From the archive
Friday, 16 October
Folded from 14 sources that day.
Sorted by quality and by how many sources covered it, the same for everyone.
-
Beam: a 501B open-weight MoE model for coding and agents
Reflection's first model, Beam, is a 501B-parameter mixture-of-experts model with 23B active per token, a 1M-token context and Apache 2.0 licensing. It has day-0 support in Transformers, vLLM and SGLang, and Reflection offers $10,000 in Hub credits for projects.
-
The NIST Elliptic Curves Seeds Bounty Has Been Claimed
An Anthropic team solved the decades-old mystery of the NIST curve seeds: they came from the POSIX drand48 generator seeded with the time in 1997–98. The $36,864 bounty went to ISRG.
-
Ontology-Driven Observability: Building an End-to-End Knowledge Graph at Netflix Scale
Netflix explains how a knowledge graph over its MELT observability data cut a typical investigation from about four hours by making system relationships machine-readable.
-
Improving navigation experience in v0 with Next.js 16.4
Vercel explains how Partial Prefetching in Next.js 16.4 and a new optimizer Skill made v0 navigations feel like a single-page app, with cached content ready before the click and on return visits.
-
What agents mean for Rust (and vice versa)
A programming language designer argues that Rust fits agentic coding well and explores why. The post also covers where Rust could build on those strengths and why language design still matters.