There is a quiet tension building in the research community, and it surfaced this week in a candid post from a researcher frustrated by the NeurIPS 2026 Concept & Feasibility Track. The premise of the track is admirable, it was designed to welcome ideas that could not be fully validated in a single paper, acknowledging that some of the most interesting concepts need room to breathe. Yet this researcher's experience tells a different story: reviewers praised their originality but penalized them for insufficient experimental scope, directly contradicting the track's stated policy. No reviewer responded to the rebuttal. No engagement followed. The researcher is left wondering if this track even functions as intended. It is a specific complaint, but it echoes a broader pattern we have seen elsewhere, including how Neurosurgery Match Requirements Highlight Growing Pressure on Medical Students and how Navigating AI/ML Job Requirements: A Shift in Expected Skills reveal growing gaps between stated rules and actual expectations.
What stands out here is the gap between policy and practice. The C&F track was meant to be a safe harbor for ambitious, unfinished work, an acknowledgment that not every good idea arrives fully formed with a complete experimental appendix. But if reviewers default to the same standards they apply to full-paper submissions, the track loses its purpose. The researcher reports that their reviewers complained about the scope of experiments despite the track explicitly noting that validation may not be possible in a single paper. This is not a minor oversight; it suggests that either reviewers are not reading the track guidelines, or the guidelines themselves are not being enforced by area chairs. The lack of any rebuttal response only deepens the concern. A track built to encourage risk-taking is instead reproducing the very conservatism it was meant to counter.
For our readers who are researchers, students, or practitioners submitting to venues like this, the practical takeaway is uncomfortable but clear: do not assume the published guidelines will match the actual review behavior. You may be judged by criteria the track says it will not use. That does not mean you should avoid these tracks, they remain important experiments in how we evaluate research, but it does mean you should prepare for a mismatch. Treat the rebuttal period not as a conversation but as a final statement, because engagement may not come. This experience is one data point, but it is a telling one. If area chairs and senior reviewers do not actively enforce track-specific policies, then the innovation these venues promise will remain aspirational.
One specific detail worth watching is whether the NeurIPS organizers address this gap publicly after the cycle closes. If they do, it will signal a willingness to iterate on the review process itself. If they do not, the message to authors is that the track's stated purpose is secondary to whatever the reviewers decide in the moment. We would tell any reader considering this track to submit, but to submit with eyes open, and to budget for the possibility that your idea will be judged by rules other than the ones printed on the page. The future of data management and AI research depends on ideas that are bold enough to be incomplete, but only if the review process is bold enough to accept them.