Why this exists

A calmer way to follow a fast field

The Latest in AI helps researchers and engineers keep up with a research landscape that changes faster than any static reading list can.

A living map, not a folder tree

Papers are embedded and grouped into stable topics. The topic names and membership can evolve as new work arrives, while the stable identity means a follow survives a rename or merge.

Built for unfinished data

The ingestion pipeline is still growing. Affiliations, figures, references, and generated summaries appear only when the pipeline has real data for them. The interface does not invent substitutes.

How research answers are grounded

The research agent starts with indexed metadata and abstracts, and can inspect available arXiv full-text passages and extracted figures when a question needs more evidence. Paper-specific claims link back to the cited records so you can inspect the original source.

What comes next

The research agent, followed feed, and saved library all build on the same living index, so every answer and recommendation can lead back to real papers.