

Connor Leahy on AI Safety and Why the World is Fragile
21 snips Jan 26, 2023
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Introduction
00:00 • 2min
How Could AI Go Wrong?
01:40 • 3min
The Minimum Viable Catastrophe
04:12 • 6min
Is It Just Because People Are Too Nice or Too Good?
09:59 • 2min
Why AIs Would Be Different Than Humans?
12:04 • 3min
The Existence Proof That the World Is Unstable
15:20 • 4min
Encoding Human Preferences and the Ontology of Weight Shifts
18:59 • 2min
Open AI
20:36 • 2min
The Fragment of the Fragile World
22:47 • 4min
Don't Be Stupid
26:49 • 2min
What Would We Do if We Were Not Stupid?
29:12 • 4min
Is There a Future for AI?
33:25 • 4min
How to Help Governments Not Be Stupid
37:16 • 4min
Is DARPA a Smart World?
41:10 • 2min
How Could Government Intervention Go Wrong?
43:14 • 2min
I Safety Funding for AI Alignment Research
45:13 • 6min
The Best Funding Model for a Smart Society
51:03 • 2min
Impact Markets
52:43 • 3min
Is There a for Profit Model?
55:28 • 5min
What Do You Think About Win Four Clauses?
01:00:42 • 2min
What Do Signals Matter?
01:02:59 • 2min