[physicsguy]

It’s only got easier to run models locally though, if performance continues to improve on smaller models, something like running ollama locally might be ‘good enough’ for a lot of people.