All published AI news and analysis.

Endava is leveraging AI agents, ChatGPT Enterprise, and Codex to enhance software delivery processes. The company aims to automate workflows and foster an AI-native culture within its enterprise.

The `awesome-ai-agents-2026` GitHub repository provides a curated list of AI agent frameworks, tools, platforms, and resources. It aims to be a definitive guide for the year 2026, which the repository's description labels as the year AI agents went mainstream.

A new research paper investigates the use of transformer model activations for selecting in-context examples in large language models. The study, which includes a comprehensive analysis using Llama-3.2-3B and Qwen2.5-3B, found that MLP outputs based on massive activations or statistical moments do not correlate with example quality or task performance, suggesting that this approach is not effective for in-context learning.

Researchers introduce GeM-NR, a novel training-free method for multi-view image editing that supports significant changes to scene geometry and appearance. The approach leverages depth estimation and multi-reference editing models to ensure consistency across views.

OpenContracts introduces a programmable citation graph for documents, enabling humans and AI agents to collaborate. The platform offers a self-hosted, MIT-licensed solution for structured extraction, AI agent integration via Model Context Protocol (MCP), and a GraphQL/REST API.

Researchers introduce DistIL, a distributional variant of the DAgger imitation learning algorithm designed to leverage rich feedback beyond simple binary rewards for training reasoning models. This new approach offers theoretical guarantees for monotonic policy improvement and regret, outperforming existing methods in scientific reasoning, coding, and mathematical problem-solving tasks.

A new text generation model, `rupeshpanda/gita-text-generation-gpt2`, has been identified on Hugging Face. Built on the GPT-2 architecture and utilizing the Transformers library, this model is designed for text generation tasks.

Wasmer leveraged OpenAI's Codex and GPT-5.5 to develop a Node.js runtime for edge computing. This integration reportedly accelerated development by 10x to 20x, enabling project completion in weeks rather than months.

Thunderbolt is an open-source, cross-platform AI client designed for on-premises deployment, allowing users to choose their models and own their data. Currently in active development and targeting enterprise customers, it supports local and on-prem models and is available on major desktop and mobile platforms.

OpenAI has introduced new capabilities to GPT-Rosalind, an AI model designed for life sciences research. The enhancements focus on improving biological reasoning, medicinal chemistry expertise, genomics analysis, and experimental workflow design.