Tech News
Pi coding agent: config folder is out of place on Linux
A GitHub issue about Pi coding agent's non-compliant Linux config folder placement sparks debate, with users citing it as a reason to explore alternatives despite its performance strengths.
NVIDIA Nemotron 3.5 Lightning now available in Amazon SageMaker JumpStart
NVIDIA Nemotron 3.5 Lightning, an open model built for high-volume agentic workloads, is now available in Amazon SageMaker JumpStart.
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things
Friday's big release was Qwen 3.8 27B, an Apache 2 licensed 27B parameter vision-capable LLM from Alibaba's Qwen research lab.
AI-Generated GitHub Copilot “Autofix” Allowed Compromise of Snowflake's Jira
A security researcher exploited a GitHub Copilot autofix suggestion to compromise Snowflake's Jira, highlighting risks in AI-generated CI/CD code.
Launch HN: Speko (YC S26) – OpenRouter for Voice AI
Speko launches as an OpenRouter-like platform for voice AI, dynamically selecting optimal STT, LLM, and TTS model combinations based on user constraints and benchmarks.
GitHub Repos
open-multi-agent/open-multi-agent
TypeScript AI agent orchestration framework with dynamic workflows.
# TypeScript# agent framework# agent orchestrationOpenHands/OpenHands
🙌 OpenHands: AI-Driven Development
# TypeScript# agent# artificial intelligencego-kratos/blades
Blades is a Go-based multimodal AI Agent framework.
# Go# agent# agent frameworkResearch Papers
Large Models for Small Devices: Recent Advances and Empirical Analysis of Edge AI Deployment
This paper surveys and tests various model compression techniques (like pruning and quantization) for running AI models on small devices, finding that no single method works best and that compression can sometimes hurt performance or even make models look better than they are.
ATLAS: Scaffold-Free Algorithm Synthesis by LLMs via Embedding-Guided Quality-Diversity Search
ATLAS is a new method that uses large language models and quality-diversity search to automatically design complete algorithms for combinatorial optimization problems without needing a predefined scaffold. It shows promising results on four NP-hard problems, producing diverse...
A Declarative-Procedural Perspective on Expert Routing in Bilingual Mixture-of-Experts Language Models
This paper explores whether Mixture-of-Experts (MoE) language models, which use specialized sub-networks, develop linguistically structured routing when learning two languages. The authors find that a model trained on mixed English-German data shows stronger specialization...