Large Language Models Continue To Impress

As you all know, I’ve been experimenting a lot with AI, and I’m truly impressed with the software that I can currently run on the computer. That has really got me wondering: in just two years’ time, the speed of consumer computers would have almost doubled, and VRAM in GPUs will also increase. With that comes more tokens and computing power for Large Language Models (LLMs).

ChatGPT has made some huge strides and has garnered the attention of even “normies,” but in the underground world of LLMs, there are far more powerful and uncensored models to choose from. Vicuna and Falcon are two of them that are very popular but have not garnered nearly as much attention. Some models excel beyond 30 gigabytes of storage space, which generally means the model has a much larger framework to work with.

Imagine, even with today’s technology, where ChatGPT-like software can write code for you (with some mistakes), write articles, and even books. And if you wanted, you could even ask it to create a song for you, not just in English but in other languages. LLMs have become very impressive, but this is just the start of the AI revolution. I predict that in less than a year, we will have AI models that can keep up a conversation with you. They may remember the smallest details of your conversation, even if days or even weeks have passed since the subject was brought up.

Almost every single week, there is a new update, a new model to choose from with some different changes, and you can train your own. Things are going to get extremely incredible when regular people can have an AI assistant with them at all times. It is scary but impressive.

When it comes to AI models, I trust open-source models the most. I think many consumers should be wary of corporate models, but that doesn’t mean you should fear them. Experiment and try it out yourself!