I went with a more clinical approach and used models that were available a half year ago but I also was interested in using LLMs to write unit tests. You can learn the details of that experiment at https://www.infoq.com/articles/llm-productivity-experiment/ but the net of what I found was that LLMs improve developer productivity in the form of unit test creation but only marginally. Perhaps I find myself a bit skeptical on the claims from that Assembled blog on significant improvement.
Comments
I went with a more clinical approach and used models that were available a half year ago but I also was interested in using LLMs to write unit tests. You can learn the details of that experiment at https://www.infoq.com/articles/llm-productivity-experiment/ but the net of what I found was that LLMs improve developer productivity in the form of unit test creation but only marginally. Perhaps I find myself a bit skeptical on the claims from that Assembled blog on significant improvement.