Onward, a Fundrise Production 56: Will AI do your job? with Raza Habib, Member of Technical Staff at Anthropic
19 snips
Mar 25, 2026 
Raza Habib, Member of Technical Staff at Anthropic and former Humanloop CEO, explains why AGI is harder to define than expected and why recent model gains feel transformative. He traces AI’s fast timeline, predicts near-term automation for senior engineering work, and discusses economic, geopolitical, and policy implications in short, sharp takes.
AI Snips
Chapters
Transcript
Episode notes
Define AGI By Capabilities Not Buzzwords
- AGI lacks a single agreed definition, so measure it by capability thresholds like "economically transformative" tasks instead of a label.
- Raza endorses Shane Legg's view: intelligence is the ability to achieve goals across many domains, and practical thresholds let you test for AGI.
How A Researcher's Skepticism Turned To Conviction
- Raza's AGI view evolved through his PhD and repeated surprises from deep learning successes that kept disproving his skepticism.
- By the time he joined Anthropic he was already convinced the technology was likely transformative.
Teach Purpose And Craft Over Job Panic
- Teach kids values, meaning, and community as much as career skills because many cognitive jobs may be automated in 10–20 years.
- Raza suggests focusing on life purpose, craft, and distribution of societal benefits, not just high-paying jobs.