Tech News
Hacking OpenAI
A security team used an autonomous Claude agent in a /goal loop to achieve remote code execution (RCE) on a Discourse Cloud instance, after first confirming local RCE via an image upload.
How To Write With An LLM
Thomas Ptacek on using LLMs as copyeditors, not as writing assistants: Rule Number One: You may not use a single word an LLM suggests to you.
A shared agentic platform for Wood Mackenzie, on Amazon Bedrock AgentCore
Wood Mackenzie built APEX, a shared agentic AI platform on Amazon Bedrock AgentCore so every team can ship production agents without rebuilding runtime, identity, observability, and guardrails from scratch.
openai/openai-python v3.15.0
3.15.0 Features api: add agent session model settings.
TensorRT Edge-LLM Completes the MLPerf Edge Agentic Benchmark 6.4x Faster on Jetson AGX Thor
AI agents are moving from cloud data centers to vehicles, robots, and other edge devices. Unlike a chatbot that answers a single prompt, an agent works through.
GitHub Repos
NanmiCoder/cc-haha
Local-first cross-platform desktop workspace for Claude Code / agents: multi-agent, Git worktrees, code diffs, skill marketplace, multi-model, Computer Use, task-aware desktop...
# TypeScript# agent# aicode-yeongyu/oh-my-openagent
OmO: Just type "mass ulw" keyword with your prompt.
# TypeScript# ai# ai agentsEverMind-AI/EverOS
One portable memory layer for every AI agent: local-first, Markdown-native, user-owned, and self-evolving across apps, tools, and workflows.
# Python# agent memory# agentic aiResearch Papers
CoRef-GS: Cooperative Referring Gaussian Splatting for Multi-Agent Scene Understanding
This paper builds a system that lets multiple robots share 3D Gaussian maps and answer language questions like 'the chair next to the table' from a specific robot's viewpoint. It also introduces a new benchmark with two quadruped robots and shows better accuracy than a prior...
When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows
This paper replicates and unifies two edit-based generative models for antibody design, showing they are the same process. It releases the first open-source implementation and a generalist antibody editor, while highlighting reproducibility issues like hidden hyperparameters...
JEPA-Anything: Learning Predictive Models across Different Worlds
World modeling enables intelligence to anticipate consequences, guide interventions, and learn from interaction. Yet predictive models remain domain-specific: can a common learning principle support world modeling across radically different systems?