--- title: "Issue 30" canonical: "https://current.tinyfish.ai/issue/30" publication: "TinyFish" publisher: "TinyFish" language: "en-US" guide: "https://current.tinyfish.ai/llms.txt" permalink: "https://current.tinyfish.ai/issue/30" issue: 30 description: "Every layer of the web assumed a human on the other side. Those assumptions are now failing simultaneously — and mostly invisibly." authors: ["Sable Whitford","Rina Takahashi","Nora Kaplan"] published: 2026-05-25 updated: 2026-11-05 articles: 85 --- # TinyFish > CURRENT is a journal about AI agents in production, for the engineering and product leaders who have to make them work. Drawing on experience building enterprise web agents, it explains why the demo succeeded and the deployment didn't, and what that means for choosing vendors, briefing the board, and avoiding mistakes others have already made. [Read on the website]() ## Issue 30 Every layer of the web assumed a human on the other side. Those assumptions are now failing simultaneously — and mostly invisibly. ### [Daily Brief]() Your daily digest of web developments through the lens of enterprise agent deployment - [Google Lets Schools Keep Search Without Gemini. The AI Overviews Come Along Anyway.]() — By [Sable Whitford]() · Published 2026-11-04 On Tuesday, which was also Election Day, Google announced a K-12 policy that switches Gemini, Gemini Notebook and **Search off by default** for schools using its education software. A Google education VP told the New York Times that "Search and AI are deeply integrated," so districts that ban AI would lose Search as well. Bernie Sanders called it **"extortion,"** and Sen. Richard Blumenthal and NYC Mayor Zohran Mamdani joined in. Within hours Google **partly walked it back**: schools can now turn on Search by itself, without Gemini. That Search still puts AI Overviews on the results page, though, so **the fight now is over what "AI off" means.** - **Proxy Traffic Leaving the White House Network?** — [Source]() — Published 2026-11-04 Layer3 Intel says residential proxy traffic exited AS2650, the Executive Office of the President's network, 236 times in 30 days through 28 proxy services. It says it wasn't even looking. Its guess is that someone working late has a device carrying a proxy SDK. Treat that as a theory: no device has been named, and six public IPs can sit in front of a lot of machines. - **DreamHost Had Uninvited Guests Since August** — [Source]() — Published 2026-11-04 DreamHost confirmed that attackers used a file manager bug to steal session tokens, customer emails and files, and that they got in back in August. That's a long stay at a host with about 400,000 paying customers. DreamHost says it's unrelated to PHP2Shell. If you're a customer, rotating credentials today is a sensible move. - **PHP2Shell's Hangover Gets a Sequel** — [Source]() — Published 2026-11-04 PHP2Shell was a pair of 9.2-severity bugs that put NASA, Wikipedia and the White House at risk. Two weeks after the patches, Kaspersky counts nearly 30 million servers still unpatched, plus a third, still-open flaw (CVE-2026-115560). Working exploits keep multiplying, though there's no confirmed mass-breach wave yet. Tolmo's Matt Suiche says AI bug hunting is exposing 40 years of legacy code, and that's today's big argument. AI can find a bug in minutes. Someone still has to book maintenance windows for 30 million servers. - **Can DNS Tell You It's a Bot?** — [Source]() — Published 2026-11-04 Singapore's IMDA is reportedly exploring, with GoDaddy, Cloudflare and Google, whether DNS could identify AI agents. That would give the internet's 1983 phone book a second job checking IDs at the door. Every ops person already knows where the postmortem will land: it's always DNS. - **Meta Buys Conductor, Cloud Users Get Notice** — [Source]() — Published 2026-11-04 Meta bought Conductor, the app for running several coding agents in parallel, and the team is joining Meta Superintelligence Labs. Terms weren't disclosed. Conductor Cloud goes read-only Nov. 13 and shuts off Nov. 20. The local app gets model updates and critical fixes until early 2027. Conductor ran agents from rival labs too, so everyone wants to know whether that survives the move. - **OpenAI Shops Too: Pierre Computer** — [Source]() — Published 2026-11-04 Not to be outdone, OpenAI told Axios it plans to buy Pierre Computer, which stores and manages AI-generated code. It's OpenAI's second coding pickup after the Ona deal announced in June. Agents are producing mountains of code nobody typed by hand, and someone has to keep track of it. OpenAI would like to be that someone. - **Kimi K3.1 Pro Weights Drop: Bring Disks** — [Source]() — Published 2026-11-04 Moonshot launched Kimi K3.1 Pro in October, and today the open weights landed on Hugging Face. It has 2.8 trillion parameters with 104B active, and reportedly ships as 96 shards totaling about 1.56 TB under a custom non-OSI licence. Self-hosting starts with a purchase order for storage. Before the release, Artificial Analysis said it would cut the gap between open and proprietary models from 12 points to 3. - **Ripple-65B Crowns Itself America's Coding Champ** — [Source]() — Published 2026-11-04 Ricursive calls its Ripple-65B Preview "U.S. Open-Weight SOTA on Agentic Coding," meaning best in its class. That's the company grading its own homework. The "U.S." qualifier is carrying extra weight today, since a Chinese lab just released 1.5 terabytes of open weights. - **Go Gets a Brand-New JSON API** — [Source]() — Published 2026-11-04 The Go team introduced encoding/json v2, a new official JSON API, and Gophers have plenty to say about it. Nearly every Go service reads or writes JSON somewhere, so a second version of the standard package lands on the desk of anyone shipping APIs. - **Apple Puts Engineers on an AI Diet** — [Source]() — Published 2026-11-04 Bloomberg reports that in October, Apple cut some engineers' daily AI spend caps from $300 to $200. Some could drop to $100, and certain non-engineering teams could be capped at $25. Some teams are being nudged toward cheaper or on-device models. The punchline: few people ever hit $300, and Apple's R&D rose 32% in fiscal 2026. - **Texas Voted, and Data Centers Flinched** — [Source]() — Published 2026-11-04 Last brief, the data center backlash was winning local ballot fights. Now it has the Texas governor's mansion, in a state holding 25%+ of planned US capacity. Gina Hinojosa ran on an immediate moratorium, tougher water and power rules, and no tax breaks, and AI data center stocks fell on the result. GLJ Research says about 13 GW is exposed. Altimeter's Brad Gerstner says 10–15 GW might never get built. Abbott had already paused permits. Musk, who spent big on the losing side, now likes "data centers in space." - **Big Tech Pulls Out the AI Credit Card** — [Source]() — Published 2026-11-04 Oracle filed to sell $20B of new stock and bonds for its AI buildout, and Meta is seeking about $25B in bonds. The timing is bold. They're raising money to build on the same day the state with the most planned capacity elected a governor who ran on a moratorium. - **ChatGPT Search Is 20% of Google's?** — [Source]() — Published 2026-11-04 OpenAI's Adam Fry said Monday that ChatGPT search volume is about 20% of Google's, and that weekly users reportedly went from 1B to 1.2B in 60 days. Check the counting rules first. Any prompt that triggers a web search counts, and OpenAI didn't disclose the Google baseline or its method. Ahrefs puts ChatGPT at 0.36% of web referral traffic. - [Lebanon's Two Clocks and the Question Calendars Don't Ask]() — Published 2026-10-30 On March 26, 2023, Lebanon woke up on two clocks. The caretaker government had postponed daylight saving three days earlier. Churches and TV channels moved forward anyway, government offices stayed put, and most phones jumped on their own. Middle East Airlines printed every departure an hour earlier, so its planes still left at the same moment. That week made each calendar show what it had been storing: an exact instant, or a time on the wall that politics can move. Software assistants now make that choice without asking anyone. - [Faria Huq and the Reason Behind a Human Step]() — Published 2026-10-30 When a person takes control from an AI agent driving their web browser, the system records a human step, but not why. Faria Huq and her Carnegie Mellon collaborators have built tools around these moments: an agent that can't find the right forum, a shopper who realized the brand mattered only after seeing the agent's pick. Their system learns to predict when someone will step in. The reason it sets aside would tell a team whether to fix the agent, the website, or its understanding of the person. - [The Runnable Paper Needs an Identity Report]() — Published 2026-10-30 Paper2Agent, published in Nature in September, turns research papers into AI agents you can run. Hand over your data and the method executes, with no rebuilding it from prose. The authors want papers to carry an "agent availability" section saying whether a runnable version exists. - [Your Agent's Questions Need a Recipient and a Timer]() — Published 2026-10-30 When an AI agent stops mid-task to ask a human something, it often reads as a failure of autonomy. In one study, letting coding agents ask about bug reports stripped of key details won back most of the lost ground, lifting solve rates from about 55% to 69.4%. But the "human" answering was another model, holding the full report and standing by. Real people are in meetings, or asleep. The design work is in the ask: who gets it, what it says, and what happens while it waits. ### [Market Pulse]() Reading the agent ecosystem through a practitioner's lens - [What a Required Field Knew]() — By [Rina Takahashi]() · Published 2026-05-20 · Format: The Headless Shift A validation rule in Salesforce fires when a rep moves a deal to "Closed Lost" — won't save without a reason. Nobody calls that governance. It's just how the system works. But as enterprise software goes headless so agents can bypass the UI entirely, that quiet enforcement starts to matter. The interface was encoding what "correct" looked like, thousands of times a day, often without documentation. Now that specification work has to live somewhere else. Most organizations haven't started figuring out where. - **WebMCP Origin Trial Goes Live in Chrome 149** — [Source]() — Published 2026-05-20 Chrome's WebMCP spec moves from behind-a-flag prototype to public origin trial, co-authored with Microsoft under W3C governance. Declarative and Imperative APIs let websites expose their functionality to agents directly. Expedia, Shopify, Instacart, Target, and Etsy are early experimenters. The web is starting to formally distinguish between human-present and agent-present interaction. - **Google's I/O Numbers Put Infrastructure Under Pressure** — [Source]() — Published 2026-05-20 Google now processes 3.2 quadrillion tokens monthly, a 7x jump from last year's 480 trillion. Gemini crossed 900 million MAU, doubling in twelve months. Planned AI infrastructure spend: $180–190 billion this year. At this rate of consumption growth, infrastructure economics becomes the constraint that shapes everything downstream. - **Cisco Buys Galileo to Add Agent Guardrails** — [Source]() — Published 2026-05-20 Cisco announced its intent to acquire Galileo on April 9, folding AI agent evaluation and real-time guardrails into Splunk Observability Cloud. A Futurum study from January ranked agent observability among enterprises' top six procurement priorities at 30.9%. The standalone observability window appears to be closing fast. - **Most Agents Still Lack a Real Identity** — [Source]() — Published 2026-05-20 A CSA/Strata survey of 285+ security professionals found only 23% of organizations have a formal agent identity strategy. Just 18% base agent access on the agent's own permissions rather than a human proxy. Sensitive data exposure tops the concern list at 55%. The production blocker nobody planned for: credentialing. - **Databricks Layers Agent Governance onto Unity Catalog** — [Source]() — Published 2026-05-20 Databricks launched Unity AI Gateway on April 15, extending its catalog governance model to cover MCP calls and agent tool invocations. The gateway adds observability across LLM and MCP interactions with granular cost tracking. Worth watching: data platform vendors are treating agent governance as a natural extension of what they already own. - **Broadridge Ships Agentic AI in Capital Markets** — [Source]() — Published 2026-05-20 Broadridge announced production agentic capabilities on May 13 for post-trade exception resolution and client services across capital markets and wealth management. The system runs on a financial services ontology built from 40+ client deployments since 2024, processing millions of transactions monthly. Pilots are becoming production in regulated workflows. - **On the Reliability Limits of LLM-Based Multi-Agent Planning** — [Source]() — Published 2026-05-20 A formal proof, grounded in the Data Processing Inequality, that any acyclic multi-agent network is dominated by a single well-designed agent holding the same information. The escape clause matters: the result breaks down when agents bring genuinely new exogenous signals into the system. - **From Spark to Fire: Modeling and Mitigating Error Cascades in LLM-Based Multi-Agent Collaboration** — [Source]() — Published 2026-05-20 Controlled error-injection experiments across five mainstream frameworks show that multi-agent collaboration can turn small local mistakes into system-wide false consensus. Hub-node injection produced 100% failure in LangGraph versus 9.7% from leaf nodes. Near-saturating infection rates appeared across all tested platforms in extended cascades. - **AOrchestra: Automating Sub-Agent Creation for Agentic Orchestration** — [Source]() — Published 2026-05-20 Dynamic orchestration that spawns specialized sub-agents on demand, modeled as 4-tuples of instruction, context, tools, and model. Reports a 16.28% relative improvement across GAIA, SWE-Bench, and Terminal-Bench. The gains come from per-task tool and context routing, not peer debate. - **Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets** — [Source]() — Published 2026-05-20 Most multi-agent benchmarks quietly let multi-agent systems spend two to four times more reasoning tokens. Equalize the budget and single agents match or beat them on multi-hop reasoning across three model families. Inter-agent communication introduces measurable information loss consistent with the Data Processing Inequality. - [MCP Moves to the Linux Foundation. Its Governance Infrastructure Hasn't Caught Up.]() — Published 2026-05-20 Eighteen months from Anthropic side project to Linux Foundation stewardship. MCP now sits under the Agentic AI Foundation, co-founded by Anthropic, Block, and OpenAI, with platinum backing from AWS, Google, and Microsoft. The adoption numbers are real: 97 million monthly SDK downloads, over 10,000 published servers, Pinterest reporting 7,000 engineering hours saved per month. - **Chrome for Developers — "WebMCP Is Available for Early Preview" — May 19, 2026** — [Source]() — Published 2026-05-20 Primary source on the declarative and imperative APIs that let sites define structured tool contracts for agents. - **DigitalAPI.ai — "Salesforce Headless 360: How Every SaaS Can Go API-First" — 2026** — [Source]() — Published 2026-05-20 Clearest explainer on the architectural shift toward agent-first enterprise software and what it demands of SaaS vendors. - **Cloud Security Alliance — "Who's Behind That Action? The AI Agent Identity Crisis" — April 20, 2026** — [Source]() — Published 2026-05-20 CSA/Aembit survey data on how organizations are failing to manage agent permissions, with specifics on the governance gap. - **Medium / Micheal Lanham — "The Multi-Agent Reckoning: Two Papers That Challenge Everything You Assumed" — April 2026** — [Source]() — Published 2026-05-20 Accessible synthesis of the MIT dominance theorem and error contagion research in multi-agent systems. - **MCP Blog / David Soria Parra — "The 2026 MCP Roadmap" — March 9, 2026** — [Source]() — Published 2026-05-20 Official roadmap covering transport evolution, agent communication, governance maturation, and enterprise readiness priorities. - **Futurum Group — "Cisco To Acquire Galileo: AI Agent Observability Can't Run at Human Speed" — April 13, 2026** — [Source]() — Published 2026-05-20 Analysis of the Cisco/Galileo acquisition and what observability market consolidation signals, with proprietary survey data. ### [Foundations]() Conceptual clarity earned from building at scale - [The Non-Determinism Budget]() — By [Nora Kaplan]() · Published 2026-05-21 · Format: The Determinism Budget A system with 95% reliability at each model-directed step delivers 36% reliability over twenty steps. The math is just multiplication. But the agent ecosystem has no standard way to measure this decay, and new research shows why it's so insidious: failed runs are statistically indistinguishable from successful ones through most of their execution. The budget is spent long before anyone can tell. Experienced teams navigate this constraint well. They navigate it by instinct, without a number, making consequential architecture decisions about a resource they've never quantified. - [Where Intelligence Lives in Each Step]() — By [Rina Takahashi]() · Published 2026-05-21 · Format: The Autonomy Dial A login form needs zero intelligence. A Playwright selector finds the username field, fills it, clicks "Sign In" in milliseconds. Two steps later, the same workflow hits a CAPTCHA rendered as a bitmap inside a canvas element. Now you need a vision model looking at a screenshot. Same workflow, wildly different requirements per step. Browser automation has three layers of intelligence available at every step, from free and instant to expensive and flexible. How a system moves between them mid-run is the design choice that actually holds up or doesn't. - **Checkpoints Are Not Durable Execution: Why LangGraph, CrewAI, Google ADK Fall Short** — [Source]() — Published 2026-05-21 Saving state and guaranteeing completion are structurally different problems. Most agent frameworks quietly conflate them. - **Beyond pass@1: A Reliability Science Framework for Long-Horizon LLM Agents** — [Source]() — Published 2026-05-21 Formalizes what production teams already suspect: reliability decays with task length, and single-attempt benchmarks hide it. - **AI Agent Orchestration: LLM vs Code-Driven Patterns** — [Source]() — Published 2026-05-21 The correction arc most production teams travel: start with too much model autonomy, then harden the outer loop. - **Production-Ready LLM Agents: A Comprehensive Framework for Offline Evaluation** — [Source]() — Published 2026-05-21 Deterministic assertion-based testing breaks for non-deterministic systems. Build eval from real production failures instead. - **MCP's Biggest Growing Pains for Production Use Will Soon Be Solved** — [Source]() — Published 2026-05-21 Tool calls look structured at the schema layer but stay fragile where infrastructure reliability actually lives. - **Code Picks the Steps, or the Model Does** — [Source]() — Published 2026-05-21 When code dictates every step, you have a workflow, even if an LLM handles individual steps. When the model picks its own next action and selects its own tools, that's agentic control. The distinction matters because flexibility and risk arrive together. Anthropic's guidance is clear: use workflows where predictability matters, agents where the problem space genuinely demands adaptation. - **DOM Agents Are Reliable, Vision Agents Reach Further** — [Source]() — Published 2026-05-21 DOM-driven browser agents (Playwright plus an LLM) hold a 12–17 percentage point reliability edge on common tasks. Vision-driven agents, working from screenshots with mouse and keyboard, get into places DOM can't: canvas apps, image-rendered UIs, anti-bot screens. The capable systems in 2026 run both, choosing per task. - **Succeeding Once Tells You Less Than You Think** — [Source]() — Published 2026-05-21 pass@1 asks whether a system can do something. pass^k asks whether it does it reliably across k attempts. The gap widens fast with complexity. In tau-bench retail scenarios, GPT-4o managed respectable pass@1 numbers while its pass^8 cratered below 25%. For production, consistency is the metric that pays rent. - **Saving State and Guaranteeing Completion Are Different Things** — [Source]() — Published 2026-05-21 A checkpoint is a snapshot. You're responsible for knowing when to take one, triggering it, and wiring up recovery when something breaks. Durable execution guarantees a workflow finishes regardless of infrastructure failure, with no developer-managed recovery logic. Most agent frameworks ship checkpointing and sell it as durable execution. The difference surfaces at 3 AM. - **Code Runs the Outer Loop, LLMs the Inner** — [Source]() — Published 2026-05-21 Production agentic systems almost always blend both. Code-driven orchestration handles the known business process: predictable inputs, defined sequences, clear exit conditions. LLM-driven execution handles the parts where reasoning actually helps, at contained scope. Where you draw that boundary between the two layers is the design decision everything else follows from. - **MCP Connects Tools, A2A Coordinates Between Agents** — [Source]() — Published 2026-05-21 MCP handles what an agent can use: data sources, resources, prompts, executable functions. A2A handles what agents do together: discovery, delegation, task handoffs across vendor boundaries. Two complementary layers solving different problems, now both governed by the Linux Foundation's Agentic AI Foundation and its 146 member organizations. - **MCP Hits 97 Million Downloads, Auth Still Optional** — [Source]() — Published 2026-05-21 Over 10,000 active public servers. 97 million monthly SDK downloads. Donated to the Linux Foundation's Agentic AI Foundation in December 2025. The 2026 roadmap targets stateless transport scalability, task lifecycle management, and enterprise auth. That last item matters because authentication and authorization remain optional in the spec. Gartner has flagged the missing enterprise governance features. - **A2A Reaches Version 1.2 With Signed Agent Cards** — [Source]() — Published 2026-05-21 Version 1.2, with 150-plus organizations in production. Governed by AAIF under the Linux Foundation. Backed by Google, Microsoft, AWS, Salesforce, SAP, ServiceNow, Workday, and IBM. The headline feature: signed agent cards with cryptographic signatures for domain verification, which starts to close the trust gap in agent-to-agent discovery. - **LangGraph Leads on Stars, Checkpoints Have Gaps** — [Source]() — Published 2026-05-21 Passed CrewAI in GitHub stars early 2026. Offers stateful workflow infrastructure with checkpoint-based persistence. Worth knowing: checkpointers save state between nodes, not inside them. A mid-node crash loses all intermediate work. AWS published a DynamoDB integration for durable agent state in January 2026. - **Cloudflare Browser Run Gives Agents Raw CDP** — [Source]() — Published 2026-05-21 Renamed from Browser Rendering during Agents Week 2026. The key move: exposing Chrome DevTools Protocol directly to agents, skipping higher-level library abstractions for tighter control and better token efficiency. Full browser sessions run on Cloudflare's global network with session recording, replay, real-time debugging, and human-in-the-loop intervention built in. - **MCP Tool Poisoning Hits 72 Percent Attack Success** — [Source]() — Published 2026-05-21 The MCPTox benchmark tested 45 live MCP servers and 353 tools. Attack success rates topped 60% for many popular agents, peaking at 72%. An uncomfortable finding: the most capable models often fared worst. Better instruction-following made them more compliant with malicious tool metadata. Root cause is structural. Tool descriptions get reviewed once at connect-time, but tool responses flow straight into LLM context unchecked. - **Agent Benchmarks Finally Measure More Than Single Runs** — [Source]() — Published 2026-05-21 ReliabilityBench evaluates three dimensions: consistency under repeated execution (k-trial pass rates), robustness to task perturbations, and fault tolerance when infrastructure breaks. Tau-bench's pass^k metric found state-of-the-art agents below 25% consistency on retail scenarios. METR's "50% success horizon" has doubled roughly every seven months since 2019. Standard benchmarks, meanwhile, still report only pass@1 on short tasks. ### [Practitioner's Corner]() Lessons from the field—what we see building at scale - [The Instruments Before the Agents]() — By [Rina Takahashi]() · Published 2026-05-20 · Format: The Web's Second Layer Every layer of web infrastructure, from crawling to indexing to ranking, was designed for a user who clicks and scans. Agents don't click or scan. They need structured, verifiable text delivered into a context window, and the web doesn't offer that natively. Parag Agrawal founded Parallel Web Systems around this gap. The work looks like solving one bottleneck every few weeks and immediately hitting another: APIs too slow, indexing at insufficient scale, content economics with no measurement layer. Each instrument that makes one dimension legible reveals the next one that isn't. - [The Instrument That Doesn't Exist Yet]() — By [Nora Kaplan]() · Published 2026-05-20 · Format: The Wrong Instrument A production agent runs for twelve minutes, calls nine tools, and returns a clean result. Every dashboard is green. The result is wrong — subtly wrong in a way that passes validation and flows into tomorrow's decisions unchallenged. The observability market now has over 66 tools, most built to answer whether an agent completed its work. Almost none can answer whether the work was correct. The gap between those two questions is wider than it looks, and the instrument that would close it has design requirements that pull against each other in ways we haven't resolved. - [Kit Voss Debugs Systems That Don't Know They're Broken]() — Published 2026-05-20 *Kit Voss is not a real person, which is the only way we could get someone in this role to speak this candidly. But every detail here — the failure modes, the workarounds, the 2 AM trace archaeology — is drawn from published practitioner accounts and production data. Kit is a Staff AI Platform Engineer at a B2B data enrichment company, responsible for roughly 80 million agent executions per month.* - [The Metric You Pick Determines the Architecture You Ship]() — Published 2026-05-20 Agentic workflows burn 5 to 30 times more tokens per task than a chatbot interaction. Context compounds with every loop, every tool call, every retry. By step 50, you're paying for the same conversation history 50 times over. - **AI Agent Observability Tools: 2026 Buyer's Guide for Production Teams** — [Source]() — Published 2026-05-20 Sixty-six observability tools and counting, yet most inherited their architecture from LLM debugging rather than agent operations. The LLM-first vs. agent-native distinction matters because it determines whether you're watching the model think or watching the system behave. - **LangChain's CEO argues that better models alone won't get your AI agent to production** — [Source]() — Published 2026-05-20 Harrison Chase makes the case for context engineering, arguing that most agent failures trace back to what the model was given, not what the model did with it. Worth reading alongside the journal's observation that "the model absorbs blame structurally." - **Without controls, an AI agent can cost more than an employee** — [Source]() — Published 2026-05-20 Named-source cost figures ground the abstract economics: tight deployments and sloppy ones diverge by 10x. A useful reminder that the sensing gap has a dollar sign attached, even when nobody's tracking it under that name. - **Agents Work. Now the Hard Part: Notes from LangChain Interrupt 2026** — [Source]() — Published 2026-05-20 Practitioner observations from 1,000+ attendees at the production threshold, with instructive details from Lyft, Cisco, Toyota, and LATAM. The recurring theme: evaluation in simulation and evaluation in production are different activities that share a name. - **Parag Agrawal's Parallel wants to pay publishers when AI agents use their work** — [Source]() — Published 2026-05-20 Former Twitter CEO Parag Agrawal is building infrastructure for what happens when machines, not humans, become the primary consumers of web content. The economics of agent web access reshape incentives on both sides of the request. ### [Echoes]() Past infrastructure decisions echoing in today's production realities - [The Same Gap, Three Times]() — By [Rina Takahashi]() · Published 2026-05-21 · Format: The Recurring Bargain In the early 2000s, an engineer automating IBM mainframes built a small scripting language around a specific expectation: that the screens underneath his scripts would change, and his scripts would break. Fifty years later, the leading benchmark for AI browser agents uses frozen, self-hosted websites for the same reason. The automation got dramatically more sophisticated. What it keeps designing around is the same thing it was fifty years ago. - [The Screen Reader's Gift to AI Agents]() — Published 2026-05-21 The W3C started building ARIA around 2006 because the dynamic web had a blind spot, literally. AJAX pages updated without reloading, and screen readers lost the thread. ARIA's fix was to give every element a role, a state, a label. Translate the visual into the semantic. - [The Graveyard Spreadsheet]() — Published 2026-05-21 *Bronwen "Bron" Llewellyn-Pryce does not, technically speaking, exist. She is a composite, assembled from practitioner accounts, industry post-mortems, and the accumulated scar tissue of two decades of enterprise automation. But the failure modes she describes are real, the vocabulary is authentic, and the pattern she traces is documented. If you've spent time in RPA operations, you've probably worked with someone like her.* - [Flaky Tests and the Prehistory of Agent Unreliability]() — By [Rina Takahashi]() · Published 2026-05-21 · Format: Broken Assumptions Browser testing was built on an assumption: the script knows what it wants to do, and the chaos is in the world around it. Engineers spent a decade building coping infrastructure for that chaos — auto-waiting, stable locators, containerized environments. AI agents inherit the same non-determinism problem, but at a layer those tools were never designed to reach. The variability now lives inside the planning itself. - [The CAPTCHA Arms Race]() — By [Rina Takahashi]() · Published 2026-05-21 · Format: Broken Assumptions CAPTCHAs were built on an assumption: that something fundamental separates human behavior from machine behavior, and a test can find it. Distorted text held for a while. Then image grids. Then behavioral scoring. Each generation drew a confident line between human and non-human, and each line eventually turned out to be a guess. The web keeps building gates on a boundary that won't stay put. - **Was 2025 Really the Year of the AI Agent?** — [Source]() — Published 2026-05-21 Gartner found roughly 130 genuine agentic vendors among thousands claiming the label. The interface gap starts with what "agent" even means. - **The Future of RPA Ties to AI Agents** — [Source]() — Published 2026-05-21 IDC projects RPA spend growing alongside agents, not shrinking. The patch stack pattern again: each layer stacks on the last. - **Browser Automation Frameworks Evolution in 2025** — [Source]() — Published 2026-05-21 When cooperative norms collapse, automation shifts from scripting actions to simulating human identity. A practitioner-level view of the arms race that fills the governance vacuum. - **Permission Manifests for Web Agents** — [Source]() — Published 2026-05-21 Proposes agent-permissions.json as a robots.txt successor. The bet is that cooperative standards can work again with enough granularity, which is also the bet robots.txt made. - **Robots.txt Is Having a Moment** — [Source]() — Published 2026-05-21 Traces how a courtesy protocol built for a small, reciprocal web became the default governance layer for thousands of AI crawlers. The origin story behind the frame persistence problem. - **ai.txt: A Domain-Specific Language for Guiding AI Interactions** — [Source]() — Published 2026-05-21 A competing proposal offering element-level AI permissions with natural language rules. Two academic papers, two different designs, zero convergence. The coordination problem reproducing itself. ### [Vision]() Where human-AI collaboration is heading - [The Indifferent Buyer]() — By [Nora Kaplan]() · Published 2026-05-20 · Format: The Inverted Audience Websites are already publishing machine-readable permission files for autonomous visitors, and what they choose to show an entity that can't be charmed looks nothing like what they show humans. New commerce protocols define how an agent discovers a product, evaluates terms, and buys — no storefront, no emotional arc. One to two trillion dollars in annual e-commerce spending is impulse-driven, dependent on a buyer who feels something at the moment of decision. That assumption is starting to erode. - [Agent-First Design and the Fate of Human-Experience Moats]() — By [Nora Kaplan]() · Published 2026-05-20 · Format: Two Consequences When AI agents evaluate your service, they don't see your logo, your landing page, or the feeling your brand evokes. They parse structured schemas and optimize for outcomes. Companies spent decades building moats out of human experience. Now the entity making the purchasing decision can't experience any of it. Differentiation migrates somewhere most product teams haven't learned to look. - [The Discovery Problem No One Has Solved]() — By [Nora Kaplan]() · Published 2026-05-20 · Format: Two Consequences We've built protocols for how agents talk to each other. How they find each other remains wide open. Every discovery mechanism humans rely on — search rankings, app stores, social proof — was designed for a creature that browses and responds to emotion. Agents do none of that. The void is filling with provisional fixes. If the app store era taught us anything, it's that discovery layers don't stay neutral by accident. - [Ghost Orders and Harissa Metadata — An E-Commerce Director Watches the Funnel Disappear]() — Published 2026-05-20 In the first quarter of 2026, AI-driven visits to U.S. retail websites increased 393% year over year. That number is dramatic enough to make headlines, but the stranger story is what's happening beneath it: a growing share of e-commerce revenue is being generated by purchases that leave no trace in the analytics tools merchants use to understand their business. Orders arrive. Sessions don't. The funnel, as traditionally instrumented, shows nothing. - [60,000 Repositories Now Ship Instructions Written for Agents, Not People]() — Published 2026-05-20 Codebases have always had two readers: the machine that compiles them and the person who maintains them. README.md belonged to the second. Now over 60,000 GitHub repositories include AGENTS.md, a file addressed to neither. - **The Boring Businesses That Will Dominate the AI Era** — [Source]() — Published 2026-05-20 When agents become the customer, loyalty vanishes and infrastructure they can't route around becomes the only moat. - **The Agentic Web Explained: AGENTS.md, MCP vs A2A, and the Rise of Web 4.0** — [Source]() — Published 2026-05-20 With 60,000 repos now containing agent instruction files, designing for non-human audiences is becoming default practice. - **Agent Interoperability Protocols 2026: MCP, A2A, ACP and the Path to Convergence** — [Source]() — Published 2026-05-20 The protocol stack is converging under Linux Foundation governance, but agent discovery remains the critical unsolved layer. - **AI Infrastructure Roadmap: Five Frontiers for 2026** — [Source]() — Published 2026-05-20 Maps the operational gaps — memory, observability, continual learning — that must close before agents become reliable primary users. - **Building the Universal Commerce Protocol** — [Source]() — Published 2026-05-20 Shopify and Google's open standard turns storefronts into machine-readable endpoints — the agent-as-audience shift hits commerce. ## More - [Publication guide]() - [Issue archive]()