The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Compositional ML and the Future of Software Development with Dillon Erb - #520
12 snips
Sep 20, 2021 
Dillon Erb, CEO of Paperspace, dives into the world of compositional machine learning and its potential to revolutionize software development. He discusses the evolution from experimental ML to a disciplined engineering framework and the challenges of integrating cloud technology. The conversation highlights the debate around the role of notebooks versus traditional coding practices in data science. Also on the table is Paperspace’s innovative new product, Workflows, which aims to streamline ML application development into a more cohesive system.
AI Snips
Chapters
Transcript
Episode notes
ML as Software Engineering
- Machine learning is shifting from an experimental process to one requiring engineering rigor.
- This shift necessitates a change in mindset and tooling for AI development.
Compositional AI's Rise
- Compositional AI, remixing existing models like first-order motion or CLIP, drives innovation.
- This approach allows developers to create novel applications by combining pre-trained models.
Compositional AI Examples
- A lip syncing app built on the first-order motion model went viral, demonstrating compositional AI's power.
- Artists are remixing OpenAI's CLIP model with others, creating fantastical art and NFTs.