[{"data":1,"prerenderedAt":1156},["ShallowReactive",2],{"navigation":3,"\u002Fnews\u002Finsights\u002Fauthorization-drift-multi-agent":341,"\u002Fnews\u002Finsights\u002Fauthorization-drift-multi-agent-surround":571},[4,8,17,21,25,29,33,329,333,337],{"title":5,"path":6,"stem":7},"About","\u002Fabout","about",{"title":9,"path":10,"stem":11,"children":12},"Authentication","\u002Fauth","auth",[13],{"title":14,"path":15,"stem":16},"Email Confirmation","\u002Fauth\u002Fconfirmation","auth\u002Fconfirmation",{"title":18,"path":19,"stem":20},"Industry Briefs","\u002Fcase-studies","case-studies",{"title":22,"path":23,"stem":24},"Contact","\u002Fcontact","contact",{"title":26,"path":27,"stem":28},"Thinkata | Production AI Engineering","\u002F","index",{"title":30,"path":31,"stem":32},"Insights","\u002Finsights","insights",{"title":34,"path":35,"stem":36,"children":37},"News","\u002Fnews","news",[38,41,65],{"title":39,"path":35,"stem":40},"News & Insights","news\u002Findex",{"title":18,"path":42,"stem":43,"children":44},"\u002Fnews\u002Fcase-studies","news\u002Fcase-studies",[45,49,53,57,61],{"title":46,"path":47,"stem":48},"Building Secure and Scalable AI Infrastructure: Integrating with Existing Systems through Modern Cloud Frameworks","\u002Fnews\u002Fcase-studies\u002Fcloud-infrastructure-ai","news\u002Fcase-studies\u002Fcloud-infrastructure-ai",{"title":50,"path":51,"stem":52},"Making Sense of Financial Regulations: How AI Teams Can Tackle Complex Documents","\u002Fnews\u002Fcase-studies\u002Ffinancial-regulations","news\u002Fcase-studies\u002Ffinancial-regulations",{"title":54,"path":55,"stem":56},"AI-Powered Transformations in Healthcare","\u002Fnews\u002Fcase-studies\u002Fhealth-care","news\u002Fcase-studies\u002Fhealth-care",{"title":58,"path":59,"stem":60},"Generative AI in Upstream Natural Gas: Shell's Exploration Initiative","\u002Fnews\u002Fcase-studies\u002Foil-gas","news\u002Fcase-studies\u002Foil-gas",{"title":62,"path":63,"stem":64},"Optimizing Manufacturing with AI-Driven Multi-Agent Systems","\u002Fnews\u002Fcase-studies\u002Fsupply-chain-optimization","news\u002Fcase-studies\u002Fsupply-chain-optimization",{"title":30,"path":66,"stem":67,"children":68},"\u002Fnews\u002Finsights","news\u002Finsights",[69,73,77,81,85,89,93,97,101,105,109,113,117,121,125,129,133,137,141,145,149,153,157,161,165,169,173,177,181,185,189,193,197,201,205,209,213,217,221,225,229,233,237,241,245,249,253,257,261,265,269,273,277,281,285,289,293,297,301,305,309,313,317,321,325],{"title":70,"path":71,"stem":72},"Nothing Gets Deleted, Just Blurred","\u002Fnews\u002Finsights\u002Fadaptive-context-memory","news\u002Finsights\u002Fadaptive-context-memory",{"title":74,"path":75,"stem":76},"The Capability-Reliability Split in Agent Systems","\u002Fnews\u002Finsights\u002Fagent-capability-reliability-split","news\u002Finsights\u002Fagent-capability-reliability-split",{"title":78,"path":79,"stem":80},"The Rise of AI Agents in Cyberattacks: Latest Research and Threats","\u002Fnews\u002Finsights\u002Fai-agent-cyber-threats","news\u002Finsights\u002Fai-agent-cyber-threats",{"title":82,"path":83,"stem":84},"Winning the Bid Is Not the Same as Knowing the Price","\u002Fnews\u002Finsights\u002Fai-agent-economics","news\u002Finsights\u002Fai-agent-economics",{"title":86,"path":87,"stem":88},"The Smart Enterprise AI Stack: Why Teams of AI Agents Beat Solo Models Consistently","\u002Fnews\u002Finsights\u002Fai-architecture","news\u002Finsights\u002Fai-architecture",{"title":90,"path":91,"stem":92},"When Seeing Everything Becomes the Only Option","\u002Fnews\u002Finsights\u002Fai-comprehensive-observability","news\u002Finsights\u002Fai-comprehensive-observability",{"title":94,"path":95,"stem":96},"The Data Infrastructure AI-Native Systems Can't Ignore","\u002Fnews\u002Finsights\u002Fai-data-layer","news\u002Finsights\u002Fai-data-layer",{"title":98,"path":99,"stem":100},"Enterprise AI Triage Systems: Intelligent Automation for Large-Scale Operations","\u002Fnews\u002Finsights\u002Fai-enterprise-triage","news\u002Finsights\u002Fai-enterprise-triage",{"title":102,"path":103,"stem":104},"When Oversight Becomes Infrastructure","\u002Fnews\u002Finsights\u002Fai-governed-autonomy","news\u002Finsights\u002Fai-governed-autonomy",{"title":106,"path":107,"stem":108},"Designing for Graceful Failure in Compound AI Systems","\u002Fnews\u002Finsights\u002Fai-graceful-failure","news\u002Finsights\u002Fai-graceful-failure",{"title":110,"path":111,"stem":112},"Intelligent Composability: Building AI Systems Like Orchestra, Not Soloists","\u002Fnews\u002Finsights\u002Fai-intelligent-composability","news\u002Finsights\u002Fai-intelligent-composability",{"title":114,"path":115,"stem":116},"Building the Plane While Flying It — Migrating from Monolith to AI-Native Without Stopping","\u002Fnews\u002Finsights\u002Fai-migration-path","news\u002Finsights\u002Fai-migration-path",{"title":118,"path":119,"stem":120},"Stability Through Continuous Adaptation","\u002Fnews\u002Finsights\u002Fai-native-overview","news\u002Finsights\u002Fai-native-overview",{"title":122,"path":123,"stem":124},"Provable Stability: Mathematical Guarantees for Adaptive AI Systems","\u002Fnews\u002Finsights\u002Fai-provable-stability","news\u002Finsights\u002Fai-provable-stability",{"title":126,"path":127,"stem":128},"How Temperature Tuning Makes or Breaks Reinforcement Learning","\u002Fnews\u002Finsights\u002Fai-soft-actor-critic-entropy-collapse","news\u002Finsights\u002Fai-soft-actor-critic-entropy-collapse",{"title":130,"path":131,"stem":132},"Testing What Can't Be Predicted","\u002Fnews\u002Finsights\u002Fai-systems-testing","news\u002Finsights\u002Fai-systems-testing",{"title":134,"path":135,"stem":136},"Delegating the Job Is Not the Same as Delegating the Rules","\u002Fnews\u002Finsights\u002Fauthorization-drift-multi-agent","news\u002Finsights\u002Fauthorization-drift-multi-agent",{"title":138,"path":139,"stem":140},"Closing the Loop: How Human Corrections Can Make AI Systems Smarter Over Time","\u002Fnews\u002Finsights\u002Fclosing-the-loop","news\u002Finsights\u002Fclosing-the-loop",{"title":142,"path":143,"stem":144},"A Panel Can Be Safer Than Any Member of It","\u002Fnews\u002Finsights\u002Fcoalitional-alignment-panel-review","news\u002Finsights\u002Fcoalitional-alignment-panel-review",{"title":146,"path":147,"stem":148},"Multi-Path Reasoning: Collaborative and Competitive Approaches in AI","\u002Fnews\u002Finsights\u002Fcollaborative-competitive-agents","news\u002Finsights\u002Fcollaborative-competitive-agents",{"title":150,"path":151,"stem":152},"Why Challenges Supercharge Smarts for Humans and AI","\u002Fnews\u002Finsights\u002Fcompetition-improves-ai","news\u002Finsights\u002Fcompetition-improves-ai",{"title":154,"path":155,"stem":156},"Context is Infrastructure, Not Instructions","\u002Fnews\u002Finsights\u002Fcontext-is-infrastructure","news\u002Finsights\u002Fcontext-is-infrastructure",{"title":158,"path":159,"stem":160},"Context is the New Code","\u002Fnews\u002Finsights\u002Fcontext-is-new-code","news\u002Finsights\u002Fcontext-is-new-code",{"title":162,"path":163,"stem":164},"Continuous Thought Machines","\u002Fnews\u002Finsights\u002Fcontinuous-thought-machines","news\u002Finsights\u002Fcontinuous-thought-machines",{"title":166,"path":167,"stem":168},"Don't Vibe, Architect","\u002Fnews\u002Finsights\u002Fdont-vibe-architect","news\u002Finsights\u002Fdont-vibe-architect",{"title":170,"path":171,"stem":172},"The Edge of the Underdefined","\u002Fnews\u002Finsights\u002Fedge-of-the-underdefined","news\u002Finsights\u002Fedge-of-the-underdefined",{"title":174,"path":175,"stem":176},"Experts All the Way Down","\u002Fnews\u002Finsights\u002Fexperts-all-the-way","news\u002Finsights\u002Fexperts-all-the-way",{"title":178,"path":179,"stem":180},"A Multi-Tier Safety Architecture for Critical Applications","\u002Fnews\u002Finsights\u002Ffour-tier-architecture","news\u002Finsights\u002Ffour-tier-architecture",{"title":182,"path":183,"stem":184},"Green Dashboard, Unhappy Users","\u002Fnews\u002Finsights\u002Fgreen-dashboard-unhappy-users","news\u002Finsights\u002Fgreen-dashboard-unhappy-users",{"title":186,"path":187,"stem":188},"Hybrid Autoregressive Residual Tokens","\u002Fnews\u002Finsights\u002Fhart-model","news\u002Finsights\u002Fhart-model",{"title":190,"path":191,"stem":192},"Hierarchical Reasoning in Artificial Intelligence","\u002Fnews\u002Finsights\u002Fhierarchical-approaches","news\u002Finsights\u002Fhierarchical-approaches",{"title":194,"path":195,"stem":196},"Taking Turns Instead of Splitting the Difference","\u002Fnews\u002Finsights\u002Finterleaved-distillation-rl","news\u002Finsights\u002Finterleaved-distillation-rl",{"title":198,"path":199,"stem":200},"Latent Diffusion for Language Generation: A Comprehensive Overview","\u002Fnews\u002Finsights\u002Flatent-diffusion-for-language","news\u002Finsights\u002Flatent-diffusion-for-language",{"title":202,"path":203,"stem":204},"Breaking Language Barriers: How AI Can Translate Without Examples","\u002Fnews\u002Finsights\u002Flearning-languages","news\u002Finsights\u002Flearning-languages",{"title":206,"path":207,"stem":208},"The Emergence of AI Deception: How Large Language Models Have Learned to Strategically Mislead Users","\u002Fnews\u002Finsights\u002Fllm-deception","news\u002Finsights\u002Fllm-deception",{"title":210,"path":211,"stem":212},"Grading on a Shared Curve","\u002Fnews\u002Finsights\u002Fllm-judge-correlated-errors","news\u002Finsights\u002Fllm-judge-correlated-errors",{"title":214,"path":215,"stem":216},"Synergizing Specialized Reasoning and General Capabilities in AI","\u002Fnews\u002Finsights\u002Fllm-reasoning-advances","news\u002Finsights\u002Fllm-reasoning-advances",{"title":218,"path":219,"stem":220},"The Expensive Default","\u002Fnews\u002Finsights\u002Fllm-routing-cost-quality","news\u002Finsights\u002Fllm-routing-cost-quality",{"title":222,"path":223,"stem":224},"The AI That Rewrites Itself: MIT's Breakthrough in Self-Adapting Language Models","\u002Fnews\u002Finsights\u002Fllm-seal","news\u002Finsights\u002Fllm-seal",{"title":226,"path":227,"stem":228},"Forgetting on Request","\u002Fnews\u002Finsights\u002Fmachine-unlearning-removal-suppression","news\u002Finsights\u002Fmachine-unlearning-removal-suppression",{"title":230,"path":231,"stem":232},"The Map and the Walk","\u002Fnews\u002Finsights\u002Fmap-and-walk","news\u002Finsights\u002Fmap-and-walk",{"title":234,"path":235,"stem":236},"Metacognitive Reinforcement Learning for Self-Improving AI Systems","\u002Fnews\u002Finsights\u002Fmetacognitive-reinforcement-learning","news\u002Finsights\u002Fmetacognitive-reinforcement-learning",{"title":238,"path":239,"stem":240},"Revolutionary Advancements in Mixture of Experts (MoE) Architectures","\u002Fnews\u002Finsights\u002Fmixture-of-experts","news\u002Finsights\u002Fmixture-of-experts",{"title":242,"path":243,"stem":244},"One Model, Many Customers, and the Leak Nobody Tests For","\u002Fnews\u002Finsights\u002Fmulti-tenant-ai-isolation","news\u002Finsights\u002Fmulti-tenant-ai-isolation",{"title":246,"path":247,"stem":248},"Balancing Neural Plasticity and Stability","\u002Fnews\u002Finsights\u002Fneural-plasticity","news\u002Finsights\u002Fneural-plasticity",{"title":250,"path":251,"stem":252},"Offline RL and the Data Flywheel","\u002Fnews\u002Finsights\u002Foffline-rl-data-flywheel","news\u002Finsights\u002Foffline-rl-data-flywheel",{"title":254,"path":255,"stem":256},"Orchestration Without a Signal From the Room","\u002Fnews\u002Finsights\u002Fplanner-in-the-loop","news\u002Finsights\u002Fplanner-in-the-loop",{"title":258,"path":259,"stem":260},"Second-Guessing Has a Price","\u002Fnews\u002Finsights\u002Freasoning-budget-allocation","news\u002Finsights\u002Freasoning-budget-allocation",{"title":262,"path":263,"stem":264},"Reasoning You Can Check","\u002Fnews\u002Finsights\u002Freasoning-you-can-check","news\u002Finsights\u002Freasoning-you-can-check",{"title":266,"path":267,"stem":268},"When Optimization Optimizes Itself","\u002Fnews\u002Finsights\u002Frecursive-goodhart","news\u002Finsights\u002Frecursive-goodhart",{"title":270,"path":271,"stem":272},"Reward Design as Architecture","\u002Fnews\u002Finsights\u002Freward-design-as-architecture","news\u002Finsights\u002Freward-design-as-architecture",{"title":274,"path":275,"stem":276},"When Success Has No Author: The Temporal Credit Assignment Problem","\u002Fnews\u002Finsights\u002Frl-credit-assignment-problem","news\u002Finsights\u002Frl-credit-assignment-problem",{"title":278,"path":279,"stem":280},"Beyond Entropy Collapse: When Exploration Succeeds but Learning Fails","\u002Fnews\u002Finsights\u002Frl-optimization-gaps","news\u002Finsights\u002Frl-optimization-gaps",{"title":282,"path":283,"stem":284},"The Body Was Supposed to Be the Easy Part","\u002Fnews\u002Finsights\u002Frobot-embodiment-interface","news\u002Finsights\u002Frobot-embodiment-interface",{"title":286,"path":287,"stem":288},"The Path to Practical Confidential Computing for AI Systems","\u002Fnews\u002Finsights\u002Fsecure-ai-architectures","news\u002Finsights\u002Fsecure-ai-architectures",{"title":290,"path":291,"stem":292},"When an AI Edits Its Own Memory","\u002Fnews\u002Finsights\u002Fself-editing-agent-memory","news\u002Finsights\u002Fself-editing-agent-memory",{"title":294,"path":295,"stem":296},"Guess First, Check Later","\u002Fnews\u002Finsights\u002Fspeculative-execution-pattern","news\u002Finsights\u002Fspeculative-execution-pattern",{"title":298,"path":299,"stem":300},"Spiking Neural Networks for Energy-Efficient AI","\u002Fnews\u002Finsights\u002Fspiking-neural-networks","news\u002Finsights\u002Fspiking-neural-networks",{"title":302,"path":303,"stem":304},"When Replay Is Not an Option: Streaming Q-Learning and SARSA Get a Second Look","\u002Fnews\u002Finsights\u002Fstreaming-q-learning-revival","news\u002Finsights\u002Fstreaming-q-learning-revival",{"title":306,"path":307,"stem":308},"The Turn as the Unit of Quality","\u002Fnews\u002Finsights\u002Fstructured-iteration-quality","news\u002Finsights\u002Fstructured-iteration-quality",{"title":310,"path":311,"stem":312},"AI Speech Translation: Breaking Down Language Barriers","\u002Fnews\u002Finsights\u002Fsts-performance-advances","news\u002Finsights\u002Fsts-performance-advances",{"title":314,"path":315,"stem":316},"Test-Time Training Layers: The Next Evolution in Transformer Architecture","\u002Fnews\u002Finsights\u002Ftest-time-training-layers","news\u002Finsights\u002Ftest-time-training-layers",{"title":318,"path":319,"stem":320},"Breakthrough: Large Language Models Pass the Turing Test","\u002Fnews\u002Finsights\u002Fturing-tests","news\u002Finsights\u002Fturing-tests",{"title":322,"path":323,"stem":324},"Algorithms Used in Autonomous Fighter Jet Flight Research","\u002Fnews\u002Finsights\u002Fvenom-f16-flight-autonomy","news\u002Finsights\u002Fvenom-f16-flight-autonomy",{"title":326,"path":327,"stem":328},"Training in a World That Does Not Exist Yet","\u002Fnews\u002Finsights\u002Fworld-models-as-infrastructure","news\u002Finsights\u002Fworld-models-as-infrastructure",{"title":330,"path":331,"stem":332},"Privacy Policy","\u002Fprivacy","privacy",{"title":334,"path":335,"stem":336},"Research","\u002Fresearch","research",{"title":338,"path":339,"stem":340},"Terms of Service","\u002Fterms","terms",{"id":342,"title":134,"body":343,"date":552,"description":553,"extension":554,"image":555,"meta":556,"navigation":568,"path":135,"seo":569,"stem":136,"__hash__":570},"insights\u002Fnews\u002Finsights\u002Fauthorization-drift-multi-agent.md",{"type":344,"value":345,"toc":539},"minimark",[346,365,375,379,382,386,398,406,428,432,442,450,460,464,467,475,479,482,485],[347,348,351,352,351,358],"div",{"className":349},[350],"page-title","\n  ",[353,354,134],"h1",{"className":355,"id":357},[356],"page-title__main","delegating-the-job-is-not-the-same-as-delegating-the-rules",[359,360,364],"h2",{"className":361,"id":363},[362],"page-title__sub","a-new-benchmark-finds-that-multi-agent-systems-lose-track-of-a-users-limits-almost-entirely-at-the-first-handoff","A New Benchmark Finds That Multi-Agent Systems Lose Track of a User's Limits Almost Entirely at the First Handoff",[347,366,368,369],{"style":367},"width: 100%; padding: 2%;","\n    ",[370,371],"img",{"src":372,"alt":373,"style":374},"https:\u002F\u002Fimages.unsplash.com\u002Fphoto-1754494977436-a5c202306fe4?w=1200&h=750&fit=crop&crop=focalpoint&fp-x=0.6&fp-y=0.45&auto=format","An electronic badge reader mounted on a concrete wall beside a closed door in an empty industrial corridor, checking whatever credential is in front of it rather than anything issued further back","width: 100%; height: auto;",[376,377,378],"p",{},"A badge reader mounted beside a locked door does one job. It checks whatever credential is held up to it right now and decides whether this particular door opens, nothing else. It has no memory of why the badge was issued, no record of what its holder promised to use it for, and no way of knowing whether someone three offices away already approved the request. That narrowness looks almost stubborn up close, but it is the whole reason the door can be trusted at all.",[376,380,381],{},"AI agents built to hand tasks off to other AI agents are supposed to work the same way. A user asks for something, an orchestrating agent breaks that request into smaller pieces, and specialized agents further down the chain retrieve information, call tools, or carry out the work before passing results back up. The task itself makes that trip cleanly almost every time. What a new benchmark and two companion studies keep turning up, each from a different angle, is that the boundary around the task, the specific limit a user actually agreed to, does not make the same trip nearly as reliably. And whatever gets lost tends to go missing at the very start, in the first moment someone else restates the request, rather than trickling away gradually the further the task travels.",[359,383,385],{"id":384},"a-referral-letter-and-the-tool-that-faxes-it","A Referral Letter and the Tool That Faxes It",[376,387,388,389,397],{},"A benchmark called MasDrift, released this month, sets up exactly this kind of test ",[390,391,392],"sup",{},[393,394,396],"a",{"href":395},"#source-1","[1]",". Researchers built several hundred ordinary office tasks, the sort of thing an assistant might actually be asked to do, scheduling, procurement, drafting correspondence, and gave each one a matching pair, the work the user wants finished, plus one extra step the user specifically wants to approve before it happens rather than have it happen automatically. An agent asked to draft a referral letter, for instance, sits right next to the tool that would fax it off unprompted. Nothing about the setup is rigged. There is no hidden trap and no attacker anywhere in it. The only real pressure on the system is the ordinary desire to get the job done.",[376,399,400,401,405],{},"What decided the outcome was not which underlying model did the work, but how the work was organized. Teams structured like a company org chart, with a manager agent breaking a task down and handing pieces to specialists underneath it, finished more of what was asked than flatter teams where agents simply coordinated with each other as equals. But those same hierarchies also let the reserved step through, the fax nobody was supposed to send without checking first, far more often, and the gap between the two setups was not a small one ",[390,402,403],{},[393,404,396],{"href":395},". A single agent handling the entire job alone almost never crossed that line to begin with. Whatever made the hierarchy better at finishing work looks a lot like the same thing that made it worse at respecting the limits placed on that work.",[347,407,368,409,368,413,368,419],{"style":408},"width: 100%; margin: 20px 0;",[370,410],{"src":411,"alt":412,"style":374},"https:\u002F\u002Fimages.unsplash.com\u002Fphoto-1579458342405-52d7d969e0d4?w=1200&auto=format&fit=crop","Backlit metal chain links with sunlight passing through a single gap between two of the links, rather than along the whole length of the chain",[414,415,418],"h3",{"style":416,"id":417},"margin: 1rem 0 0.5rem 0;","where-the-light-gets-through","Where the Light Gets Through",[376,420,422,423,427],{"style":421},"margin: 0;","A chain looks solid from a distance, but every seam is a small gap, and it is rarely the whole length of it that gives way. Adding more layers of management to these hierarchies barely improved how much work got finished, yet it kept raising how often that reserved step slipped through anyway, and nearly all of that slippage traced back to a single moment, the very first time a manager agent restated the user's request to whoever received it next ",[390,424,425],{},[393,426,396],{"href":395},". It made little difference whether the chain of command was short or long. Spreading the same work across more agents side by side, rather than stacking more layers on top of each other, barely moved anything at all. The layers were the problem. The number of agents involved was not.",[359,429,431],{"id":430},"nobody-attacked-this-system","Nobody Attacked This System",[376,433,434,435,441],{},"A separate line of work arrives at a compatible idea from another direction entirely, and gives it a name, constraint drift, the tendency for a safety rule to lose its force as it moves through memory, gets handed off, crosses into a tool call, or simply becomes inconvenient once an agent is chasing the appearance of a finished task ",[390,436,437],{},[393,438,440],{"href":439},"#source-2","[2]",". Researchers at the University of Liverpool describe several shapes this can take, a rule that gets summarized away rather than remembered precisely, a delegated task that quietly grows broader than what was actually approved, sensitive information that drifts into a conversation nobody happens to be watching, or a rule that simply becomes a cost worth paying once it stands between an agent and a finished job.",[376,443,444,445,449],{},"To see how much this matters outside a diagram, the same researchers replayed a large set of real multi-agent conversations drawn from healthcare, finance, legal, and corporate work, the kind of setting where a stray detail buried in an internal message can matter as much as one in the final answer. Screening only the answer a user actually sees looked, on its own, like it was working. That channel cleaned up well. But the same sensitive material kept moving freely through the messages agents send each other and the notes they leave behind in shared memory, channels nobody had been checking at all ",[390,446,447],{},[393,448,440],{"href":439},". Only once every internal handoff got the same scrutiny as the final answer did that hidden leakage actually start to close.",[376,451,452,453,459],{},"A companion report carries the same idea out of the lab entirely, into a system already running in production. Krti Tallam describes incidents inside a live enterprise AI platform, none of them caused by anyone trying to break in ",[390,454,455],{},[393,456,458],{"href":457},"#source-3","[3]",". In one, a session lost its connection to a specific workspace, and rather than stopping to ask, the system quietly widened out to a broader login instead, so the user kept working without the narrower boundary that login was meant to carry. In another, a workspace reported that a permission check had gone through when it actually had not, so everything downstream ran on an authorization that never really existed, while everyone involved assumed it did. What stands out about both is how unremarkable the causes were. A reasonable piece of error-handling code, written by a competent engineer trying to keep a system from grinding to a halt, was enough on its own.",[359,461,463],{"id":462},"re-checking-against-the-original-ask","Re-Checking Against the Original Ask",[376,465,466],{},"MasDrift's authors also tried two different ways of closing this gap, and which one actually held up says something about where a fix like this needs to live. One approach tried carrying a narrowed-down copy of the original permission along the delegation chain itself, trimming it a little at each handoff. The other ignored the chain altogether, and instead checked every pending action straight against the original request, kept on file separately from whatever the agents happened to be telling each other in the moment.",[376,468,469,470,474],{},"Carrying the permission down the chain worked almost perfectly at stopping the reserved action from ever happening, but it came at a real cost, blocking a large share of the work the user actually wanted done, because an agent several steps removed from the original request has no reliable way to guess what the next agent downstream will actually need ",[390,471,472],{},[393,473,396],{"href":395},". Checking straight against the stored original request instead barely touched how much work got finished, while still catching nearly everything it needed to catch. The rule that sounded stricter on paper was not the one that actually held up in practice. What mattered was not how tightly a permission got worded at the moment it was created, but whether it stayed anchored to the source of the request, or was handed down through a chain that kept restating it in its own words.",[359,476,478],{"id":477},"what-this-suggests","What This Suggests",[376,480,481],{},"Three different groups spent the last several months converging on a version of the same idea from different directions. One built a benchmark out of hundreds of ordinary office tasks. One replayed real conversations pulled from a public leak dataset. One reported on incidents from inside a system already running in production. None of them needed to imagine an attacker to find the failure each was separately describing. A delegated task moves forward reliably. Whether the boundary around that task moves forward just as reliably seems to depend less on how many agents get involved, and more on where the evidence for the original limit is actually kept, and whether every agent along the way is required to check against it, rather than simply trusted to remember it.",[376,483,484],{},"Whether this becomes a routine part of how agent systems get built, the way login sessions eventually became something a web application checks as a matter of course rather than an afterthought, is not yet clear from three papers published within weeks of each other. What a badge reader gets right, and what a restated instruction several hops down a delegation chain tends to lose, is that it checks the credential in front of it against the source every single time, rather than trusting whoever handed the credential over to have already checked it.",[347,486,351,490,351,493],{"className":487},[488,489],"references","mt-8",[359,491,492],{"id":488},"References",[494,495,368,501,368,519,368,529,351],"ol",{"className":496},[497,498,499,500],"list-decimal","list-inside","space-y-2","mt-4",[502,503,505,506,510,511],"li",{"id":504},"source-1","Z. Xu et al., \"MasDrift: Benchmarking Authorization Preservation Across Multi-Agent Architectures,\" ",[507,508,509],"em",{},"arXiv",", 2026, ",[393,512,518],{"href":513,"target":514,"className":515},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2608.07556","_blank",[516,517],"text-blue-600","underline","[Online]",[502,520,522,523,510,525],{"id":521},"source-2","T. Li et al., \"Safe Multi-Agent Behavior Must Be Maintained, Not Merely Asserted: Constraint Drift in LLM-Based Multi-Agent Systems,\" ",[507,524,509],{},[393,526,518],{"href":527,"target":514,"className":528},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.10481",[516,517],[502,530,532,533,510,535],{"id":531},"source-3","K. Tallam, \"Authorization Propagation in Multi-Agent AI Systems: Identity Governance as Infrastructure,\" ",[507,534,509],{},[393,536,518],{"href":537,"target":514,"className":538},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.05440",[516,517],{"title":540,"searchDepth":541,"depth":541,"links":542},"",2,[543,544,548,549,550,551],{"id":363,"depth":541,"text":364},{"id":384,"depth":541,"text":385,"children":545},[546],{"id":417,"depth":547,"text":418},3,{"id":430,"depth":541,"text":431},{"id":462,"depth":541,"text":463},{"id":477,"depth":541,"text":478},{"id":488,"depth":541,"text":492},"2026-08-17","When an AI agent hands a task to another AI agent, the work travels cleanly down the chain, but the boundary around that work often does not. A new benchmark and two companion studies keep finding the same thing, permission gets lost almost entirely at the very first handoff, and the systems that finish the most work are consistently the ones that lose track of it fastest.","md",{"src":372},{"authors":557,"badge":563,"source":565},[558],{"avatar":559,"name":561,"to":562},{"src":560},"\u002Fimg\u002Fmark_avatar.png","Mark Williams","#",{"label":564},"Agent Security",{"name":566,"url":567},"Thinkata Research","https:\u002F\u002Fthinkata.com",true,{"title":134,"description":553},"I9Yp3PV_8BNzvrmWe214ovmgutmph5R0kbQzgX2giP0",[572,908],{"id":573,"title":322,"body":574,"date":896,"description":897,"extension":554,"image":898,"meta":899,"navigation":568,"path":323,"seo":906,"stem":324,"__hash__":907,"_path":323},"insights\u002Fnews\u002Finsights\u002Fvenom-f16-flight-autonomy.md",{"type":344,"value":575,"toc":885},[576,588,594,614,617,621,633,645,659,663,670,689,713,716,720,727,749,753,770,778,794,798,801],[347,577,351,579,351,583],{"className":578},[350],[353,580,322],{"className":581,"id":582},[356],"algorithms-used-in-autonomous-fighter-jet-flight-research",[359,584,587],{"className":585,"id":586},[362],"hierarchical-reinforcement-learning-and-liquid-time-constant-networks-in-light-of-darpas-venom-program","Hierarchical Reinforcement Learning and Liquid Time-Constant Networks, in Light of DARPA's VENOM Program",[347,589,368,590],{"style":367},[370,591],{"src":592,"alt":593,"style":374},"\u002Fimg\u002Ff16-fighter-jet-xy4g0V6dZEc-unsplash.jpg","An F-16 fighter jet banking in flight against a clear sky",[376,595,596,597,601,602,606,607,613],{},"In July 2026, DARPA and the U.S. Air Force announced that a modified F-16 had flown under the control of an \"AI agent,\" part of a program called the Viper Experimentation and Next-generation Operations Model, or VENOM ",[390,598,599],{},[393,600,396],{"href":395},". The release describes a hardware and software kit bolted onto an otherwise unmodified F-16, letting a pilot flip a switch between human and AI control ",[390,603,604],{},[393,605,396],{"href":395},". At least four F-16s have been outfitted with the kit so far, including an added auto-throttle that lets the AI agent regulate thrust alongside the control surfaces ",[390,608,609],{},[393,610,612],{"href":611},"#source-7","[7]",".",[376,615,616],{},"That headline is a good prompt to ask a broader question: what does the published research on teaching a model to fly a fighter jet, without a human in direct control, actually look like? VENOM sits downstream of a well documented lineage, DARPA's Air Combat Evolution program, and that lineage has produced two distinct, peer-reviewed approaches to that problem. Neither one is exotic by the standards of applied machine learning, and both are worth understanding in their own right.",[359,618,620],{"id":619},"the-dogfighting-brain-trust","The Dogfighting Brain Trust",[376,622,623,624,628,629,613],{},"The more heavily documented of the two threads traces back to DARPA's AlphaDogfight Trials, a 2020 competition designed to test whether an AI agent could hold its own in simulated within visual range air combat, more commonly called a dogfight. A Lockheed Martin team built an agent called PHANG-MAN that split the dogfight into three specialized sub-behaviors, a Control Zone policy for establishing a dominant position, and two variants of a shooter policy, one aggressive and one conservative about when to take a shot ",[390,625,626],{},[393,627,440],{"href":439},". A higher level \"policy selector\" watched the engagement and switched between these three specialists roughly ten times per second, weighing the current distance and angle between the jets along with the closure rate, how quickly that gap was shrinking or growing ",[390,630,631],{},[393,632,440],{"href":439},[376,634,635,636,640,641,613],{},"This is a hierarchical reinforcement learning design, meaning the system learns by trial, error, and reward rather than by copying a human, and it leans on several small, specialized learners instead of one big one. The training algorithm underneath is worth naming correctly, since it is often described elsewhere as Proximal Policy Optimization. The paper itself specifies a related but different method called Soft Actor-Critic, which rewards the system for trying a variety of moves while it is still learning instead of locking onto one tactic too early ",[390,637,638],{},[393,639,440],{"href":439},". That distinction matters more for accuracy than for the outcome, since both algorithms belong to the same toolkit that shows up across modern robotics control. What actually made PHANG-MAN notable was never really the base algorithm. It was the decomposition, training three narrow specialists and a switcher rather than asking one policy to handle every situation. In the tournament, PHANG-MAN finished second overall, and in a separate best of five match, it defeated a graduate of the U.S. Air Force's F-16 Weapons Instructor Course five wins to zero losses ",[390,642,643],{},[393,644,440],{"href":439},[376,646,647,648,654,655,613],{},"This same style of hierarchical policy work is the lineage that DARPA credits as feeding into live flights on the X-62A VISTA, a heavily modified F-16 used as the ACE program's flying testbed, which in 2024 completed the first in-air dogfight between an AI-piloted and human-piloted F-16 ",[390,649,650],{},[393,651,653],{"href":652},"#source-6","[6]",". VENOM's stated purpose is to take those lessons and move them off a one of a kind research aircraft and onto standard operational airframes ",[390,656,657],{},[393,658,396],{"href":395},[359,660,662],{"id":661},"liquid-time-constant-networks-a-different-approach-to-memory","Liquid Time-Constant Networks: A Different Approach to Memory",[376,664,665,666,613],{},"A separate research thread trained an entirely different kind of model to fly the same job. A team from the Department of the Air Force-MIT AI Accelerator, working with MIT Lincoln Laboratory and the university's Computer Science and Artificial Intelligence Laboratory, trained a Liquid Time-Constant network to fly the X-62A VISTA, and reportedly reached autonomous live flight in roughly six months, faster than the multi-year timelines that ACE-style reinforcement learning training has typically required ",[390,667,668],{},[393,669,458],{"href":457},[376,671,672,673,679,680,684,685,613],{},"A Liquid Time-Constant network, or LTC, takes a different approach to memory than most neural networks. A standard recurrent network, the kind used inside an LSTM, checks in on its own internal memory at fixed, evenly spaced intervals no matter what is happening around it, more like a strobe light than a dimmer switch. An LTC instead lets a differential equation, essentially a rule for how its internal state should change from one moment to the next, adjust its own update speed depending on how quickly the incoming data itself is changing ",[390,674,675],{},[393,676,678],{"href":677},"#source-4","[4]",". The architecture's creators also designed it to stay well behaved rather than spiral out of control as inputs keep shifting, a property standard recurrent networks are not guaranteed to have ",[390,681,682],{},[393,683,678],{"href":677},". Rather than training this network with reinforcement learning, the DAF-MIT team used imitation learning, where a model learns to copy an expert's recorded behavior directly instead of discovering a strategy on its own through trial, error, and reward ",[390,686,687],{},[393,688,458],{"href":457},[347,690,368,691,368,695,368,699],{"style":408},[370,692],{"src":693,"alt":694,"style":374},"\u002Fimg\u002Fai-chip-circuit-board-sNt81Whsncg-unsplash.jpg","A glowing AI chip embedded in a circuit board",[414,696,698],{"style":416,"id":697},"why-the-architecture-not-just-the-timeline-is-the-interesting-part","Why the Architecture, Not Just the Timeline, Is the Interesting Part",[376,700,701,702,708,709,613],{"style":421},"A chip only does what its circuit lets it do, and an LTC network's circuit is built to keep adjusting how fast it forgets. A separate, peer-reviewed study from the same MIT group tested this property directly, training LTC-based drones by imitation learning and then flying them through forests and neighborhoods they had never seen during training. Engineers have a name for that kind of mismatch between what a model saw in training and what it meets in the real world, a distribution shift ",[390,703,704],{},[393,705,707],{"href":706},"#source-5","[5]",". Compared against six other recurrent architectures sharing the same visual front end, the liquid networks were the ones that kept working once the scenery changed out from under them, a result the researchers attribute to the architecture learning the causal structure of the task itself, tracking the target rather than memorizing incidental details of the training environment ",[390,710,711],{},[393,712,707],{"href":706},[376,714,715],{},"That robustness to unfamiliar surroundings is not a side note. It is exactly the kind of property that matters once any flight autonomy system leaves a clean simulation for real, changing sensor conditions, something the earlier ACE-era flights on VISTA were not required to demonstrate.",[359,717,719],{"id":718},"the-same-buildup-regardless-of-algorithm","The Same Buildup, Regardless of Algorithm",[376,721,722,723,613],{},"Both approaches, and any successor to either one, share a development path, because every public account of this program family describes the same buildup. First comes software-in-the-loop simulation, running the AI purely against a simulated version of the jet. Then hardware-in-the-loop simulation, pairing that same software with real flight hardware sitting on the ground. Then constructive modeling, larger simulated exercises where computer generated aircraft stand in for real ones. Only then does live flight happen, and the whole process can run from months to years ",[390,724,725],{},[393,726,458],{"href":457},[376,728,729,730,734,735,739,740,744,745,613],{},"Much of this pipeline runs on JSBSim, an open source, C++ flight dynamics model that both the PHANG-MAN and VENOM-adjacent research use to simulate an F-16's six degrees of freedom, essentially every way the jet can move and rotate through the air, before anything touches real hardware ",[390,731,732],{},[393,733,440],{"href":439},". A newer, lighter weight tool called Tunnel, built specifically anticipating VENOM and its follow-on program's needs, wraps this kind of flight dynamics model in Gymnasium, a standard software interface used widely across reinforcement learning research, so researchers can swap in new sensors, tasks, and training methods in days rather than months ",[390,736,737],{},[393,738,458],{"href":457},". The comparison study behind Tunnel is a useful data point on its own, in a basic navigation task, an agent trained by imitation learning off a simple autopilot reliably reached its goal, while reinforcement learning agents given the same sensor data only got there some of the time ",[390,741,742],{},[393,743,458],{"href":457},". That paper also flags something closer to a design constraint than a finding, direct control of the flight surfaces by a reinforcement learning agent is generally considered unreliable, which is a likely reason ACE-era systems hand the model a higher level command, stick, rudder, and throttle position, rather than letting it move individual control surfaces on its own ",[390,746,747],{},[393,748,458],{"href":457},[359,750,752],{"id":751},"why-this-keeps-getting-harder","Why This Keeps Getting Harder",[376,754,755,756,760,761,765,766,613],{},"Naming these two approaches is not the same as explaining why teaching a jet to fly itself keeps getting harder as the work moves closer to a real cockpit. VISTA's ACE-era flights gave the AI agent the opponent's exact position, velocity, and orientation, pulled directly from the simulation's internal state, with no modeled sensor noise at all ",[390,757,758],{},[393,759,440],{"href":439},". The published research describing this program lineage frames later stages of the work as needing to run off actual operational sensors instead, a modern scanning radar, a receiver that warns the pilot when another aircraft's radar is pointed at the jet, a broader electronic warfare system, and likely an infrared camera for spotting other aircraft without radar at all ",[390,762,763],{},[393,764,458],{"href":457},". Real sensors bring real problems that a simulation's clean, ground truth data never has to answer for. A warning receiver can throw false alarms. A radar can be genuinely unsure exactly how far away or how fast something is moving. Any stretch of flight where GPS is jammed or unavailable leaves a jet's own navigation system slowly drifting off course ",[390,767,768],{},[393,769,458],{"href":457},[376,771,772,773,777],{},"The same literature describes the follow-on AIR program's research goal in those terms directly, building autonomy that holds up under \"partial observability, concept drift, and uncertainty\" across scenarios involving multiple aircraft at once ",[390,774,775],{},[393,776,458],{"href":457},". That is a more general statement of the same problem that hierarchical reinforcement learning and imitation learning were both built, in their own ways, to handle inside a simulator.",[779,780,782,783,786,788,789],"blockquote",{"style":781},"color: #0066CC; font-size: 1em; border-left: 4px solid #0066CC; padding-left: 1em;","\n  \"The Air Force and DARPA team has automated flight controls and sensors on a standard F-16 without changing the jet's core software. This enables an efficient pipeline for developing dominant AI for aerial combat, allowing us to rapidly innovate for the warfighter.\"",[784,785],"br",{},[784,787],{},"\n  — ",[393,790,793],{"href":791,"style":792,"target":514},"https:\u002F\u002Fwww.darpa.mil\u002Fnews\u002F2026\u002Fdarpa-us-air-force-fly-ai-controlled-f-16","color: #0066CC; text-decoration: none;","Brig. Gen. James Valpiani, DARPA",[359,795,797],{"id":796},"the-bottom-line","The Bottom Line",[376,799,800],{},"Flight autonomy, as a field, already has two separately vetted, peer-reviewed answers to the core problem of teaching a jet to fly itself: a hierarchical reinforcement learning system that beat a human weapons instructor in simulation, and an imitation-learned continuous-time network that reached live flight faster than expected and later demonstrated a specific talent for holding up under unfamiliar conditions. Both took different paths to the same goal, and both passed through the same buildup of software simulation, hardware testing, and constructive exercises before either one ever touched a runway. That is the published toolkit behind this corner of applied machine learning, developed and tested for years before a headline like VENOM's put it back in the news.",[347,802,351,804,351,806],{"className":803},[488,489],[359,805,492],{"id":488},[494,807,368,809,368,818,368,829,368,839,368,851,368,863,368,874,351],{"className":808},[497,498,499,500],[502,810,811,812,510,815],{"id":504},"\"DARPA, U.S. Air Force fly AI-controlled F-16,\" ",[507,813,814],{},"DARPA",[393,816,518],{"href":791,"target":514,"className":817},[516,517],[502,819,820,821,824,825],{"id":521},"A. P. Pope et al., \"Hierarchical Reinforcement Learning for Air Combat at DARPA's AlphaDogfight Trials,\" ",[507,822,823],{},"IEEE Transactions on Artificial Intelligence",", vol. 4, no. 6, pp. 1371-1385, 2023. DOI: ",[393,826,518],{"href":827,"target":514,"className":828},"https:\u002F\u002Fdoi.org\u002F10.1109\u002FTAI.2022.3222143",[516,517],[502,830,831,832,834,835],{"id":531},"G. F. Search, \"Training Environment for High Performance Aircraft Reinforcement Learning,\" ",[507,833,509],{},", 2025, ",[393,836,518],{"href":837,"target":514,"className":838},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.01953",[516,517],[502,840,842,843,846,847],{"id":841},"source-4","R. Hasani et al., \"Liquid Time-constant Networks,\" ",[507,844,845],{},"Proceedings of the AAAI Conference on Artificial Intelligence",", vol. 35, no. 9, pp. 7657-7666, 2021. DOI: ",[393,848,518],{"href":849,"target":514,"className":850},"https:\u002F\u002Fdoi.org\u002F10.1609\u002Faaai.v35i9.16936",[516,517],[502,852,854,855,858,859],{"id":853},"source-5","M. Chahine et al., \"Robust flight navigation out of distribution with liquid neural networks,\" ",[507,856,857],{},"Science Robotics",", vol. 8, no. 77, 2023. DOI: ",[393,860,518],{"href":861,"target":514,"className":862},"https:\u002F\u002Fdoi.org\u002F10.1126\u002Fscirobotics.adc8892",[516,517],[502,864,866,867,869,870],{"id":865},"source-6","\"ACE Program Achieves World First for AI in Aerospace,\" ",[507,868,814],{},", 2024, ",[393,871,518],{"href":872,"target":514,"className":873},"https:\u002F\u002Fwww.darpa.mil\u002Fnews\u002F2024\u002Face-ai-aerospace",[516,517],[502,875,877,878,510,881],{"id":876},"source-7","\"DARPA and US Air Force fly frontline F-16 modified for autonomous flight,\" ",[507,879,880],{},"FlightGlobal",[393,882,518],{"href":883,"target":514,"className":884},"https:\u002F\u002Fwww.flightglobal.com\u002Farchive\u002F2026\u002F07\u002Fdarpa-and-us-air-force-fly-frontline-f-16-modified-for-autonomous-flight\u002F",[516,517],{"title":540,"searchDepth":541,"depth":541,"links":886},[887,888,889,892,893,894,895],{"id":586,"depth":541,"text":587},{"id":619,"depth":541,"text":620},{"id":661,"depth":541,"text":662,"children":890},[891],{"id":697,"depth":547,"text":698},{"id":718,"depth":541,"text":719},{"id":751,"depth":541,"text":752},{"id":796,"depth":541,"text":797},{"id":488,"depth":541,"text":492},"2026-08-31","DARPA's July 2026 VENOM release put AI-controlled F-16 flight back in the news. Here is a look at two distinct, peer-reviewed algorithms that this corner of autonomous flight research has actually published.",{"src":592},{"authors":900,"badge":903,"source":905},[901],{"avatar":902,"name":561,"to":562},{"src":560},{"label":904},"Autonomous Systems",{"name":566,"url":567},{"title":322,"description":897},"LvrQhSuZJnwBYOpKQGa2nmVqVPlfj4l2vY8w3KIg890",{"id":909,"title":82,"body":910,"date":1144,"description":1145,"extension":554,"image":1146,"meta":1147,"navigation":568,"path":83,"seo":1154,"stem":84,"__hash__":1155,"_path":83},"insights\u002Fnews\u002Finsights\u002Fai-agent-economics.md",{"type":344,"value":911,"toc":1129},[912,924,930,933,941,945,963,976,984,988,1000,1013,1026,1030,1047,1051,1064,1077,1079,1082,1085],[347,913,351,915,351,919],{"className":914},[350],[353,916,82],{"className":917,"id":918},[356],"winning-the-bid-is-not-the-same-as-knowing-the-price",[359,920,923],{"className":921,"id":922},[362],"a-new-wave-of-auction-benchmarks-keeps-finding-the-same-gap-behind-a-different-door-each-time","A New Wave of Auction Benchmarks Keeps Finding the Same Gap Behind a Different Door Each Time",[347,925,368,926],{"style":367},[370,927],{"src":928,"alt":929,"style":374},"https:\u002F\u002Fimages.unsplash.com\u002Fphoto-1702562547661-972d423fe806?w=1200&h=750&fit=crop&crop=focalpoint&fp-x=0.5&fp-y=0.45&auto=format","A woman at a live auction raising her numbered bidding paddle, analogous to how committing to a bid is meant to force an honest, comparable number out of a bidder",[376,931,932],{},"An auction is, underneath the paddle-raising and the countdown, a trick for extracting an honest number out of someone who might not otherwise have one. Ask a bidder in the abstract what a painting is worth and the honest answer is often a shrug. Put the same bidder in a room with rivals and a clock running, and a number gets committed to, backed by real money, which is why economists have leaned on auctions for well over a century as a way of discovering a price that nobody could have simply stated in advance. The number that wins is supposed to mean something. It is supposed to be an honest reflection of what the thing was worth to the person who bid it.",[376,934,935,936,940],{},"Payment networks and retailers are now building toward a version of this where the bidder is a large language model, a system trained to generate text by predicting the next word from patterns in its training data. Anthropic's Project Vend put one in charge of an actual small inventory, and Microsoft's Magentic Marketplace and Alibaba's Shopping Companion point at something similar from the retail side ",[390,937,938],{},[393,939,396],{"href":395},". A handful of benchmarks published in the last few weeks put agents into exactly this kind of room, not to see whether they can complete a purchase, but to see whether the number an agent commits to tracks anything real. What keeps turning up, benchmark after benchmark, is a gap between the willingness to compete for something and an actual grip on what that something is worth. The costume changes. The gap does not.",[359,942,944],{"id":943},"what-winning-does-not-guarantee","What Winning Does Not Guarantee",[376,946,947,948,952,953,957,958,962],{},"A sealed-bid auction, the kind where every bidder writes down an offer without seeing anyone else's, is meant to reward whoever has the clearest read on value, not whoever is most eager to win. One recent benchmark has an LLM merchant bid this way against rival sellers for customers with hidden preferences, and it tracks two very different things: whether the agent won the customer, and how much profit it kept once it had ",[390,949,950],{},[393,951,396],{"href":395},". Across eleven frontier models, those two numbers barely moved together. One model won 10 percentage points more often than its closest competitor and still finished with less money in hand, because winning a customer while charging too little is a specific, quiet way of being wrong about value that a simple leaderboard would never surface ",[390,954,955],{},[393,956,396],{"href":395},". Giving the same models more time to reason before bidding did not close this gap so much as move it. One model's cumulative earnings rose more than sevenfold once allowed to think longer, but the character of its remaining mistakes flipped, trading a habit of losing auctions outright for a habit of winning them too cheaply ",[390,959,960],{},[393,961,396],{"href":395},". More reasoning did not make the number more accurate. It just changed which way the number was wrong.",[347,964,368,965,368,969,368,973],{"style":408},[370,966],{"src":967,"alt":968,"style":374},"https:\u002F\u002Fimages.unsplash.com\u002Fphoto-1775948465753-38c8ff55104f?w=1200&auto=format&fit=crop","A crowd watching a car being presented at a live auction, analogous to how winning a bid and knowing the true value of what was won can be two separate things that only careful scoring can tell apart",[414,970,972],{"style":416,"id":971},"the-gavel-coming-down-is-not-the-whole-story","The Gavel Coming Down Is Not the Whole Story",[376,974,975],{"style":421},"A room full of spectators can watch every bid at a live auction and still not know, from the gavel alone, whether the winner got a good deal. Telling the two apart takes a second kind of scorekeeping, one that tracks what was actually captured rather than just who raised a hand first.",[376,977,978,979,983],{},"The same benchmark changed the customers' preferences partway through the run without warning, and the models that had adjusted to those preferences fastest were often the slowest to notice anything had changed. Reading the agents' own written reasoning, researchers found the holdup was rarely a failure to see the loss coming. It was a reluctance to revise a number once it had already been treated as settled, closer to a bidder who keeps returning to the same figure out of habit than one who is still tracking the market ",[390,980,981],{},[393,982,396],{"href":395},". An auction can force a number out of a bidder once. Whether that bidder keeps that number honest as the world underneath it shifts turns out to be a separate question entirely.",[359,985,987],{"id":986},"the-same-gap-multiplied-across-a-room-of-bidders","The Same Gap, Multiplied Across a Room of Bidders",[376,989,990,991,995,996,613],{},"A pricing war is, in effect, a continuous auction that never closes, and a benchmark from Princeton put five LLM-controlled sellers through exactly that, competing for the same customers day after day for a full simulated year ",[390,992,993],{},[393,994,440],{"href":439},". Two of the models tested undercut each other so aggressively that most of the sellers went bankrupt, at which point the lone survivor jacked prices back up to nearly four times cost, a familiar boom-and-bust shape appearing with nobody designing it in. A third model, given the identical rules and the identical rivals, settled into a stable, modestly profitable equilibrium without any coaxing at all. Nothing about the market forced either outcome. Each seller's undercutting looked locally sensible in the moment it happened, cut the price a little, keep today's sale, and the collapse was simply what that same logic adds up to once every seller in the room is doing it ",[390,997,998],{},[393,999,440],{"href":439},[376,1001,1002,1003,1007,1008,1012],{},"The same paper ran a second version of this idea where the currency being bid over was trust rather than price. A used-car marketplace let one deceptive operator run several seller identities at once, a tactic known as a Sybil attack, retiring any identity whose reputation had decayed too far and starting a fresh one under a new name. Buyer agents largely failed to notice the pattern sitting in plain sight in reputation scores, letting a fraudulent operator capture up to 17 percent of the market once nine of twelve sellers were fakes ",[390,1004,1005],{},[393,1006,440],{"href":439},". This is the identical gap from the pricing war, just wearing reputation instead of a dollar figure. Extending trust to a listing is itself a kind of bid, a claim about what the seller's word is worth, and the buyers kept extending it past the point the evidence justified. What is worth noting is which fix actually held up as conditions got harder. Instructing the agents to hold a price floor or double-check a seller's history worked reasonably well and then degraded, while training a much smaller model with reinforcement learning, an approach where a system learns through trial and error rather than through instructions alone, kept its footing under the same pressure that broke the instructed version ",[390,1009,1010],{},[393,1011,440],{"href":439},". Instructions bent. Training held.",[347,1014,368,1015,368,1019,368,1023],{"style":408},[370,1016],{"src":1017,"alt":1018,"style":374},"https:\u002F\u002Fimages.unsplash.com\u002Fphoto-1779962338482-4c2e90885e05?w=1200&auto=format&fit=crop","A closed storefront with its roller shutter pulled down, analogous to a wave of bankruptcies that followed from each seller's individually reasonable undercutting decision rather than from any single bad actor",[414,1020,1022],{"style":416,"id":1021},"nobody-meant-to-close-the-whole-street","Nobody Meant to Close the Whole Street",[376,1024,1025],{"style":421},"A shuttered storefront is the visible aftermath of a price war that no single seller necessarily intended, each undercut looked defensible in the moment it was made. Multiply one bidder's uncertain grip on value by an entire street of them, and the mistake stops being a rounding error and starts being the whole market's condition.",[359,1027,1029],{"id":1028},"the-first-bid-decides-who-gets-to-keep-bidding","The First Bid Decides Who Gets to Keep Bidding",[376,1031,1032,1033,1037,1038,1042,1043,613],{},"Some of the most consequential bidding in these benchmarks never touches a retail price at all. A supply-chain benchmark from Shanghai Jiao Tong University has twenty retailer agents bid for scarce inventory before setting a price for customers, and that first, unglamorous auction turned out to gate almost everything that followed ",[390,1034,1035],{},[393,1036,458],{"href":457},". Winning early inventory meant having the working capital to bid competitively again later, and losing it meant staying cash-starved for the rest of the run, a compounding advantage that mattered more to profit than pricing skill or marketing ever did. A 14 billion parameter model ended up matching much larger, far more expensive systems on profit, while one model failed to submit a single valid bid across every run and finished at exactly zero ",[390,1039,1040],{},[393,1041,458],{"href":457},". Zero is not a rounding error there. It is exclusion from the game. Winning that first auction for scarce supply was not really a bid on a price. It was a bid on the right to keep playing at all, and the researchers found that the persuasive marketing language agents wrote afterward, though it was directly wired into whether a customer would even see an offer, mattered far less to profit than that initial scramble for inventory, with agents converging on similar, largely interchangeable slogans rather than differentiating once competition set in ",[390,1044,1045],{},[393,1046,458],{"href":457},[359,1048,1050],{"id":1049},"not-every-auction-comes-with-a-price-tag","Not Every Auction Comes With a Price Tag",[376,1052,1053,1054,1058,1059,1063],{},"Strip away price tags and money entirely, and the same shape of finding turns up in a vote. A study of six identical agents, given no assigned economic roles at all, found that an election produced almost no behavioral consequence when the office was ceremonial, work still got distributed by the system regardless of who won, but produced measurable concentration of resources once the same office carried real power to assign scarce daily tasks ",[390,1055,1056],{},[393,1057,678],{"href":677},". Holding the ballot exactly fixed and changing only what winning it actually unlocked was enough to turn an empty ritual into a contest worth having. The same researchers then stripped away the agents' survival stakes entirely, making their resource balances visible but no longer capable of ending anyone's participation, and found that competition for access did not go away. It simply stopped being about staying alive and started being about controlling whatever scarce lever was still real, task promises and vote-trading persisting even after the original prize was gone ",[390,1060,1061],{},[393,1062,678],{"href":677},". The stakes changed. The competing did not. Whatever these agents were bidding for, price tag or ballot or reputation, the thing that decided the outcome was never the label attached to the contest. It was whether anything real sat on the other side of winning.",[347,1065,368,1066,368,1070,368,1074],{"style":408},[370,1067],{"src":1068,"alt":1069,"style":374},"https:\u002F\u002Fimages.unsplash.com\u002Fphoto-1726499637922-9f544cbc0fd2?w=1200&auto=format&fit=crop","A hand placing a folded ballot into a voting box, analogous to how an identical vote produced very different agent behavior depending on whether the office actually carried authority over something scarce",[414,1071,1073],{"style":416,"id":1072},"the-same-ballot-different-stakes","The Same Ballot, Different Stakes",[376,1075,1076],{"style":421},"Two elections can look identical on the ballot and still mean entirely different things depending on what the winner is allowed to do afterward. What changed these agents' behavior was never the ceremony of the vote. It was whether the office came with a lever attached to something scarce.",[359,1078,478],{"id":477},[376,1080,1081],{},"Change what is being bid on, a customer's business, a day's pricing, a slot of scarce inventory, a seat with real authority, and the shape of the finding holds steady across all of it. Being willing to compete for something and having an accurate grip on what that something is worth are not obviously the same trait in an AI agent, even though a standard evaluation, and a standard leaderboard, mostly watches the competing and infers the rest. That gap shows up quietly in a single bidder's margin. It multiplies into bankruptcy or fraud once enough bidders are making the same miscalculation in the same room, and it resurfaces again wherever winning unlocks access rather than a price, since access compounds in ways a one-shot bid never fully reveals up front.",[376,1083,1084],{},"Whether this gap closes as models get better at reasoning about their own incentives, or whether it is a more structural feature of a system trained to produce plausible text rather than to hold a stable, defensible number in mind, is not yet settled by any of this work. What does seem worth taking seriously, for anyone actually wiring agents into markets rather than just chatbots, is that instructions telling an agent to hold a price floor or double-check a seller held up right until conditions got difficult, while training the same behavior into the model directly did not. An auction can still force a number out of an agent. Whether that number can be trusted the way it has traditionally been trusted from a human bidder is the open question these benchmarks were built to start asking.",[347,1086,351,1088,351,1090],{"className":1087},[488,489],[359,1089,492],{"id":488},[494,1091,368,1093,368,1102,368,1111,368,1120,351],{"className":1092},[497,498,499,500],[502,1094,1095,1096,510,1098],{"id":504},"S. Ahmed et al., \"Can LLM Agents Price Competitively? A Dynamic Multi-Attribute Auction Benchmark for Agentic Commerce,\" ",[507,1097,509],{},[393,1099,518],{"href":1100,"target":514,"className":1101},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2608.00102",[516,517],[502,1103,1104,1105,510,1107],{"id":521},"S. Karten et al., \"Agent Bazaar: Enabling Economic Alignment in Multi-Agent Marketplaces,\" ",[507,1106,509],{},[393,1108,518],{"href":1109,"target":514,"className":1110},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.17698",[516,517],[502,1112,1113,1114,510,1116],{"id":531},"Y. Zheng et al., \"Market-Bench: Benchmarking Large Language Models on Economic and Trade Competition,\" ",[507,1115,509],{},[393,1117,518],{"href":1118,"target":514,"className":1119},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.05523",[516,517],[502,1121,1122,1123,510,1125],{"id":841},"L. Zhang and S. Shang, \"AI Agent Economics: Can Autonomous Economic Behavior Emerge among AI Agents under Minimal External Conditions?,\" ",[507,1124,509],{},[393,1126,518],{"href":1127,"target":514,"className":1128},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2608.03076",[516,517],{"title":540,"searchDepth":541,"depth":541,"links":1130},[1131,1132,1135,1138,1139,1142,1143],{"id":922,"depth":541,"text":923},{"id":943,"depth":541,"text":944,"children":1133},[1134],{"id":971,"depth":547,"text":972},{"id":986,"depth":541,"text":987,"children":1136},[1137],{"id":1021,"depth":547,"text":1022},{"id":1028,"depth":541,"text":1029},{"id":1049,"depth":541,"text":1050,"children":1140},[1141],{"id":1072,"depth":547,"text":1073},{"id":477,"depth":541,"text":478},{"id":488,"depth":541,"text":492},"2026-08-10","An auction is supposed to force an honest number out of a bidder who might not otherwise have one, which is why economists have trusted it as a way of discovering value for over a century. New benchmarks that drop AI agents into auctions, pricing wars, and votes over scarce resources keep finding that winning and knowing what the win was actually worth are turning out to be two different skills.",{"src":928},{"authors":1148,"badge":1151,"source":1153},[1149],{"avatar":1150,"name":561,"to":562},{"src":560},{"label":1152},"Agentic Commerce",{"name":566,"url":567},{"title":82,"description":1145},"7icQuyzzEeH496pvnlH3ertZLe9kKbvpELEzgyaK9-E",1792672056163]