Grandma Speaks Faster Than LLMs

only retards use self hosted llms 10 toks/s? lmao my 90yr old grandma can speak faster than qwen2.5b-creamy-abliterated-4bit-gguf

A dark-mode social media post mocks people who self-host small language models, joking that a 90-year-old grandmother can speak faster than a model generating only 10 tokens per second. It is shared as a sarcastic reaction to slow local LLMs and overly technical AI setups.

• @gentoomemes / 18277 2025-08-18