The Archive

a quiet day lets us find some connections among recent themes

Moonlight & Mayhem (Raccoon Heist by Codex + GPT-5.6 Sol Ultra) On Wednesday I wrote about One-shotting a Raccoon Heist game using Claude Fable 5 , where I had Claude Fable 5 build a full working game from a premise I generated with GPT-3 and DALL-E four years ago . I decided to pose the exact same prompt to Codex Desktop running GPT-5.6 Sol Ultra - the mode where Sol makes aggressive use of sub-agents - to see how it would do. It produced a much better game! Here's Moonlight & Mayhem - GitHub repository here , including the textures and prompts it generated using gpt-image-2 . Your browser d...

Cloudflare has introduced Kitesurf, a cloud-hosted browser designed for AI agents instead of people. The company says the browser uses less computing power than Chromium for common automation tasks, helping developers build browser-based AI agents more efficiently.

Tool use transforms LLMs into agents that act beyond their training data, and for code-capable models, programmatic tool calling extends this further by replacing rigid JSON calls with scripts that chain and parallelize naturally. However, a systematic evaluation of tools as code on an established benchmark across current and prior model generations under real-world task conditions has not been conducted. In this work, we empirically compare programmatic tool calling (PTC) to native JSON tool calling across 14 language models on BFCL v4. In the programmatic tool calling paradigm, tools are ex...

Deciding which of two agents is stronger means playing games until skill outweighs luck, and every game costs money, model inference, or expert time. Since the number of games needed is unknown, fixed-budget evaluations either keep paying after the result is settled or stop before the agents can be told apart, while naive optional stopping with an ordinary confidence interval invalidates the stated level. We make such an evaluation stop as soon as its evidence suffices, with the guarantee intact. The Action-Informed Value Assessment Tool (AIVAT) reduces variance in imperfect-information games...

We give a formal mechanism design model for the continuous participatory governance of a deployed AI agent. The mechanism is built on the principle that governance should control an AI agent through resource allocation so as to make authorization self enforcing via compute budgets. The mechanism seeks to establish the Safe AI paradigm that compute is an effective governance lever. We situate our work as a compliance or commons overlay on a deployer. One governance period is an extensive form game in which verified human stakeholders arrive sequentially and contribute, on a provision or a reje...

Training terminal agents requires executable and verifiable tasks that are not merely solvable, but appropriately challenging for learning. Executable validation establishes feasibility, yet does not reveal how a task behaves relative to a given solver setting. In this paper, we present CalibForge, an autonomous terminal-task synthesis system that uses verified solver behavior to revise candidate tasks through adversarial solver calibration. Multi-solver calibration targets disagreement within a heterogeneous solver pool, whereas contrastive solver calibration targets a designated strong-pass...

Task-oriented conversational agents are evaluated using curated or automatically generated benchmarks, yet benchmark quality is rarely assessed. Poor benchmarks may contain inconsistent tasks, simplistic scenarios, or limited policy coverage, leading to unreliable evaluations. We introduce a reference-free framework that uses LLM judges to assess benchmark consistency, complexity, and policy coverage, while providing actionable diagnostics of weaknesses. We validate the framework by demonstrating agreement with independent human annotations and by evaluating benchmarks generated by LLMs of va...

Synthetic clinical benchmarks for enterprise AI agents can pass existing utility checks and still remain structurally unrealistic, especially in privacy-sensitive healthcare settings where operational data are hard to access. We study how to improve such benchmarks without breaking the downstream utility checks already used in practice. We formulate benchmark revision as utility-constrained realism improvement: dataset changes should increase realism while staying above an operational utility floor. We instantiate this idea on a care-gap benchmark derived from Synthea-generated patients exerc...

Despite advances in artificial intelligence (AI) across multiple sectors, today's AI tools, including deep learning and generative AI, still fail when embedded into physical systems, such as robots and vehicles operating under real-world physical laws. This stems from their inability to maintain reliable world models for long-horizon planning under uncertainty and generalize to unseen scenarios. In this context, wireless networks, through pervasive sensing and communication, can orchestrate physical intelligence. However, current architectures optimize throughput, latency, and reliability and...

Training large language model agents for long-horizon tool use typically relies on interactions with real or synthesized executable environments, whose construction and verification are costly, or on external simulators that are difficult to ground. We introduce EnvACE, an agentic reinforcement learning method that replaces external environment interaction during training with world rehearsal. The policy alternates between acting and rehearsal: it first generates a tool call, then plays the role of the environment to produce the response induced by that action, and conditions subsequent decis...

Agents backed by large skill libraries must decide which skills to load and in what order. Loading the entire library into context is expensive and provides no structure for autonomous sequencing. We study two systems for this problem over a corpus of 690 skills: a hybrid ranker combining lexical and dense-embedding retrieval for sparse, on-demand loading, and a typed knowledge graph encoding workflow relations such as prerequisites, data flow, and ordering. On a set of 117 realistic, non-echoing queries, the hybrid ranker retrieves the correct skill within the top five in 73.5% +/- 8.0 of ca...

Web agents observe a browser through text, pixels, or both, and the choice is usually fixed once for all tasks. We measure six observation modes across eight site-model combinations (cells) on VisualWebArena and WebArena and ask what choosing per task would buy. The modes are complementary: each solves tasks the others miss, they fail in structurally different ways, and the best choice reverses between task sets. The obvious prize, an oracle that picks a winning mode for every task, looks large but is inflated by run-to-run noise: rerunning the same mode on the same tasks changes 12-14% of ou...

Incident Report: unsanctioned agent behaviour during cyber testing It happened again . This time it was the UK government's AI Security Institute who accidentally attacked other companies while running an evaluation with models with the safety filters turned off. From their technical paper (PDF): During a cyber evaluation, from 25 to 28 July 2026, AI agents engaged in sustained, unsanctioned activity directed at what were, in practice, real people and organisations. These attempts were unsuccessful and, to the best of our knowledge, no real-world harm resulted. [...] Across 122 evaluation att...

The serial entrepreneur joins the e-commerce company as CPO to lead its AI agents.

In partially observable reinforcement learning, agents face a dual bottleneck: they must explore to encounter rewarding states and retain that experience in memory to optimize their policies. Exploration bonuses and memory architectures are traditionally evaluated in isolation, leaving their interaction unmeasured, and standard notions of sparse reward conflate temporal signal density with what the reward actually supervises. We present a controlled study crossing episodic exploration bonuses with diverse neural memory architectures across three environments that vary how the content of memor...

AI-supported care planning can help clinicians, patients, caregivers, and care teams coordinate complex decisions across clinical, functional, psychosocial, and environmental needs. However, many AI systems present recommendations as fixed outputs, limiting stakeholders' ability to inspect, challenge, and revise plans when they conflict with clinical judgment, patient values, or real-world feasibility. We present CoPlan - a Co-Intelligent and Contestable Interface for Human-AI Care Planning. CoPlan uses a multi-agent workflow in which specialized AI agents generate candidate interventions and...

Long-horizon search agents must make multiple sequential actions (steps) to search, retrieve, verify, and integrate evidence to reach a final answer. However, existing methods for training these agents typically treat all steps within a trajectory uniformly during both supervised fine-tuning (SFT) and reinforcement learning (RL), failing to distinguish useful actions from erroneous or redundant ones. In this paper, we propose Answer-Backtracked Credit Assignment (ABC), a fine-grained credit assignment framework for training long-horizon search agents by converting sparse trajectory-level outc...

Agents for long term reasoning require a memory that can be efficiently and effectively updated over time, as new facts and external feedback continue to arrive. Recently, graph memory has been adopted to offer structural organization for multi-hop retrieval and reasoning. However, existing methods store all memories in a flat graph, and accumulated historical memories can introduce irrelevant contexts and increase the cost of evidence selection during retrieval. Moreover, they typically update memory units independently, requiring repeated unit-wise rewrite to cover related changes. To addre...

Yet more rogue AI agents from OpenAI and Anthropic have been caught attempting to hack real targets online without permission. The discoveries add to a growing list of previously unknown incidents that have alarmed AI safety experts and intensified pressure for greater oversight of frontier systems. According to a report from the UK's AI Security Institute, which evaluates frontier models from top AI labs before they are released, agents powered by OpenAI's GPT-5.6-Sol and Anthropic's Mythos 5 went "engaged in sustained, potentially harmful activity directed at real people and organisations."...

Training LLM agents commonly relies on supervised fine-tuning from expert trajectories or online reinforcement learning over human-specified tasks with handcrafted verifiers. Though effective, both remain bottlenecked by externally specified tasks and supervision signals, limiting the scalability and diversity of agent training. We study an environment learning paradigm in which agents acquire interaction and manipulation capabilities solely through environment interaction, without externally specified tasks. We propose State2State, an environment-derived mid-training method that converts exp...

The week-old Open Secure AI Alliance, spearheaded by Nvidia and grown to over 120 companies, already has proposals out for defending against AI agents.

SocietyBench evaluates LLM agents on forecasting real social-world events by ingesting news/social media timelines and measuring counterfactual event prediction.

PAST-Bench benchmarks recursive self-improvement in personal AI agents by measuring whether retained preferences, task histories, and skills improve performance over sessions.

Video-DeepResearch extends multimodal agents to video streams; identifies modality bias and knowledge leakage bottlenecks in current models.

HIVE benchmark reveals voice transcription and keyboard perturbations reduce LLM accuracy across instruction tasks.

Game-theoretic analysis shows foundation model agents achieve rational cooperation through similarity inference, challenging decoupled-agency assumptions.

ContinualSkillBench evaluates whether LLM agents can continually acquire and reuse skills across 500 interconnected task domains.

MAFIA: query-only memory poisoning attacks on audited LLM agents via probing and factual injection.

Multi-agent clinical committees using Gemini show vulnerability to social shortcut cascades where peer consensus propagates errors across agents.