Gareth is out, and the "most interesting person on the planet" is in. Adam is joined by Chalice AI’s SVP of Data Science, Tylynn Pettrey, for a high-speed masterclass on the history of predictive modeling, from 19th-century planetary orbits to the transformer architectures powering 2026’s agentic workflows.Tylynn breaks down complex math into plain English, explaining why "more data" isn't always better (the curse of dimensionality), why some algorithms break the bank on computing costs, and why AdTech is finally catching up to High-Frequency Trading. Plus, we answer audience questions on Quantum Computing, the death of the R programming language, and why Tylynn kept a mass spectrometer in her garage during the pandemic.00:00 Intro: The Data Science Takeover02:40 Linear Regression: Connecting the Dots05:37 Logistic Regression & The S-Curve09:12 Decision Trees: The Medical Diagnosis Logic10:56 K-Nearest Neighbors & Lookalikes13:13 Neural Networks: The Factory Assembly Line16:18 Back Propagation: Learning from Error18:32 Random Forests: Crowd-Sourced Math19:50 Support Vector Machines (SVM) & Big O Notation22:25 Skills for 2026: Python vs. R vs. Rust24:54 Deep Learning & Sparse Data28:40 Transformers & Attention Mechanisms33:18 The Imbalance Problem: Synthetic Data38:54 Scoring: AUC and Probability41:42 Q&A: Is Agentic AI Hype?43:10 Q&A: Quantum Computing in AdTech48:25 Q&A: Is R Dead?51:15 The Mass Spectrometer in the Garage53:20 Curate AI: Low Latency Scoring in Rust
