ML research on arXiv tops 100 daily submissions

The sheer volume of ML research has transformed arXiv from a discovery engine into a signal-to-noise nightmare.

• –Annual submissions in cs.AI and cs.LG grew by over 30% year-over-year, driven by the generative AI gold rush.
• –The doubling rate of AI research papers is now roughly 24 months, significantly outpacing human reading capacity.
• –Researchers are increasingly reliant on social filtering via X/Twitter and newsletters to identify high-impact work.
• –Influx of AI-generated content and incremental "wrapper" papers is diluting research quality and straining peer review systems.
• –Recent shifts in arXiv policies and its 2026 transition to an independent nonprofit reflect the massive infrastructure burden of this scaling.