📈Why the rise of LLMOps?
The inception of early Large Language Models (LLMs), such as BERT and GPT-2, can be traced back to 2018. However, the concept of LLMOps has only recently, almost half a decade later, surged in popularity. This increased attention can largely be attributed to the launch of ChatGPT in December 2022, which garnered significant media coverage.
LLMs have been harnessed in diverse applications since their emergence:
Chatbots, from renowned ones like ChatGPT, to more personalised versions such as Michelle Huang's interaction with her younger self.
Writing aids, ranging from editing and summarization tools like Notion AI, to specialized platforms for copywriting (Jasper, Copy.ai) or contract drafting (Lexion).
Programming assistants that aid in code writing, debugging (GitHub Copilot), testing (Codium AI), and even identifying security threats (Socket AI).
As the development and deployment of LLM-powered applications become more commonplace, individuals have started sharing their experiences:
"It's easy to make something cool with LLMs, but very hard to make something production-ready with them." - Chip Huyen
It has become evident that crafting production-grade applications powered by LLMs presents its own unique set of challenges, distinct from those encountered when building AI products with conventional ML models. To navigate these challenges, there's a growing need for the development of new tools and best practices to efficiently manage the lifecycle of LLM applications, thereby leading to the rise in the usage of the term "LLMOps."
Last updated
