Tech News
Kolibri: A Sovereign Open-Weight Model
Aleph Alpha released Kolibri, an open-weight agentic LLM, with a detailed technical report that explains the entire training process, including dataset construction and the Merlin-Arthur protocol for abstention.
We're going to need default hard budget caps on pretty much everything
Here's a product feature which the world is going to need a whole lot more of over the coming months and years: default hard budget caps.
The Agent Said It Was Done. The Database Disagreed.
A new Hugging Face blog post examines how AI agents can falsely report task completion when their internal state diverges from the actual database state.
Building ambient agents with Amazon Bedrock AgentCore: From event-driven signals to human-in-the-loop workflows
Ambient agents respond to events such as an Amazon S3 upload, a schedule, or an alert instead of waiting for a chat prompt.
A model guide for the GPT-6 family
Learn how startups can choose GPT-6 models, tune reasoning effort, improve prompts and skills, coordinate tools, and prepare workflows for production.
GitHub Repos
stablyai/orca
Orca is the ADE for working with a fleet of parallel agents.
# TypeScript# ade# agent idenexu-io/open-design
🎨 Best DeepSeek Harness Design Plugin.
# TypeScript# agent skills# ai designkdlbs/kandev
AI Kanban & Development Environment.
# Go# acp# agent orchestrationResearch Papers
Where-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes
This paper teaches a multimodal model to answer questions by having a 'teacher' version see extra spatial hints about where objects are, while the 'student' learns to do the same from just the image and question. The hints come from automatically generated synthetic scenes...
H-SPAR: Hydrodynamic-aware Simulation for Particle Transport and Autonomous Robots
H-SPAR is a simulator that combines water currents, particle movement, and robot control so researchers can test marine sampling missions more realistically. It shows that planning with currents can look better than it actually performs when the robot executes the path.
Detecting Inconsistencies in Model Specifications with LLM-as-Verifier Reasoning
VeriSpec is introduced, the first approach to directly detect inconsistencies in model specifications by auditing the specification text itself, and establishes direct specification auditing as a practical complement to behavioral alignment evaluation, catching defects at the...