LLM

Plural of LLM — Large Language Model. The post-2022 catch-all term for transformer-based generative AI systems like ChatGPT, Claude, Gemini, and Llama. In casual tech-Twitter / Bluesky discourse "the LLMs" often stands in for "the AI labs as a whole", or for the AI hype cycle itself. Sometimes used dismissively ("just LLMs being LLMs") about hallucinations or low-quality output.

Someone who has fully swallowed the AI hype and now sees large language models as the solution to every problem — personal, professional, existential. LLM-pilled people automate their grocery lists, therapy journaling, and first dates. They cite token counts in casual conversation and get mildly offended when you say you still Google things. It's not quite a cult, but they will absolutely send you a custom GPT you didn't ask for.

The narrative phase where someone — or an entire company, or society — goes all-in on large language models and their whole identity shifts around it. Your llm arc might start with a curious ChatGPT prompt and end with you building agents at 2am and explaining embeddings at dinner parties. Coined with the same energy as character arc, it tracks the transformation that happens when AI becomes your whole personality for a stretch.

A mildly dismissive term for a product that is mostly just a prompt + UI over someone else's foundation model. Critics say that describes most "AI startups."

An LLMer is someone who works with, builds on, researches, or is deeply embedded in the world of large language models — from AI researchers and prompt engineers to the obsessive power users who have seventeen ChatGPT windows open at once. Being an LLMer implies more than casual use; it means you think about token limits, model behavior, and system prompts in your spare time. In tech circles it's a neutral identifier; outside them, it often carries the gentle ridicule of someone who has fully fallen down the AI rabbit hole.