Weekly

Frontier AI Research

Papers, architectures, and the ideas that will define the next wave.

For readers who check arxiv before Twitter. New architectures, benchmark-breaking results, mechanistic interpretability, and the research that mainstream coverage misses. Technical, specific, and assumes you know what a loss function is.

Latest Issue

Frontier AI: Week of March 31 -- April 7, 2026

Neuro-symbolic efficiency claims hit the headlines, Mamba-3 proved SSMs belong at scale, and Google's KV cache trick might actually change your inference bill.The past seven days delivered a rare combination: a genuinely surprising efficiency result, a mature architecture paper that closes the SSM-v...

Read This Issue April 7, 2026

What we cover

  • Key arxiv papers each week
  • Architecture innovations
  • Scaling laws & compute research
  • Mechanistic interpretability
  • Benchmark results & analysis

Who it's for

ML researchers, PhD students, and senior engineers who read papers.