Wojciech Zaremba
Showing 1–2 of 2 transcripts.
- Lex Fridman2h 52m
Wojciech Zaremba: OpenAI Codex, GPT-3, Robotics, and the Future of AI | Lex Fridman Podcast #215
OpenAI co-founder Wojciech Zaremba leads the organization's language and code generation teams while articulating a "maximize compute" resolution to the Fermi Paradox and defining consciousness as self-referential metacompression. He advocates for expanding human action space through advanced AI to mitigate existential risks, citing recent successes in robotics and natural language programming alongside a cautious approach to deploying empathetic models in high-stakes environments. Zaremba further proposes a distributed power strategy for artificial general intelligence involving tax equity and universal basic income, emphasizing that iterative public critique is preferable to withholding powerful systems until they are theoretically perfect.
- Y Combinator55 min
An AI Primer with Wojciech Zaremba
Wojciech Zaremba, Craig Cannon
OpenAI founder Wojciech Zaremba details the organization's mission to develop artificial general intelligence through robotics manipulation and deep reinforcement learning, while highlighting the significant hurdles in defining real-world rewards compared to digital game environments. He contrasts narrow, supervised learning successes in commercial applications with the theoretical complexities of general AI, noting that current deep learning architectures struggle to match human learning speeds without massive computational resources. Zaremba concludes by predicting that rapid automation will necessitate societal shifts like Universal Basic Income and recommends specific educational resources for those seeking to understand these emerging technologies.