There's a quiet tension buried in the review process of top-tier conferences, and the question posed here deserves more than a shrug. When a reviewer asks for substantial additions that push a paper's scope beyond its stated boundaries, yet the submission is already at the page limit, the implicit suggestion is that the work belongs in supplemental materials or appendices. But that framing misses something important. Retracting a paper over this exact concern is not an overreaction; it is a strategic recognition that a conference publication can become a barrier to a later journal version. That is not a trivial administrative headache. It is a structural mismatch between what reviewers ask for and what conferences are designed to deliver.
This is where the conversation connects to a broader pattern we see across the AI research community. The pressure to pack more into a single submission often collides with the practical realities of reproducibility and data quality. Consider the experience of one practitioner who found that automated detection tools flagged genuine reviews as AI-generated, which skewed their sentiment model. That is not a niche problem; it is a signal that the tools we rely on for evaluation are themselves in need of scrutiny. Similarly, the push to expand a paper's scope in response to reviewer feedback can introduce new claims, new data, or new methods that were never part of the original contribution. And if those additions are not fully validated, they can undermine the very credibility the authors are trying to build. In the same way, the rise of edge-deployed computer vision models has forced a reckoning with what "performance" actually means when you are optimizing for mobile constraints rather than benchmark glory. The lesson across these examples is consistent: clarity about scope is not a limitation; it is a form of respect for the reader and for the scientific record.
So what should a researcher do when faced with such a review? First, recognize that a request for lengthy additions is often a sign that the work has outgrown the venue. That is not a failure on your part. It is an opportunity to step back and ask whether the journal route would serve the work better. The instinct to retract was not just about avoiding a future conflict; it was about preserving the integrity of the contribution. We would tell a reader who asked that the supplemental material should complement, not carry, the main narrative. If a reviewer's suggestions would fundamentally alter the paper's core argument or require new experiments that belong in a longer format, that is a valid reason to withdraw and resubmit elsewhere. The real question is not whether you should comply with the review, but whether the venue can honestly accommodate the work you are trying to present. And that is a decision worth making deliberately, not under pressure. Watch for the next time a reviewer asks for "just a few more details" and consider whether those details are actually a new paper in disguise.