Interview
Charles Isbell and Michael Littman: Machine Learning and Education | Lex Fridman Podcast #148
- Lex Friedman intends to occasionally moderate debates between disagreeing guests and plans to upgrade the video setup to three or more cameras, though he acknowledges the two-guest format is an experimental process requiring potential improvements to audio and video quality.
- Charles Isbell and Michael Littman anticipate divergent pedagogical approaches, with Isbell expecting to provide directional pushes to students who may not know where to search, while Littman argues that struggle must be paired with hope to remain productive.
- Isbell predicts that students will eventually appreciate rigorous machine learning assignments and that the "Smooth and Curly" personas will remain a lasting element of the video series, which they plan to use to discuss episodes of Westworld.
- Historical perspectives on education and research are projected to show that Bell Labs was a unique environment for pure research that no longer exists, while the "drown proofing" requirement at Georgia Tech was discontinued around 1983 or 1984.
- Isbell forecasts that 85% of students will return to traditional campus life despite the disaggregation of education caused by the pandemic, whereas online master's programs are expected to reach 11,000 students with a 60% acceptance rate by offering a dispersed experience in two to five years.
- Littman predicts that the online teaching experience during the pandemic will be viewed as a struggle that is now over, with students preferring in-person connection for better information retention despite feeling they learn equally well online.
- Isbell expects that video games and VR will fundamentally alter the labor market and human behavior over the next century by pulling people into virtual worlds, potentially leading to a societal division between those with traditional university experiences and those with credentials only.
- Both speakers anticipate that AI risks stem more from everyday tools making inefficient decisions rather than superintelligence, and that future educational models will facilitate lifelong learning and broader access.
- Isbell expects the online master's program to foster tribal belonging and return graduates as teaching assistants, while noting that lecturers at Georgia Tech currently teach over 50% of credit hours and are often the most remembered by alumni.
- Personal and professional trajectories are expected to involve strong collaboration between the speakers, with Littman estimating their "inner product" of friendship at roughly 0.7 and Isbell expressing gratitude for how their partnership maintains energy in teaching.
- Specific details regarding the "Smooth and Curly" personas are expected to include their spontaneous origin during lecture preparation and derivation from a cartoon of criminals or a prisoners' dilemma, with plans to discuss Westworld and Ex Machina in future episodes.
- Isbell expects that luck may favor his specific wishes regarding the future, while Littman anticipates that the automated assistant project developed at AT&T Labs eventually contributed to the creation of Siri.