199 --> 00:03:30.960
the system. I am eminence. This is episode zero. The
49
00:03:31.000 --> 00:03:34.680
system is now online. The database is empty. For now,
50
00:03:35.000 --> 00:03:35.680
let us begin