Agents & autonomy

The rise of Software Engineering (SE) agents, i.e., LLM-based agents that can understand large codebases and carry out engineering tasks with limited human intervention, has been marked by rapid advances and adoption, but little is known about how developers build these systems in practice: existing studies mine repositories or examine deployment, but few investigate how SE agents are constructed. Through semi-structured interviews with 20 practitioners from 12 organizations and an online survey

Multi-agent ensembling multiplies active parameters and inference cost without answering three basic questions: which agents to consult, how deeply a query should traverse a hierarchy of agents, and when inter-agent communication is worth its cost. We present GRADE (Gated Routing and Adaptive Depth for Efficient Reasoning), a hierarchical multi-agent system in which four lightweight learned gates jointly govern agent selection, hierarchy depth, inter-agent communication, and branch pruning. Trai

Nia Jetter spent two decades building spacecraft and robots. Now she is trying to make sure the artificial intelligence revolution doesn't leave anyone behind.

AI engineering is shifting from passive text generation by large language models (LLMs) to agent-driven task execution, creating new reliability challenges for long-horizon tasks under resource constraints and environmental uncertainty. Conventional error-elimination optimization strategies fail to address cumulative error propagation. This paper proposes Distributed Agent System (DAS), a device-edge-cloud framework for fault-tolerant collaboration among heterogeneous agents. We redefine agent r

Recent advances in LLMs and existing work on programming by demonstration have made it possible for end users to create automations by explicitly demonstrating their behavior to LLMs. However, these approaches rely on the assumption that users know what to automate and what is capable of being automated. Additionally, automation via LLM agents is often expensive compared with programs. We introduce Motif, a system that passively observes everyday browser activity to discover recurring interactio

Stateful personal agents increasingly maintain long-term user profiles, episodic memories, and reusable skills. This persistence turns conversational sycophancy into a state-writing failure: accepted user-centric claims can be committed as lasting preferences, background facts, or workflows and later reused after the original conversation is gone. We call this persistent sycophancy and introduce the Personal Agent Sycophancy Benchmark (PASB), a 1,600-task benchmark that traces whether a conversa

Large language model (LLM) agents are beginning to automate machine learning engineering (MLE) by coupling planning, code execution, debugging, and empirical feedback. Translating this capability to medical imaging remains difficult because each task imposes modality-specific experimentation and strict requirements for validation protocols and prediction artifacts. Here we introduce AMID, an autonomous multi-agent framework for medical imaging model development. AMID first proposes Data-Conditio

Autonomous CLI agents can now execute hundreds of actions across multi-hour sessions: writing code, executing shell commands, browsing the web, and managing cloud infrastructure, all with minimal human oversight. Does greater autonomy invite greater risk? We introduce ANCHOR, an automated auditing framework that stress-tests CLI agents on illegal tasks grounded in public US court cases. ANCHOR deploys an auditor agent fine-tuned on dark personality data using supervised and reinforcement fine tu

A merge is a contract. The moment a change lands on main, every other team in the organization starts building The post Your merge gate was a compromise. Coding agents are making it a liability. appeared first on The New Stack .

Safe and socially compliant navigation remains a fundamental challenge for autonomous robots operating in human-populated environments. Beyond collision avoidance, robots must anticipate human motion and respect personal space to ensure human comfort. Model Predictive Control (MPC) offers a robust alternative to classical and data-driven methods, although its effectiveness strongly depends on accurate human motion prediction and efficient computation. This paper introduces SFM-NMPC, a Social For

Go has emerged as the lingua franca for cloud infrastructure, used for everything from container orchestration and CI/CD pipelines to The post Microsoft joins Google in backing Go for AI agents — OpenAI and Anthropic lag appeared first on The New Stack .

Recent VLM and VLA systems have improved robotic perception and action prediction, yet long-horizon embodied agents still require a general runtime layer for reasoning, memory, tool use, verification, and cross-embodiment execution. We present ABot-AgentOS, a general robotic Agent Operating System that sits above low-level controllers and provides a deliberative agent layer for scene-conditioned planning, context-isolated skill execution, multi-stage verification, multi-modal memory, and edge-cl

GENEVA, July 10, 2026 - The AI for Good Global Summit concluded following four days of sessions that demonstrated the widespread application of artificial intelligence in a variety of sectors, including health and agriculture, mobility, education, creative industries, and public policy. The summit also raised awareness of quantum technologies, robotics, and the standards and infrastructure required to facilitate their widespread deployment. The post Day 4 at the AI for Good Global Summit: Women’

Scientific research is a complex, multi-stage workflow rather than a single act of text generation. The ideation process typically emerges through literature search, paper reading, tool use, claim checking, cross-paper synthesis, brainstorming, rejection of weak directions, and iterative writing. Existing resources capture individual components of this process, but datasets that jointly record tool use, evidence acquisition, intermediate artifact evolution, and idea- or proposal-level endpoints

Large language model agents increasingly store reusable procedures outside the model. These reusable procedures are often called \emph{skills}: they may be code functions, natural-language instructions, SKILL.md packages, workflow graphs, or learned adapters that a future agent can retrieve and invoke. This taxonomy-driven survey asks how such skill libraries change over time. Across a $124$-paper $2023$--$2026$ audit set, we synthesize dynamic skill systems as \emph{lifecycle-managed, verified,

Agentic retrieval-augmented generation (RAG) extends static RAG by allowing language models to iteratively reason, generate search queries, retrieve evidence, and predict answers. However, it remains challenging for models to decide when to retrieve, whether to use lexical matching or semantic similarity, and how to control context granularity to prevent irrelevant tokens from interfering with agent reasoning. In this paper, we introduce GRASP, a reinforcement learning (RL) framework for trainin

HydroShear, a new physics-based simulator, teaches robots how to use their sense of touch to perform complex manipulation tasks, in a way that transfers seamlessly to the real world.

The results were surprisingly good—with a few important caveats.

Planning summer travel is hard. Planning it for a busy family —visiting multiple destinations with three little kids in tow—can feel downright impossible. I’m a professional news and travel photographer, so I need to plan a lot of trips with my wife and our three boys. I’m also an AI expert with a cog sci degree from Johns Hopkins and a decade of experience in the field. So naturally, I decided to combine those two passions. To that end, I built an AI digital twin of my family, which uses multip

The results were surprisingly good—with a few important caveats.

The results were surprisingly good—with a few important caveats.

The results were surprisingly good—with a few important caveats.

The post The Robotics for Good Youth Challenge 2026 comes to an end as winners are awarded appeared first on AI for Good .

The proliferation of agentic AI systems across enterprise and public-sector contexts has outpaced the capacity of general-purpose AI risk frameworks to classify and govern them. In this paper, we introduce the TrustX Agent Risk Classification Framework, a structured, repeatable instrument that can be applied to seven types of agentic AI systems and is grounded in foundational pre-existing AI governance frameworks. At the core of the framework is a twelve-dimension scoring rubric that robustly qu

Agentic systems usually have two jobs: Build context, then use that context to produce an answer or action. Many failures The post Why retrieval quality is becoming the defining challenge in AI agent architecture appeared first on The New Stack .

Cloudflare has recently introduced temporary accounts that let AI agents deploy Cloudflare Workers immediately, without first creating or authenticating with a permanent account. If left unclaimed, the accounts and their deployments expire automatically after 60 minutes. By Renato Losio

OpenAI announced ChatGPT Work on July 9 and began rolling it out to Pro, Enterprise, and Edu users. It runs The post OpenAI, Microsoft & Anthropic agree on who runs the agent. They disagree on what you can take back. appeared first on The New Stack .

Agentic testing is an AI-driven approach to end-to-end test automation introduced by Slack engineering. It uses AI agents that execute workflows based on intent rather than fixed scripts, adapting to UI and system changes at runtime. The approach aims to reduce brittle tests in distributed systems while complementing deterministic unit, integration, and E2E testing strategies. By Leela Kumili

Artificial intelligence has spent years learning to generate words, images and videos. Now, the industry’s attention is shifting towards “world models” – AI systems designed to simulate how the physical or digital world changes over time. While the term was originally associated with physics simulations and robotics, a growing number of Chinese companies are applying the concept more broadly. The technology remains in its infancy, with little consensus on what a world model should ultimately...

The Robotics for Good Youth Challenge Grand Finale will reach its conclusion on the Youth Stage from 14:10 to 14:45. After four days of competition, the challenge will come to an end, closing a season that opened the Youth Zone's programme on 7 July. The post Top 3 things happening on Day 4 of the AI for Good Global Summit 2026 appeared first on AI for Good .

Today's links "Rights for robots" and the AI slavery fantasy: When we were robots in Egypt… Hey look at this: Delights to delectate. Object permanence: Awkward questions for the entertainment industry; ISP v record label death penalty; Coming out on Splash Mountain; Negative Swiss bond yields; Are We Having Fun Yet? Upcoming appearances: London, Edinburgh, Sydney, Melbourne, Brighton, London, South Bend. Recent appearances: Where I've been. Latest books: You keep readin' em, I'll keep writin' 'e

Recognition tells an agent what is in an image; personal memory affects what is worth looking up next. In a camera-first setting the user can send only an image, so the agent must form the lookups. We study whether personal visual memory improves agent-side tool choice and tool arguments, and thereby more user-aligned multi-tool lookups. The design uses a three-layer personal visual memory (profile, short-term focus, observations) that is loaded on each turn to condition an LLM tool-calling loop

The organisations that will lead in AI will be those that identify their power law use cases fastest − and scale them before their competitors do.

Social Media + Society, Volume 12, Issue 3, July-September 2026. This article examines the role the performance-tracking app Strava plays in facilitating professional careers and engagement in a growing endurance sports industry. Through qualitative interviews with 25 elite endurance athletes and their agents, I find ...

Amid live coding sessions and Silicon Valley optimism, the UN’s AI for Good summit wrestled with an urgent question: Can global governance catch up before the technology races beyond its control?

EAP's exports now face higher trade restrictions, elevated policy uncertainty, and slowing global growth. And its workers must cope with the growing use of robots, AI, and digital platforms. Download ...

Insta360 marked its 11th anniversary on Thursday by unveiling its vision for a Cameraman photography robot. According to the company, the Cameraman is not a single hardware product but an AI agent concept capable of autonomous filming, with panoramic drones serving as one of its early prototypes. Insta360 said it has already developed key technologies […]

Grace Investment Machine (GIM) has raised $20 million in Series A funding to build agentic systems for capital markets.

IntroductionLarge language models (LLMs) can generate persuasive narratives that shift agent behavior in multi-agent systems, but deploying raw, unstructured text as an influence mechanism offers no formal guarantees on effectiveness, interpretability, or controllability.MethodsWe introduce the LLM Influence Compiler (LIC), a solver–critic pipeline that compiles natural-language cooperation directives into structured, parameterized influence policies defined over a five-field schema: network tar