

What Wall Street didn't see coming
May 23, 2025
Discover how Wall Street's perception of artificial intelligence evolved from skepticism to fascination, driven by a pivotal Goldman Sachs report. Delve into the significance of quality data in machine learning, highlighting its advantage over mere quantity. Explore the complexities of deep learning, where less feature engineering is required, and critique the reliance on horizontal APIs. The discussion emphasizes the future importance of context awareness in enhancing digital assistants.
AI Snips
Chapters
Transcript
Episode notes
Early Wall Street AI Blindspots
- Wall Street initially ignored AI advances like generative models and reinforcement learning in 2016.
- These themes have since become central to AI discussions, highlighting early underestimations.
Generative Models & Efficiency
- Generative models create new data unlike discriminative models that categorize data.
- Efficiency gained from AI isn't just optimization but also from models needing less data and computation.
Reinforcement Learning Impact
- Reinforcement learning solves tasks through trial and error, crucial for robotic control and similar problems.
- Despite current limited production use, its real-world applications are approaching.