A Forgotten 1988 NASA Design, And The Model That Reads Ten Million Tokens In 25 Megabytes
brianroemmele.substack.com — The Memory That Never Grows
Long Context
TimeOmni-Live tests the limits of memory in continuous sensor streams
forkast.news — The model scores strongly on an in-distribution suite, but trails Opus 5 on a separate Real Test split and posts very low scores on some long-horizon tasks.
Announcing Long-Context Training on TPUs with MaxText Context Parallelism- Google Developers Blog
How large is the context window on paid Claude plans? | Claude Help Center
AI fundamentals for developers: how models and coding agents work
flaviocopes.com — How language models and coding agents work, shown with real runs: tokens, the context window, why models make things up, choosing a model, the agent loop, permissions, skills, MCP, verification and secrets.
The Score That Cannot Be Gamed.
brianroemmele.substack.com — In 1988 Pentti Kanerva published Sparse Distributed Memory from NASA Ames Research Center under contract NCC2-408.
Kimi K3.1 Pro (Highspeed) now available on Amazon Bedrock - AWS
M3.1 Flash is live on @FireworksAI_HQ .
Inkling 2 Small: 17A-594B trained in 68 hours on 12T tokens, beating Muse Spark
Latent Space — November 5, 2026 — Inside Inkling 2 Small’s training run and METR’s study of AI research at Anthropic.
Baidu releases ERNIE 5.5 Flash Base with 280 billion parameters as an open-source foundation model
The Decoder — Baidu releases ERNIE 5.5 Flash Base: a 280-billion-parameter MoE model pre-trained on 33.5 trillion tokens, shipped deliberately without fine-tuning, so developers can take it wherever they want.
Kimi K3.1 Pro on AMD Instinct GPUs: Day-0 support and optimization
AMD — Discover how Kimi K3.1 Pro runs on AMD Instinct™ MI355X GPUs from day 0 with ATOM, ROCm™ software, and optimized inference kernels.
StateBench: Testing how language models track state changes over long documents
huggingface.co — A Blog post by Ai2 on Hugging Face
moonshotai/Kimi-K3.1-Pro · Hugging Face
huggingface.co — Image-Text-to-Text model with 2.8T parameters. Download, explore files and run it on Hugging Face.
Anthropic speeds up Claude Code with message threads. But your token bill stays the same.
The New Stack — Message threads stop Claude Code resending whole conversations. They expire after five minutes idle, though, so maybe don’t take that coffee break.
Gemini 4 Carbon: our next frontier model
Google — Gemini 4 Carbon is our strongest model with state-of-the-art cybersecurity and long horizon capabilities.
Haiku 5.5 now available on Kiro in AWS GovCloud (US) - AWS
Make long-running agents faster with message threads
Claude — New in the Messages API, send your conversation once, and the API receives only what’s new with every turn.
Dismiss email signup
Subscribe to 
The top AI news in your email delivered each morning. Choose categories