ahartmetz 18 hours ago

1. Running "near-SOTA" models locally is cool 2. Having an LLM write the blog post is not cool