> ## Content Index > Fetch the complete content index at: https://www.dailyferment.com/llms.txt > Use this file to discover other available public pages before exploring further. # arXiv now limits you to two papers a month. It's rationing the reader. - URL: https://www.dailyferment.com/arxiv-two-submissions-per-month-moderator-time/ - Published: 2026-10-04T16:00:00.000Z - Updated: 2026-10-04T15:59:59.000Z - Description: arXiv got 40,363 submissions in September, up from 20,569 two years ago. The new two-a-month cap doesn't ration writing. It rations the volunteers who check it. - Author: Rana Bilal Zafar - Tags: AI, Engineering, #reframe On Thursday, 1 October, arXiv, the preprint server where much of AI research is posted, changed its rules. It "now limits submitters to up to two submissions per calendar month, with a limit of three total active submissions." The reason is in the numbers it published. In September 2016 arXiv received 9,869 submissions. In September 2024, 20,569\. This September, 40,363, "which in turn generated almost 9,000 support tickets." Submissions to the cs.AI category went up more than six times in two years. I publish twice a day here, with AI in the toolchain, so this one is close to home. I wanted to understand what arXiv is actually running out of. ## Why did arXiv limit submissions? Not storage, and not readers. Moderators. arXiv says "the heart of arXiv's process for detecting and rejecting low-quality papers is our many volunteer moderators", volunteers with terminal degrees in their field who check each submission before it goes public. What they're seeing, in arXiv's words: "thin papers of narrow scope", "'salami' papers, where a single work is broken up and submitted as a set of smaller papers", and "a marked increase in dense, AI-written papers." Thomas Dietterich, who chairs arXiv's editorial advisory council, put the shape of it plainly: "a relatively small proportion of authors are submitting a large number of low-quality papers and consuming a disproportionate fraction of the moderators' time." One detail tells you what the cap is really counting. Rejected papers count toward the limit, "because it is submissions that consume moderator time." ![arXiv submissions in September: 9,869 in 2016, 20,569 in 2024, 40,363 in 2026. From 1 Oct: 2 submissions a month, 3 active, rejected papers count.](https://storage.ghost.io/c/d7/b7/d7b77195-48c5-45ab-8688-51750879d1d6/content/images/2026/10/df28-arxiv.png) Monthly submissions, from arXiv's post of 1 Oct 2026. ## Isn't this just arXiv fighting AI slop? That was my first reaction, and arXiv has been doing that for a year. In October 2025 it stopped taking computer science review and position papers that hadn't passed peer review. In May it announced year-long bans for papers with unchecked AI output. This rule is different. It doesn't look at whether a paper is AI-written. It caps everyone, which means it isn't a quality filter at all. It's a budget for the one resource that didn't get cheaper when writing did: a qualified person's attention. ## Where have we seen this before? In email, in 1992\. Cynthia Dwork and Moni Naor proposed fighting junk mail by making the sender "compute a moderately hard, but not intractable, function in order to gain access to the resource." Adam Back's Hashcash built on the idea in 1997 to "throttle systematic abuse of un-metered internet resources such as email". The design rule was that the cost should be "efficiently verifiable, but parameterisably expensive to compute": cheap for the checker, costly for the sender. arXiv's cap does the same job by a blunter route. It doesn't make sending cost something in proportion to volume. It makes every author's monthly allowance the same. Someone posting forty AI-written papers and someone with two careful results in a busy month now hit the same wall. That's the criticism Prachee Avasthi made the day after: "How much science enters the record each month is now set by human bandwidth instead of by how much science is being done." She'd rather arXiv detect bad actors than cap everyone. I think she's right about the cost and arXiv is right about the constraint. A detector needs moderators too. ## What does it say for anyone publishing with AI? That output is not the bottleneck any more, the checker is. It applies inside a team as much as on arXiv. When a tool lets one engineer open five times as many pull requests, the reviewers are the arXiv moderators. A cap on submissions per person is the crude fix. Making the author do more of the checking before submitting is the better one, and it's what the rejected-papers-count rule is quietly pushing people toward. It connects to [the OpenAI contractors piece](https://www.dailyferment.com/openai-contractors-fired-using-ai-rlhf-human-feedback/), which was about what happens when the human checkers themselves start using AI. ## What I'm confident of, and what I'm not The numbers and the rules are established from arXiv's post. That the cap rations moderator time is arXiv's own framing. That a flat quota charges careful and careless authors the same is my inference, and whether it improves quality is a guess until arXiv publishes what happened. The claim, in one sentence: when writing gets cheap, the scarce thing is the person who checks it, and a flat per-author quota is the bluntest way to price their time: it stops the flood, and charges the careful author the same as the careless one. ## Sources Kat Boboris, arXiv, "Fair Moderation, Equitable Access, and AI: arXiv's Updated Rate Limit Policy", 1 October 2026\. [https://blog.arxiv.org/2026/10/01/updated-rate-limit-policy/](https://blog.arxiv.org/2026/10/01/updated-rate-limit-policy/?ref=dailyferment.com) arXiv, Content moderation. [https://info.arxiv.org/help/moderation/index.html](https://info.arxiv.org/help/moderation/index.html?ref=dailyferment.com) arXiv, review articles and position papers in CS, 31 October 2025\. [https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/](https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/?ref=dailyferment.com) Jose Antonio Lanz, Decrypt, 3 November 2025\. [https://decrypt.co/347196/arxiv-blocks-ai-generated-survey-papers-flood-trashy-submissions](https://decrypt.co/347196/arxiv-blocks-ai-generated-survey-papers-flood-trashy-submissions?ref=dailyferment.com) Alina Maria Stan, The Next Web, 17 May 2026\. [https://thenextweb.com/news/arxiv-ai-slop-ban-researchers-preprint](https://thenextweb.com/news/arxiv-ai-slop-ban-researchers-preprint?ref=dailyferment.com) Prachee Avasthi, 2 October 2026\. Cynthia Dwork and Moni Naor, "Pricing via Processing or Combatting Junk Mail", CRYPTO 1992\. [https://web.cs.dal.ca/\~abrodsky/7301/readings/DwNa93.pdf](https://web.cs.dal.ca/~abrodsky/7301/readings/DwNa93.pdf?ref=dailyferment.com) Adam Back, "Hashcash: A Denial of Service Counter-Measure", 2002\. [https://www.hashcash.org/papers/hashcash.pdf](https://www.hashcash.org/papers/hashcash.pdf?ref=dailyferment.com)