#452 #452 – Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity
The Future of AI with Dario Amodei, Amanda Askell, and Chris Olah
The Future of AI with Dario Amodei, Amanda Askell, and Chris Olah
Dr. Lex Fridman Host
Curious computer scientist and insightful podcast host exploring diverse ideas.
Dario Amodei
Summary
In this conversation, Dario Amodei, CEO of Anthropic, discusses the scaling laws and the rapid advancements in AI capabilities. He explains how models like Claude are improving and the potential for superintelligent AI by 2026 or 2027. He also addresses concerns about AI safety, misuse, and the ethical implications of powerful AI systems. The conversation delves into the challenges of post-training, RLHF, and the role of regulation in ensuring AI safety. Amanda Askell, a researcher at Anthropic, talks about crafting Claude's character and personality, emphasizing the importance of alignment and user interaction. Chris Olah, a pioneer in mechanistic interpretability, explains the field's goals of understanding the inner workings of neural networks, highlighting the beauty and complexity of AI systems.
Highlights
- If you extrapolate the curves that we’ve had so far, right? If you say, Well, I don’t know, we’re starting to get to PhD level, and last year we were at undergraduate level, and the year before we were at the level of a high school student, again, you can quibble with what tasks and for what.
- I am optimistic about meaning. I worry about economics and the concentration of power. That’s actually what I worry about more, the abuse of power.
- The thing that makes me feel that progress will in the end happen moderately fast, not incredibly fast, but moderately fast, is that you talk to... What I find is I find over and over again, again in large companies, even in governments which have been actually surprisingly forward leaning, you find two things that move things forward.
- It feels to me like one of the questions that’s just calling out to be asked, and I’m sort of, I mean a lot of people are thinking about this, but I’m often surprised that not more are is how is it that we don’t know how to create computer systems that can do these things?
Takeaways
- AI capabilities are scaling rapidly, with superintelligent AI possible by 2026 or 2027.
- AI safety and alignment are critical to prevent misuse and ensure ethical deployment.
- Mechanistic interpretability helps us understand the inner workings of neural networks.
- Post-training techniques like RLHF and constitutional AI are key to improving AI models.
- Regulation may be necessary to ensure AI systems are developed responsibly.
Official description
Dario Amodei is the CEO of Anthropic, the company that created Claude. Amanda Askell is an AI researcher working on Claude's character and personality. Chris Olah is an AI researcher working on mechanistic interpretability.
Thank you for listening ❤ Check out our sponsors: https://lexfridman.com/sponsors/ep452-sc
See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc.
Transcript:
https://lexfridman.com/dario-amodei-transcript
CONTACT LEX:
Feedback - give feedback to Lex: https://lexfridman.com/survey
AMA - submit questions, videos or call-in: https://lexfridman.com/ama
Hiring - join our team: https://lexfridman.com/hiring
Other - other ways to get in touch: https://lexfridman.com/contact
EPISODE LINKS:
Claude: https://claude.ai
Anthropic's X: https://x.com/AnthropicAI
Anthropic's Website: https://anthropic.com
Dario's X: https://x.com/DarioAmodei
Dario's Website: https://darioamodei.com
Machines of Loving Grace (Essay): https://darioamodei.com/machines-of-loving-grace
Chris's X: https://x.com/ch402
Chris's Blog: https://colah.github.io
Amanda's X: https://x.com/AmandaAskell
Amanda's Website: https://askell.io
SPONSORS:
To support this podcast, check out our sponsors & get discounts:
Encord: AI tooling for annotation & data management.
Go to https://encord.com/lex
Notion: Note-taking and team collaboration.
Go to https://notion.com/lex
Shopify: Sell stuff online.
Go to https://shopify.com/lex
BetterHelp: Online therapy and counseling.
Go to https://betterhelp.com/lex
LMNT: Zero-sugar electrolyte drink mix.
Go to https://drinkLMNT.com/lex
OUTLINE:
(00:00) - Introduction
(10:19) - Scaling laws
(19:25) - Limits of LLM scaling
(27:51) - Competition with OpenAI, Google, xAI, Meta
(33:14) - Claude
(36:50) - Opus 3.5
(41:36) - Sonnet 3.5
(44:56) - Claude 4.0
(49:07) - Criticism of Claude
(1:01:54) - AI Safety Levels
(1:12:42) - ASL-3 and ASL-4
(1:16:46) - Computer use
(1:26:41) - Government regulation of AI
(1:45:30) - Hiring a great team
(1:54:19) - Post-training
(1:59:45) - Constitutional AI
(2:05:11) - Machines of Loving Grace
(2:24:17) - AGI timeline
(2:36:52) - Programming
(2:43:52) - Meaning of life
(2:49:58) - Amanda Askell - Philosophy
(2:52:26) - Programming advice for non-technical people
(2:56:15) - Talking to Claude
(3:12:47) - Prompt engineering
(3:21:21) - Post-training
(3:26:00) - Constitutional AI
(3:30:53) - System prompts
(3:37:00) - Is Claude getting dumber?
(3:49:02) - Character training
(3:50:01) - Nature of truth
(3:54:38) - Optimal rate of failure
(4:01:49) - AI consciousness
(4:16:20) - AGI
(4:24:58) - Chris Olah - Mechanistic Interpretability
(4:29:49) - Features, Circuits, Universality
(4:47:23) - Superposition
(4:58:22) - Monosemanticity
(5:05:14) - Scaling Monosemanticity
(5:14:02) - Macroscopic behavior of neural networks
(5:18:56) - Beauty of neural networks
Thank you for listening ❤ Check out our sponsors: https://lexfridman.com/sponsors/ep452-sc
See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc.
Transcript:
https://lexfridman.com/dario-amodei-transcript
CONTACT LEX:
Feedback - give feedback to Lex: https://lexfridman.com/survey
AMA - submit questions, videos or call-in: https://lexfridman.com/ama
Hiring - join our team: https://lexfridman.com/hiring
Other - other ways to get in touch: https://lexfridman.com/contact
EPISODE LINKS:
Claude: https://claude.ai
Anthropic's X: https://x.com/AnthropicAI
Anthropic's Website: https://anthropic.com
Dario's X: https://x.com/DarioAmodei
Dario's Website: https://darioamodei.com
Machines of Loving Grace (Essay): https://darioamodei.com/machines-of-loving-grace
Chris's X: https://x.com/ch402
Chris's Blog: https://colah.github.io
Amanda's X: https://x.com/AmandaAskell
Amanda's Website: https://askell.io
SPONSORS:
To support this podcast, check out our sponsors & get discounts:
Encord: AI tooling for annotation & data management.
Go to https://encord.com/lex
Notion: Note-taking and team collaboration.
Go to https://notion.com/lex
Shopify: Sell stuff online.
Go to https://shopify.com/lex
BetterHelp: Online therapy and counseling.
Go to https://betterhelp.com/lex
LMNT: Zero-sugar electrolyte drink mix.
Go to https://drinkLMNT.com/lex
OUTLINE:
(00:00) - Introduction
(10:19) - Scaling laws
(19:25) - Limits of LLM scaling
(27:51) - Competition with OpenAI, Google, xAI, Meta
(33:14) - Claude
(36:50) - Opus 3.5
(41:36) - Sonnet 3.5
(44:56) - Claude 4.0
(49:07) - Criticism of Claude
(1:01:54) - AI Safety Levels
(1:12:42) - ASL-3 and ASL-4
(1:16:46) - Computer use
(1:26:41) - Government regulation of AI
(1:45:30) - Hiring a great team
(1:54:19) - Post-training
(1:59:45) - Constitutional AI
(2:05:11) - Machines of Loving Grace
(2:24:17) - AGI timeline
(2:36:52) - Programming
(2:43:52) - Meaning of life
(2:49:58) - Amanda Askell - Philosophy
(2:52:26) - Programming advice for non-technical people
(2:56:15) - Talking to Claude
(3:12:47) - Prompt engineering
(3:21:21) - Post-training
(3:26:00) - Constitutional AI
(3:30:53) - System prompts
(3:37:00) - Is Claude getting dumber?
(3:49:02) - Character training
(3:50:01) - Nature of truth
(3:54:38) - Optimal rate of failure
(4:01:49) - AI consciousness
(4:16:20) - AGI
(4:24:58) - Chris Olah - Mechanistic Interpretability
(4:29:49) - Features, Circuits, Universality
(4:47:23) - Superposition
(4:58:22) - Monosemanticity
(5:05:14) - Scaling Monosemanticity
(5:14:02) - Macroscopic behavior of neural networks
(5:18:56) - Beauty of neural networks