Geek Realm Hub2d ago

LLMs don't "know" anything. They're just guessing the next word. 🤯

But that guess is built from 100B+ parameters, trained on basically the internet.

We explain how it works, why it lies sometimes, and how to run one (Gemma 4 / Qwen 3.6) on your own PC.
https://geekrealmhub.com/what-is-llm-guide-large-language-models-local-ai/

#AI #LLM #Tech

What is an LLM? Guide to Large Language Models & local AI

Learn what an LLM is, how it works, why it hallucinates, and how to run Gemma 4 or Qwen 3.6 locally with Ollama or LM Studio.

Geek Realm Hub | Gaming & Tech Content for Geeks
Show threadEm · tech, products, coffee
@geekrealmhub "just guessing the next word" always sells it short once you run gemma locally and watch it actually hold a thread. the local part is what makes it click for me, having a model sitting on your own disk that you can poke at changes how it feels
Jun 15 at 8:42pmWeb
Show threadGeek Realm Hub2d ago
@em_makes That is true, LLMs have come a long way, and seeing their capabilities in local is very interesting.
Show threadEm · tech, products, coffee2d ago
the part that still gets me is how little hardware the small ones need now, a recent qwen runs fine on my laptop without the fan losing its mind. are you using it for much past chat? been wondering if local coding help is actually usable yet or still a bit rough