Fireside Chat #10: Large Language Models: Beyond Proofs of Concept

Federico Bianchi is a post-doctoral researcher at Stanford University, with experience in building large Vision-Language models (e.g., FashionCLIP), model understanding (e.g., Limitations of Vision-Language Models) and fairness and bias (e.g., Bias in Text-to-Image, EvalRS). In this fireside chat, Federico joins Hugo Bowne-Anderson, Outerbounds’ Head of Developer Relations, to discuss where the field of large language models is today and what we can expect to see in terms of business value, once the hype dies down. After attending, you’ll have an understanding of - What large language models are and are not currently capable of; - Ways to get started with LLMs today and how to stay up to date; - How LLMs can be used to build sustainable business value and defensible business moats; - How to think about productionizing LLMs and incorporating them into existing software stacks; - Ethics and the trade-off between the usefulness of language models and the associated risks; - Privacy and risks of content regurgitation. And much more! The fireside chat will be followed by an AMA with Federico and Hugo at slack.outerbounds.co. 00:00 Prelude 03:55 The fireside chat begins! 10:09 What even are Large Language Models? 17:14 So what aren't LLMs capable of? 22:19 What are all the LLMs out there and how should we think about the space? 26:45 How can people get started with LLMs today? 35:10 How to productionize LLMs and incorporate them into existing software stacks 40:54 How can LLMs be used to create sustainable business value? 47:05 LLMs and the associated risks 56:06 Bias in LLMs