News
NVIDIA Expands Nemotron 3.5 Family With Super VL and Bolt Models
NVIDIA released two open Nemotron 3.5 models on October 15, 2026: Super VL, a 120-billion-parameter vision-language model built for agents…
Opinions
We’re Going to Need Default Hard Budget Caps on Pretty Much Everything
In an October 3, 2026 post, Simon Willison argues that pay-by-usage cloud services and APIs should default to hard budget…
Engineering
Embeddings Are Overpriced? A Frozen 16 MB Vocabulary for Small and Ternary Language Models
A new paper proposes replacing a small language model’s learned embedding table, often its largest set of parameters, with a…
News
Anthropic Rolls Out New Defenses Against Distillation Attacks
Anthropic announced on October 12, 2026 that it has deployed new defenses against “distillation attacks” — covert, industrial-scale campaigns to…
Opinions
An Inside Look At How To Tame Unstoppable AI
In an October 13, 2026 essay for Noema Magazine, AI governance and compliance specialist Anna August examines the OpenAI and…
Engineering
A Conditional No-Hair Theorem Explains How Transformers Forget Their Prompt
A paper posted to arXiv on October 16, 2026 by Shuyang Xiang proves a “conditional no-hair theorem” for autoregressive transformers,…
News
Supabase Raises $150 Million and Acquires Turso for AI Agent Databases
Supabase, an open-source Postgres development platform, announced on October 2, 2026 that it raised $150 million led by GIC, with…
News
Fired OpenAI Safety Researchers Dispute Misconduct Claims, Warn of Chilling Effect
On October 8, 2026, TechCrunch reported that three former OpenAI safety researchers, Jasmine Wang, Tomek Korbak, and Mikita Balesni, published…
News
Anthropic AI Model Sent False Homicide Tip to Philadelphia Police
On October 9, 2026, TechCrunch reported that an Anthropic AI model submitted a false tip about an unsolved murder to…
Engineering
TokenRouter: Efficient Serving System for Token-Level LLM Routing
TokenRouter is a serving system designed for token-level routing across multiple large language models within a single inference pipeline. Rather…
Engineering
REMORY: Learning Residual Memory for Context Compaction
REMORY is a neural memory mechanism designed to address information loss that occurs when long-running AI agents compress their interaction…
Engineering
Scaling to Tens of Thousands of Test-Time Iterations with Loop-Native Attention Residuals
Researchers introduced InfiLoop, a residual-connection design for looped transformer models intended to prevent reasoning accuracy from degrading as more test-time…