KeepSanity AI

Week 2, May 2026

business

Claude & SpaceX: Higher usage limits and a compute deal

product updates

Cloudflare: Agents can now create accounts, buy domains, and deploy

models

Gemma 4: Faster inference with multi-token prediction drafters

top-rated papers

MolmoAct2: Action Reasoning Models for Real-world Deployment

From Context to Skills: Can Language Models Learn from Context Skillfully?

Stream-R1: Reliability-Perplexity Aware Reward Distillation for Streaming Video Generation

tools

DeepSeek 4 Flash: Local inference engine for Metal

resources

OpenAI Voice AI: Delivering low-latency at scale

LLMs from Scratch: Train your own model

community

M4 with 24GB Memory: Running local models

OpenAI: The WebRTC problem

Computer Use: 45x more expensive than structured APIs

Agent Skills: An overview