Edge Of AI Podcast · The Edge Of Company, Inc

The Path to Human-Level Intelligence with Peter Voss

·58 min·3 clips

Voss says his company replaced 3,000 agents on Valentine’s Day, then paused commercialization to chase AGI.

1. Edge Of AI Podcast centers on Peter Voss and the path to human-level intelligence through AGI and cognitive AI. 2. Ron Levy hosts Peter Voss, who says he coined “AGI” in 2001 and built IGO, a system that thinks, learns, and reasons. 3. The episode asks what separates narrow AI from AGI and whether a cognitive approach can reach human-level intelligence. 4. Voss contrasts narrow AI examples like chess and container optimization with the original dream of a thinking machine. 5. He says narrow AI relies on the intelligence of programmers and data scientists, not on the system’s own understanding. 6. He points to ChatGPT and large language models as broad but still limited systems that need prompt engineering, fine tuning, and external databases. 7. Voss says ChatGPT can summarize, write poems, and help with coding, but it is not reliable and can “make up stuff.” 8. He cites DARPA’s “three waves of AI” and places cognitive AI in the third wave. 9. He describes the third wave as adaptive, real-time intelligence that can reason about what it is doing. 10. Voss says AGI should learn interactively like a human, generalize across domains, and use metacognition. 11. He argues that a human child uses about 20 watts and only a few million words before language, unlike models trained on 10 trillion tokens. 12. He says the right approach is to start from the ground and teach the system basic concepts first. 13. Voss compares this bootstrapping process to learning 200 words in a language, then using them to learn thousands more. 14. He says Aigo’s commercial work used short-term memory, long-term memory, and reasoning to remember prior conversations. 15. He says the system can identify a caller by caller ID, depending on customer security rules and verification requirements. 16. He describes a three-layer architecture: common-sense core knowledge, application-specific knowledge, and a user-specific outer layer. 17. Voss says the best systems should feel like one human agent remembering past interactions and changing orders without repeating work. 18. The interview style is direct, technical, and conversational, with Levy pressing for practical examples and Voss answering in long, structured explanations. 19. People interested in AGI, call-center automation, and AI architecture will get the most from this episode. 20. Listeners wanting a light, nontechnical conversation about consumer apps may skip it.

As heard by us

A clear case for AGI that lands on customer service, memory, and desk work.

Peter Voss grounds the discussion in the origins of AGI and in a practical case for why human level AI matters. The strongest parts are the plain examples: customer service, memory across conversations, and the possibility that desk bound cognitive work could be done cheaper and…

Read the full review in PlayNext →

Why you'd press play

Hear AGI from the man who coined the term.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Edge Of AI Podcast