108. Last Week In AI 2021: The (full) year in review

Published: Jan. 5, 2022, 3:21 p.m.

2021 has been a wild ride in many ways, but its wildest features might actually be AI-related. We\u2019ve seen major advances in everything from language modeling to multi-modal learning, open-ended learning and even AI alignment.

\n

So, we thought, what better way to take stock of the big AI-related milestones we\u2019ve reached in 2021 than a cross-over episode with our friends over at the Last Week In AI podcast.

\n

***

\n

Intro music:

\n

- Artist: Ron Gelinas

\n

- Track Title: Daybreak Chill Blend (original mix)

\n

- Link to Track: https://youtu.be/d8Y2sKIgFWc

\n

***

\n

Chapters:

\n
    \n
  • 0:00 Intro
  • \n
  • 2:15 Rise of multi-modal models
  • \n
  • 7:40 Growth of hardware and compute
  • \n
  • 13:20 Reinforcement learning
  • \n
  • 20:45 Open-ended learning
  • \n
  • 26:15 Power seeking paper
  • \n
  • 32:30 Safety and assumptions
  • \n
  • 35:20 Intrinsic vs. extrinsic motivation
  • \n
  • 42:00 Mapping natural language
  • \n
  • 46:20 Timnit Gebru\u2019s research institute
  • \n
  • 49:20 Wrap-up
  • \n