GOTO - The Brightest Minds in Tech Beyond the Hype: What AI Actually Can (and Can't) Do • Jodie Burchell & Michelle Frost
10 snips
Apr 24, 2026 
Jodie Burchell, Senior Data Science Developer Advocate with a background in clinical psychology, biostatistics, and NLP. They unpack what generative AI and transformers actually are. They talk about why ML fundamentals and data quality still matter. They examine ethics, bias, and practical governance. They trace AI boom–bust history and clarify AGI versus narrow AI.
AI Snips
Chapters
Books
Transcript
Episode notes
From Clinical Psychology To NLP Data Science
- Jodie Burchell described her unusual path from clinical psychology to data science and NLP.
- She trained as a clinical psychologist, did a postdoc in biostatistics, then moved into industry in 2015 and focused on NLP from 2016 onward.
AI Is Ambiguous And Often Means Generative Models
- The term AI is poorly defined and currently often means generative AI models built with transformer architectures.
- Jodie warns people conflate generative models with applied ML systems and that generative AI remains a relatively unproven branch with unclear use cases.
Handle Data Quality When Building Vector Search
- Don't ignore data quality when building vector search or retrieval systems; chunk size and domain knowledge matter.
- Jodie explains you must choose meaningful chunk sizes, handle missingness and understand collection biases for good embeddings.
