Ep 47: Chief AI Scientist of Databricks Jonathan Frankle on Why New Model Architectures are Unlikely, When to Pre-Train or Fine Tune, and Hopes for Future AI Policy

Ep 47: Chief AI Scientist of Databricks Jonathan Frankle on Why New Model Architectures are Unlikely, When to Pre-Train or Fine Tune, and Hopes for Future AI Policy

1:04:25 Nov 12, 2024
About this episode
Jonathan Frankle is the Chief AI Scientist at Databricks ($43B), which he joined through the acquisition of MosaicML in July 2023. Databricks has over 12,000 customers on the cutting edge of AI; Jonathan works to anticipate their needs and offer solutions even as the tech is rapidly evolving. [0:00] Intro[0:52] Incentives and Team Motivation at Databricks[2:40] The Evolution of AI Models: Transformers vs. LSTMs[5:27] Mosaic and Databricks: A Strategic Merger[7:31] Guidance on AI Model Training and Fine-Tuning[11:11] Building Effective AI Evaluations[16:02] Domain-Specific AI Models and Their Importance[19:37] The Future of AI: Challenges and Opportunities[25:07] Ethical Considerations and Human-AI Interaction[29:13] Customer Collaboration and AI Implementation[30:45] Navigating AI Tools and Techniques[35:41] The Role of Open Source Models[36:46] AI Infrastructure and Partnerships[48:27] Academia's Role in AI Research[52:09] Ethics and Policy in AI[57:47] Quickfire With your co-hosts: @jacobeffron - Partner at Redpoint, Former PM Flatiron Health @patrickachase - Partner at Redpoint, Former ML Engineer LinkedIn @ericabrescia - Former COO Github, Founder Bitnami (acq’d by VMWare) @jordan_segall - Partner at Redpoint
Select an episode
0:00 0:00