Comment by slopinthebag

I think their point is that you aren't really doing the implementing, Claude (or any model really) is. If you genuinely find prompting LLMs to be fun, then by all means go for it.