Transformers—a Love Letter to a New Way of Computing
I don't think we understand Transformers enough. They're the key to Large Language Model performance, and understanding them is critical to understanding how the "brains" of our AI friends work.
This is a LONG read. Not gonna lie. But I thought it was time to write a real love letter to one of the foundational technologies that is shaping our world. ChatGPT 4.5 is coming tomorrow, probably. The magic intelligence it delivers is built on a Transformer architecture, like other large language models. But most people who use these models—even in te…
Keep reading with a 7-day free trial
Subscribe to Nate’s Substack to keep reading this post and get 7 days of free access to the full post archives.