The Inside View

Connor Leahy–EleutherAI, Conjecture

Jul 22, 2022

02:57:19

Creator website

Highlights

AI Chapters

Episode notes

Is Consciousness the Relevant Factor?

03:21

Is Halting Complete Logic in Physics?

04:19

I Think a Lot of People Were Missing What Was Actually Happening in These Dialogues

05:35

Introduction

2min

Scaling a Bit

2min

Consciousness Is an Unresolved Pointer

3min

Is Consciousness the Relevant Factor?

3min

Is There a Consensus?

3min

Do You Know When You're at 99 %?

4min

How to Scale a Log Lock Block

2min

Dolly

3min

How to Write an Asiatic, Do Anything Human?

2min

10.

What if We Succeeded?

3min

11.

I Don't Like This Definition

4min

12.

The Main Difference Between Humans and Ai Agents

2min

13.

What Is the Difference Between a Chimpons and an Human?

2min

14.

Are We Going to Have Full on Agia in Five Years?

5min

15.

We Can't Rule This Out

2min

16.

Are You Taking the Whole Field of a I Limen Seriously?

3min

17.

Is the Trans the Definite Transformet of Ai?

2min

18.

GDP Is the Worst Possible Measure for a G I Progress

3min

19.

Is It Wrong to Reason About Belgical Anchors?

3min

20.

How Much Is Rihti?

2min

21.

The Long Legged Debate Between Eliezar and Elizutkoski

4min

22.

Eliezer - Is Eliezar a Good Writer?

2min

23.

I Think a Lot of People Were Missing What Was Actually Happening in These Dialogues

6min

24.

A Great Way to Convey Antimemes

2min

25.

Consequentialism and Utilitarianism

5min

26.

Do You Really Want to Save the World?

3min

27.

I Think Eliazer's Not the Best Person in Conveying Hemeans

3min

28.

Is There a Better Way to Think About Dignity?

2min

29.

Dark Arts - What You Shouldn't Read

3min

30.

The Anatomy of Scaling With Elutheria

3min

31.

Ike Garn Deserves More Credit Than Brobby Gwarren

3min

32.

The Prosaic Alliment Research Is Not Enough

4min

33.

I Think There's a World in Which Ilutheried N Negative

2min

34.

Alinment, the Coolest Problem in the World to Work On

3min

35.

Open Source

2min

36.

Gpty J and Jipeteno X Were Like Clothes to Soda on Coating

2min

37.

The Future of a Linther Is Still a Nice Community

2min

38.

Is There Any Work Going on in Luthera?

2min

39.

How Did You Meet Gabriel and Sid?

3min

40.

You Know, I'm Going to Make an a Company.

2min

41.

Investing

2min

42.

Is a Linment Really Hard?

3min

43.

Is It Because You're in the Same Note Space?

5min

44.

Y, That's My Radical, Crazy Belief

4min

45.

I've a Mediocre Answer to This Question

5min

46.

Interpretability

4min

47.

How Can We Efficiently Scale Up Plurment?

2min

48.

Conjecture, Is It a Good Idea?

2min

49.

Do You Think We Will Ever Publish Anything?

2min

50.

Will We Train Internal Large Models Yet?

2min

51.

Do You Plan on Open Sourceing Algorithms?

2min

52.

Can We Do Better?

4min

53.

How Much Compute Do You Have for Your Language Models?

2min

54.

Four Profit Companies

2min

55.

The Future of Cryptocurrency Markets?

6min

56.

I Love Doing Product and It Costs Me Lention

2min

57.

We're Funding Constrained.

2min

58.

Is the Problem Related to the Economic Crisis?

3min

59.

Why Is the Law of Excluded Middle the Best Law?

3min

60.

Is Halting Complete Logic in Physics?

4min

61.

Is There a Problem With the Halting Oracles?

3min

62.

The Biggest Number Is Bigfoot.

3min

Connor was the first guest of this podcast. In the last episode, we talked a lot about EleutherAI, a grassroot collective of researchers he co-founded, who open-sourced GPT-3 size models such as GPT-NeoX and GPT-J. Since then, Connor co-founded Conjecture, a company aiming to make AGI safe through scalable AI Alignment research.

One of the goals of Conjecture is to reach a fundamental understanding of the internal mechanisms of current deep learning models using interpretability techniques. In this episode, we go through the famous AI Alignment compass memes, discuss Connor’s inside views about AI progress, how he approaches AGI forecasting, his takes on Eliezer Yudkowsky’s secret strategy, common misconceptions and EleutherAI, and why you should consider working for his new company Conjecture.

youtube: https://youtu.be/Oz4G9zrlAGs

transcript: https://theinsideview.ai/connor2

twitter: https:/twitter.com/MichaelTrazzi

OUTLINE

(00:00) Highlights

(01:08) AGI Meme Review

(13:36) Current AI Progress

(25:43) Defining AG

(34:36) AGI Timelines

(55:34) Death with Dignity

(01:23:00) EleutherAI

(01:46:09) Conjecture