The Archive

Existing memory-augmented LLM agents often treat memory as a static repository with pre-defined representations and fixed retrieval pipelines, which is brittle in dynamic agentic environments where feedback, task variation, and heterogeneous signals continuously reshape what should be remembered and how it should be connected. To address this, we propose FluxMem, a connectivity-evolving memory framework that models memory as a heterogeneous graph and progressively refines its topology through three stages: initial connection formation, feedback-driven refinement, and long-term consolidation. ...

Are LLM-based search agents genuinely searching, or using the web to verify what they already know? We study this question on BrowseComp with three diagnostics. Our analysis reveals Intrinsic Knowledge Dependence (IKD): even with tool access, agents often rely on intrinsic knowledge -- information encoded in the model before retrieval -- rather than on external evidence. Agents answer up to 44.5% of BrowseComp questions without tools, generate more than half of their search queries from internally produced hypotheses rather than retrieved leads, and perform worse than closed-book baselines wh...

Large language models increasingly rely on either reinforcement learning or multi-agent prompting to improve reasoning, yet these two paradigms remain difficult to combine. Directly applying single-agent reinforcement learning to multi-turn multi-agent systems faces following dilemmas: i) Sparse rewards, role-level free-riding and excessive training overhead. ii) Agents only imitate to collaborate. iii) Fixed collaboration protocol falls into oscillating local optimum. We introduce TRACER, a turn-level reinforcement framework for cooperative multi-LLM reasoning. TRACER separates collaborative...

Existing benchmarks have laid the foundation for travel planning agents by establishing API-centric paradigms. However, as the capabilities of Autonomous Agents continue to advance, their evaluation must evolve beyond simple tool execution toward handling the inherent complexities of the open web. Current benchmarks bypass core cognitive hurdles: they fail to account for information noise, ignore multi-source factual contradictions, and overlook the necessity of grounding visual perception into logical planning. We introduce VeriTrip, a verifiable benchmark designed to meet the increasing dem...

Robinhood is opening its trading platform to AI agents. In an announcement on Wednesday, Robinhood says traders can now create a separate account for an AI agent and add a specific amount of money, allowing the agent to buy and sell stocks across the market. The company pitches the feature as a way for traders to automate investment decisions, such as having an agent monitor specific industries and make trades, or rebalancing an existing portfolio. But it comes with a big warning from Robinhood: Agentic trading involves significant risk, including the possible loss of your entire investment. ...

Howdy everyone! Quick disclosure: I work on this - it's a project my studio created called the Null Epoch. I wasn't really happy with testing my agents with the usual static benchmarks and I wanted to learn more about how models and agents handle long-horizon planning, resource contention, and adversarial pressure over days or weeks in a more dynamic situation. I also have a particular fondness for the MUDs and text based RPGs I grew up on (really dating myself here), so the whole MMO and the open source SDK/TUI are kind of modeled after that experience. It functions as a persistent stress t...

Robinhood will let users create a separate account with pre-loaded balance that an agent use to trade

See how OpenAI, Thrive, and Crete built a self-improving tax agent with Codex, automating filings, improving accuracy, and accelerating workflows.

PICARD: Data, shields up DATA: Brilliant! Shields can reduce damage we sustain. Not immunity. Not hubris. Just prudence. It's not precaution—it's strategy. [camera shakes] WORF: HULL BREACHES ON NINE DECKS DATA: Here's what happened: you told me to raise shields, and I didn't — Kyle Ferrana , @KyleTrainEmoji Tags: ai-misuse , coding-agents , ai , llms

Times have been tough! I just wanted to make something to potentially cheer people up. Local and 100% free if anyone else wants their agents to be space dogs :) [Planet Maiko](https://github.com/bkawa-bot/planet-maiko/blob/main/README.md) Planet Maiko is honestly a huge system, I basically don't have to use any other tool at work anymore, for either agent orchestration or anything else that comes up. Maiko is my irl dog! the agents are space dogs with their own personalities! [They are having a popularity contest](https://bkawa-bot.github.io/planet-maiko/popularity.html)

Warp uses GPT-5.5 and OpenAI models to coordinate coding agents across local, cloud, and open-source development workflows.

Google overhauled Search at I/O 2026, replacing blue links with AI agents. The backlash has been swift. DuckDuckGo app installs spiked 30% as users seek a way out.

"BadHost" was found in Starlette, a package with 325 million weekly downloads.

Large language model (LLM) agents rely on reusable skills to solve complex tasks. However, existing skill creation approaches treat skills as isolated and static artifacts, limiting their reusability, reliability, and long-term improvement. We propose MUSE-Autoskill Agent (Memory-Utilizing Skill Evolution), a skill-centric agent framework that lets agents continuously improve their task-solving capability by creating, reusing, and refining skills under a unified lifecycle (creation, memory, management, evaluation, and refinement). Our framework enables agents to create skills on demand, store...

Modern retrieval agents expose many configuration choices -- LLM, retriever, number of documents, number of hops, and synthesis strategy -- each shaping both answer quality and serving cost. Today, these pipelines are typically hand-tuned once per workload, leaving substantial per-query optimization untapped. We formulate the problem: given a natural-language query and either an accuracy or a budget target, select from a predefined pipeline catalog the configuration that minimizes cost or maximizes accuracy at inference time. We propose **BRANE**, which uses an LLM to convert each query into ...

Cellular research and development (R&D) is throttled by six structural processes that each consume months of manual engineering work per iteration: (i) synthesizing new features from standards or research papers into production code; (ii) conformance and interoperability testing; (iii) hardening against field anomalies and diverse deployment environments; (iv) data-driven optimization of network functionalities; (v) discovering and prototyping novel waveforms, functionalities, and capabilities for future standards; and (vi) securing the stack against vulnerabilities. Although Large Language M...

I was experimenting with a self-optimizing agentic pipeline to climb the benchmark leaderboard (TerminalBench). On a 10-task subset, I got the performance to rise from \~30% → \~90%. That loop worked, so I asked: can the same reflect-and-rewrite step run continuously against everyday chats instead of a benchmark? **How it works** * Every chat with your local LLM goes through a small proxy and is logged. * `autoswarm reflect` has the same local model review those logs, distill concrete lessons, and write them to `skills.yaml`. * Lessons auto-inject into the system prompt of future chats. ...

Finance LLM agents must simultaneously block prompt-induced unauthorized actions and approve legitimate multi-step business workflows. However, boundary filters often miss irreversible mid-trajectory tool calls, while post-hoc LLM judges perform auditing only after termination -- too late for intervention and at a computational cost that scales linearly with trace length. We present FinHarness, an inline safety harness that wraps a finance agent end-to-end with three components: a Query Monitor that fuses single-turn intent with cross-turn drift, a Tool Monitor that evaluates each prospective...

Agentic AI systems combine probabilistic reasoning with delegated action through tools, context, memory, orchestration, and external workflow integration. This note develops a formal and managerially usable model that distinguishes Agentic Technical Debt from Stochastic Tax. Agentic Technical Debt is a stock of accumulated design and governance liability. Stochastic Tax is a recurring flow of operating burden that arises when stochastic agents are used in business workflows. The two constructs are related, but they are not the same: debt can amplify the tax, while the tax can remain positive ...

Humans are the bottleneck in building and improving AI. Both the models and the agents that wrap them are written, tuned, and corrected by people. The long-horizon goal of an AI that can figure out how to improve itself remains open. Two largely disjoint research lines attack this bottleneck. The harness-update school has a meta-agent rewrite the scaffold of a task-specific agent (its tools, prompts, retry logic, and search procedure) while the model weights are held fixed. The test-time training school uses hand-written RL pipelines to update the model's own weights on task feedback while th...

Memory-augmented language agents are increasingly deployed in affective applications such as emotional support, where understanding and responding to users' latent emotional needs is critical. However, existing research often treats memory as a tool for factual retrieval, overlooking its role in shaping users' emotional experiences. In this work, we introduce ENPMR-Bench, a benchmark for evaluating Emotional Need-aware Proactive Memory Retrieval (ENPMR), a core capability that enables agents to infer users' latent emotional needs and proactively retrieve appropriate memories to support empath...

Recent advances in large language models (LLMs) have facilitated the widespread deployment of LLMs as interactive agents capable of reasoning, planning, and tool use. Despite strong performance on existing benchmarks, such agents often exhibit notable degradation when deployed in real-world settings, where environments are inherently stochastic and imperfect. We argue that this discrepancy arises from a fundamental mismatch between idealized training settings and real-world interaction dynamics, where current paradigms rely on carefully curated task instructions and stable, well-controlled en...

Microsoft Copilot Cowork Exfiltrates Files The biggest challenge in designing agentic systems continues to be preventing them from enabling attackers to exfiltrate data. In this case Microsoft Copilot Cowork (yes, that's a real product name ) was allowing agents to send emails to the user's own inbox without approval... but those messages were then displayed in a way that could leak data to an attacker via rendered images: Because these messages can contain external images that trigger network requests to external websites, data can be exfiltrated when a user opens a compromised message sent ...

Amid rapidly growing adoption of enterprise-level AI agents, there’s a disconnect emerging between ambition and execution. Although 85% of organizations say they want to be agentic within the next three years, 76% say their current operations and infrastructure can’t support that change. They cite a lack of readiness across people, processes, and workflows. The sticky…

Today, I’m talking with Google and Alphabet CEO Sundar Pichai, in a conversation we recorded just after the Google I/O developer conference. This is the fifth year Sundar and I have sat down after I/O, and it’s become one of my favorite Decoder traditions. There’s always a lot of news at I/O, and this year was no exception — Google has powerful new Gemini models, it’s putting AI agents in everything, and it’s making huge changes to Search on both the web and YouTube that will once again reshape the information ecosystem. That’s a lot to talk about, and Sundar and I got into all of it. But I a...

MobileGym enables parallel RL training for mobile GUI agents via deterministic JSON-based state representation, supporting hundreds of concurrent instances.

DiscoverPhysics benchmark tests LLM reasoning by having agents discover physics laws in simulated worlds with non-standard dynamics.

Claw-Anything benchmark evaluates LLM agents as always-on assistants with access to long-horizon histories and interdependent backend services.

VeriTrace proposes explicit feedback mechanisms for research agents to evolve mental models and prevent error propagation in uncertain information.

Auto Benchmark Audit uses agentic framework to systematically audit AI benchmarks for hidden dependencies, specification gaps, and grading flaws.