LLM tech stack, Open source startups, Building products for the government | Brian Raymond, cofounder and CEO of Unstructured
Infinite Curiosity Pod with Prateek Joshi
00:00
The Evolution of NLP Text Tags
A traditional NLP text tag is pre-processed and then pushed through a machine learning pipeline, single model or orchestrate it. What's changing in a lot of ways is a move from a graph database to a vector database. And so the introduction, like the reliance on embeddings and the reliance on a vector store is a key shift. The emergence of these frameworks to interact with the models, rather than just pushing the data through the models as it's available right? Right.
Transcript
Play full episode
