Why I Publish on Zenodo

A note on my choice of preprint repository, and what it means for my work.

Context

Over the past year, I have published two preprints on Zenodo:

• Crowdsourced Semantic Cache Network (DOI: 10.5281/zenodo.19401231)
• Reasoning Over Recall: Why Statute-Only RAG Degrades Performance on Doctrinal Legal Tasks (DOI: 10.5281/zenodo.19522950)

I have also presented related work at FOSSASIA Summit 2026 and Open Source Summit India 2026. I sometimes receive questions about why I chose Zenodo specifically, so I wanted to lay out my reasoning clearly.

Why Zenodo was the right fit for me

Zenodo is operated by CERN as part of its OpenAIRE infrastructure. It provides a persistent DOI, a timestamped archival record, and immediate open access to any reader regardless of institutional affiliation. For a first-time author without a university affiliation, these properties mattered a great deal to me.

My research concerns the behaviour of current-generation language models on legal reasoning tasks. The models I evaluated — GPT-5, DeepSeek R1, Gemini 2.5 Pro — are updated on short cycles, and the failure mode I documented (which I call Statutory Myopia) was something I wanted to share with the legal-AI community while the finding was still timely. Zenodo allowed me to establish a dated, citable record of the work without a lengthy submission queue, and to make it freely readable by anyone who might benefit from it — a developer building a legal research tool, a fellow student, a researcher at an institution without journal subscriptions.

Why not arXiv

I want to address this honestly, because I suspect it is the first question people have.

arXiv requires endorsement for first-time submitters in most categories: an established author in the relevant subfield must vouch for your work before it is accepted. I understand and respect the purpose of that policy — it helps maintain the quality and relevance of what appears on the platform.

However, as a Class 12 student publishing independently for the first time, I did not have an existing relationship with a published researcher who could provide that endorsement. I also did not want to delay the release of my findings by several months while I sought one, given how quickly the model landscape was shifting. Zenodo did not require that intermediary step, and so it became the practical and timely choice.

I hope that as my work develops, I will have the opportunity to submit to arXiv and to peer-reviewed venues in due course. I very much look forward to that.

What I value about traditional publishing

I want to be clear: I do not see Zenodo as a replacement for peer review. The journal and conference review process serves an essential function in science — it subjects methodology to expert scrutiny, catches errors, and strengthens arguments before they enter the permanent literature. I have enormous respect for that tradition, and I benefit from it constantly as a reader.

What Zenodo offers is something different but complementary: a way to share early-stage findings, establish a dated record, and invite community feedback before or alongside formal submission. In fast-moving fields like AI evaluation, that immediacy can be genuinely useful to other researchers who are working on related problems in real time.

What a Zenodo DOI represents

A DOI on Zenodo does not claim the imprimatur of peer review. I would never suggest otherwise. What it does provide is:

• A permanent, timestamped record of the work as it existed at the point of upload.
• A stable citation target that other researchers can reference.
• Open access without a paywall or institutional login.
• Version tracking, so corrections and revisions are transparent.
• Space for linked artifacts — code, datasets, technical appendices — alongside the paper itself.

For my work, which includes a reproducible evaluation harness with SHA-256 caching and a controlled-retrieval blacklist protocol, the ability to link those artifacts directly to the DOI is particularly valuable. Reproducibility matters to me, and Zenodo supports it well.

Looking ahead

I am seventeen years old, in my final year of school, and at the very beginning of what I hope will be a long engagement with research. I chose Zenodo because it was accessible, immediate, and aligned with the open-science values I care about. It allowed me to share my work without waiting for circumstances I did not yet have.

I do not see this as a permanent arrangement. I hope to submit future work to peer-reviewed conferences and journals, to participate in the review process as both author and, eventually, reviewer, and to contribute to the scholarly conversation in the more traditional ways that the academic community has built over centuries.

But for now — for a first-time author who wanted to share a finding about Crawford v. Washington and RAG degradation while the result was still fresh — Zenodo was the right home. And I am grateful that it exists.

— Kannan Murugapandian

Singapore, 2026