What is changing in AI, and what we are building with it. Follow the wider industry, go behind the scenes at Ashlr, or read both.
Latest briefing:
EngineeringThe briefing
AI-generated illustration
Nvidia launched an open-weight 30B mixture-of-experts transcription model that accepts plain-language prompts for context, speaker labels, translation, and audio questions. It supports 40 languages and runs on a single GPU via vLLM or NeMo-Speech.cpp.
EngineeringThe briefing
AI-generated illustration
GitHub released Copilot Fusion, a generally available system that selects and orchestrates models for coding tasks. Enterprise admins can restrict the model pool while developers choose effort levels instead of individual models.
EngineeringThe briefing
AI-generated illustration
OpenAI published coding evaluations for less common languages. Its migration tests ask whether old software’s behaviour survives a rewrite.
EngineeringThe briefing
AI-generated illustration
OpenAI is ending API access to GPT-4o, GPT-4o mini, GPT-4.1 and related models on 14 January 2027. The three-month window and behavioral differences in the recommended GPT-6 Luna replacement require code and agent workflow changes.
EngineeringThe briefing
Alphabet's Intrinsic released ROS-compatible control, motion and grasp planning under Apache 2.0 at ROSCon 2026. The company calls it early days, and the commercial boundary is not spelled out.
Your week, in perspective.
A weekly email with the developments worth understanding, what they mean for your work, and a practical next step.
Built by Ashlr, a team working on the systems and workflows behind these stories.
Original sources, clear publication dates, and an explanation of what is reporting and what is our interpretation.
Our editorial approachExplore our engineering tools, practical guides, and longer essays from Ashlr Insights.
