Ep 11: Exploring AI for application developers (with Sinan Ozdemir)

The Artifact

Link to GitHub for Sinan's book https://github.com/sinanuozdemir/quick-start-guide-to-llms

Sinan on LinkedIn: https://www.linkedin.com/in/sinan-ozdemir/

In this episode, I engage in a compelling conversation with Sinan Ozdemir, a seasoned expert in AI and LLMs who shares over a decade of experience. We navigate the fascinating progression of large language models (LLMs), discussing their deterministic and non-deterministic characteristics, the art of prompting, and the significance of model architecture and training data.

Our discussion also covers the latest strides in AI, including the creation of smaller and more efficient models capable of running on everyday devices. We delve into the intriguing world of AI benchmarks, the surprising performance of specialized models compared to industry giants like GPT-4, and the innovative approaches required for continued advancement.

Additionally, we highlight the ongoing tension between major corporations with extensive resources and smaller companies using open-source models to stay competitive. Join us to explore how the demand for personalized AI is influencing both research and corporate strategies.

SWE-bench: Can Language Models Resolve Real-World GitHub Issues? https://arxiv.org/abs/2310.06770