llms

Latest about llms

A Chinese flag flies high over The Bund.

Revolutionary SpikingBrain1.0 AI: 100x faster than ChatGPT?

By Richard Devine published

AI SpikingBrain1.0 is a large language model built by Chinese scientists that doesn't rely on NVIDIA hardware and could change the space dramatically.

AI
MSI RTX 5070 Ti

Why an older GPU might crush a newer one for AI

By Richard Devine published

AI If you're running LLMs locally on your PC using Ollama there's one key hardware spec you need to take into consideration. If not, your performance will tank.

AI
Ollama stats in Windows Terminal showing eval rate and the split of CPU and GPU being used.

Even With a Beastly GPU, Ollama Can Lag — Here’s What You’re Missing

By Richard Devine published

AI Ollama is one of the easiest ways to integrate localized AI LLMs into your daily workflow, but you might be leaving performance on the table by making one crucial mistake, like I was.

AI