Blogmark

The landscape of LLM and agentic tooling advancement is shifting too fast to think that any workflow I've figured out is going to be the most effective approach for too long.

Here is Justin Searls on this:

A skill we need to learn is how to ratchet up our ambitions until we uncover what the models truly cannot do. And each time they're improved, to once again test the fences. Confidence and certainty in one's own workflow have become trademark signs of ignorance and complacency.

This fits with my observation from this interview with developers at Anthropic:

good ideas about how to work well with LLMs have a short shelf life