TDD agent loop value debate
TDD inside the agent loop - theater or actual value?
2 experts are actively discussing the implications.
Why this matches
Research & technical analysis reaction
1 attributable expert contribution
· Sung Kim
“Is TDD worth it when coding with AI agents? No, TDD is a process that doesn't seem to “come natural” to models. It's like an uphill battle against the training data, and takes a lot of iterations on a prompt to get it to follow the process most of the time.…” evidence ↗
2 experts
1 community
1 sources clustered
“Is TDD worth it when coding with AI agents? No, TDD is a process that doesn't seem to “come natural” to models. It's like an uphill battle against the training data, and takes a lot of iterations on a prompt to get it to follow the process most of the time.…”
““I think at this point there is generally more and more evidence that being overly specific about how we want a model to do something is not a sustainable approach. Instead, we should find as many ways as we can to monitor the outcomes and give feedback.” m…”
2 experts discussed this · 3 posts
Sung Kim: Is TDD worth it when coding with AI agents? No, TDD is a process that doesn't seem to “come natural” to models. It's like an uphill battle against the training data, and takes a lot of iterations o…
Simon Willison: Interesting result, but this kind of experiment is always hard to be confident about because so much depends on the details - in this case it looks to be a comparison between these two quite length…
Simon Willison: Recent prompt guidelines from Anthropic and OpenAI recommend less detailed rules and more "use your own judgement", since the models are stricter rule followers and are hence likely to focus on the…
Open the full discussion →