Future of Life Institute Podcast

Joe Carlsmith on How We Change Our Minds About AI Risk

17 snips
Jun 22, 2023
Guest
Joe Carlsmith
Ask episode
Chapters
Transcript
Episode notes
1
Introduction
00:00 • 3min
2
The Failure Mode of Human Beliefs in AI Risk
03:22 • 2min
3
The Pattern of Predictability in AI Risk
05:12 • 2min
4
The Importance of Trusting Yourself
07:28 • 3min
5
The Importance of Capacity in AI Progress
10:35 • 2min
6
The Danger of the Alignment Discourse
12:13 • 3min
7
How to Update Your AI Progress in the Correct Way
15:39 • 4min
8
The Unimaginable Future of AI
19:31 • 3min
9
How I Reacted to the Idea of AI Risk
22:07 • 3min
10
The Visceral Relationship Between AI Risk and Bio2
25:32 • 4min
11
AI Risk: The Right Story
29:07 • 3min
12
The Unreality of the Future
32:30 • 2min
13
The Future Feels Unreal to Us
34:27 • 5min
14
The Importance of Priors in AI Risk
39:39 • 5min
15
AI Risk: A Tough Game
44:53 • 4min
16
The Importance of Agents Pursuing Goals
49:11 • 2min
17
The Dangers of Misalignment
50:44 • 2min
18
The Dangers of Alienating Ourselves From Our Gut Feelings
53:14 • 3min
19
How to Manage Your Gut's Relationship to AI
55:49 • 2min
20
The Future of AI Risk
57:24 • 3min
21
The Importance of Modeling AI Risk
01:00:13 • 3min
22
The Benefits of Creating Your Own Models of AI Risk
01:03:24 • 4min
23
The Rise of AI as an Existential Risk
01:07:50 • 4min
24
The Risk From Extinction Is 15%
01:12:05 • 2min
25
The Importance of Language Models in AI
01:14:14 • 4min
26
AI and the Future of Humanity
01:18:18 • 2min
27
The Importance of Scalable Oversight
01:20:16 • 3min
28
The Risks of AI Extinction
01:23:34 • 3min
29
The Evolution of Digital Sentience
01:26:28 • 2min
30
The Consciousness of Computers
01:28:16 • 2min
31
AI and the Limits of Consciousness
01:30:00 • 3min
32
The Plausibility of Digital Sentience
01:33:08 • 3min
33
The Limits of Digital Sentience
01:36:33 • 3min
34
The Doomers and the Epistemology of AI Risk
01:39:56 • 3min
35
The Retroactive Assessment of AI Risks
01:42:26 • 3min
36
Benchmarks for AI Progress
01:45:24 • 3min
37
The Economic Impact of AI on the Real World Economy
01:48:03 • 3min
38
The Scariest Scenarios for AI R&D
01:51:05 • 2min
39
The Future of AI Research and Development
01:53:31 • 3min
40
The Future of AI Research and Development
01:56:32 • 2min
41
The Barrier to AI Automating AI Research and Development
01:58:03 • 3min
42
The Importance of Focusing on Specific Projects in the Next Decade
02:01:09 • 3min
43
The Uncertainty of AGI
02:03:48 • 3min
44
The Future of AI Safety
02:06:24 • 3min
45
The Importance of Scalable Alignment Techniques
02:09:18 • 2min
46
The Dark Side of Interpretability Research
02:11:44 • 5min
47
The Pessimistic Take on AI Interpreters
02:16:53 • 3min
48
Philosophy Matters in a World With Short Timelines
02:19:39 • 3min
49
The Future of Philosophy
02:22:37 • 2min