Stories by tosh

Summarise cached articles with Mixtral-8x7B on pure CPU

Fast Inference of Mixture-of-Experts Language Models with Offloading

Ohm

Rails 8 will ship with a default Rubocop linter

The roadmap to Product/Market Fit maybe

How Robust Are LLMs to In-Context Majority Label Bias?

Frankenmerge

Hammock Driven Development (2010) [video]

Monolith

MLX: Speculative Decoding

Notus-7B

Dolphin-2.6-Mistral-7B

Benchmarking Apple's MLX vs. Llama.cpp

Do Androids Know They're Only Dreaming of Electric Sheep?

Copyright is not a moral right, it's a monetization strategy

Chatglm3-6B-32k

K-Quants

China Copies

MonadGPT

Listening to YouTube in the Background