The Data Exchange with Ben Lorica

Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

Mar 16, 2023
Ask episode
Chapters
Transcript
Episode notes
1
Introduction
00:00 • 2min
2
The Scaling Loss of a Model
02:14 • 2min
3
The Key Findings for NLP
04:38 • 3min
4
The Future of Computer Vision
07:43 • 1min
5
The Reaction to Your Paper From Machine Learning Researchers
09:12 • 2min
6
Scaling Loss and the Importance of Data
10:57 • 3min
7
The State of Scaling Loss in Machine Learning
13:58 • 3min
8
Do You Track Friends in Memory?
16:33 • 2min
9
The Trends in Model Size
18:12 • 2min
10
The Role of Synthetic Data in Artificial Intelligence
20:23 • 2min
11
The Future of Retrieval Mechanisms
22:31 • 2min
12
The Unpredictabilities of Transfer Learning
24:55 • 2min
13
The Rise of Generative AI
27:12 • 2min
14
The Future of Open Source Data
29:31 • 4min