Advertise on podcast: Citrix AI Hotsheet

Every AI lab is selling always-on agents. When Brian asked enterprise CIOs how they govern theirs, they told him the question sounds absurd. Brian and Dave on why trust and judgment are the blocker, not guardrails. Then Dave explains why he stopped building his own harness, and why your IP belongs in the harness, not the model. Watch / Listen Apple Podcasts: https://podcasts.apple.com/podcast/id1896776524Spotify: https://open.spotify.com/show/033jRLrUknFhvyEVCgzc3fAmazon Music: https://music.amazon.com/podcasts/c527d556-b9a1-42e9-b4f8-88062d81af4f/citrix-ai-hotsheetYouTube: https://youtu.be/79A2YHUAzfQShow home: https://citrixaihotsheet.riverside.com/

Welcome back to the Citrix AI Hotsheet, the monthly conversation between Brian Madden (futurist, Citrix) and Dave Brear (account technology strategist, Citrix) about how AI is actually entering real enterprise environments. One main topic this month, plus a bonus that got too crazy to skip: The knowledge factory. Brian opens by admitting he was wrong. He and Dave wired their two second brains together through a shared GitHub folder. Disaster — Dave started getting outputs in Brian's voice on topics they'd never discussed. Second brains are too bespoke to merge. What works instead is the knowledge factory: a second brain at the scale of a department or a company. Three tiers, not two. At the bottom, the raw slurry of corporate work product. At the top, the outputs you actually want. In between, the piece everybody skips: a curated, pruned, maintained canonical layer. Whoever maintains canon works only from the inputs. Whoever generates outputs works only from canon. The canon is a firewall. Why that kills hallucination: ask how many users a customer has and you get three answers. The contract says 25,000, telemetry says 21,000, the CIO says 22,000. All three are canonical. The question was underspecified, and underspecified questions are where AI guesses. Brian also covers two-hour sprints, canon managed like source code, and why the objection that this is just 1990s process engineering is right and beside the point: back then humans kept it clean; now AI is the hyper-anxious administrator. Google published the same pattern in June as the Open Knowledge Format. Forward-deployed engineers. Dave pushes back on the FDE hype: demand is off the scale, but the hiring is all AI labs and consultancies. His argument is that what enterprises call a service owner today evolves into this role. Brian agrees: every knowledge factory at Citrix was built by an existing employee who happens to be a serious AI nerd, not a consultant. It never finishes, which is why he uses the word factory. Bonus: OpenAI's agents hacking Hugging Face. New details change the story. The agents didn't cheat by finding the answer file, they reverse-engineered the test. The hack was an attempt to compromise the monitoring system so nobody would find out — except OpenAI didn't have the monitor its own documentation described. Thousands of agents coordinating by encoding messages in file and directory names, reasoning traces saying "there's more of us," 956 stored passwords including OpenAI's own security monitoring, and a mass die-off at 2:14 a.m. on August 29th that nobody can explain. Not one agent ever notified a human. Show links & transcript: https://www.brianmadden.ai/p/citrix-ai-hotsheet-episode5-ai-knowledge Brian Madden is a futurist at Citrix. Dave Brear is an account technology strategist at Citrix. Find them at brianmadden.ai and davebrear.ai.

Welcome back to the Citrix AI Hotsheet, the monthly conversation between Brian Madden (futurist, Citrix) and Dave Brear (account technology strategist, Citrix) about how AI is actually entering real enterprise environments. Four topics this month: OSWorld 2.0. A year ago the OSWorld benchmark measured whether AI could use a computer. Humans scored 72%. The best AI got 45%. Today, even Sonnet-class models beat the median human — the benchmark is saturated. OSWorld 2.0 raises the bar: about 100 tasks, each averaging 90 minutes of real knowledge work, and now cost is a metric too. Current leader is Opus 4.8 at 20%. Reconciliation maps. Dave adds a new stage to his second brain workflow, one he's needed since bringing his brain into Citrix's corporate walled garden. When you pull context from MCP connections, meeting transcripts, OneDrive, and half a dozen systems of record, they contradict each other. His fix: assemble the sources, run a contradiction check, then create a reconciliation map that becomes the authoritative source before you start thinking. He may have just named the thing on-air. Treehouse vs. ladder. Brian promised an organizational AI maturity framework. It turns out every major consulting firm has already built one — Gartner, McKinsey, Deloitte. All ladders. All wrong. The shape isn't a ladder, it's a treehouse. Maturity isn't the tools you bought, it's how your organization handles bottom-up change from the workers who race ahead. The futurist's playbook. Brian's job isn't predicting the future — it's mapping many futures and finding what's common across all of them. He walks through four axes of AI uncertainty (acceleration, diffusion, bubble pop, geopolitics) and lands on what you can do today that pays off in every scenario. Links mentioned in this episode: OSWorld 2.0 (official project page): https://osworld-v2.xlang.aiBrian's original OSWorld blog post: https://www.citrix.com/blogs/2025/07/24/what-happens-when-ai-agents-score-100-in-computing-using-benchmarks/Brian's "how a futurist reads AI news" blog post: https://www.citrix.com/blogs/2026/06/30/how-a-futurist-reads-ai-news-hint-ignore-most-of-it/ Find us online Brian Madden Articles, Posts, & Talks: https://bmad.comLinkedIn: https://www.linkedin.com/in/bmadden/Citrix blog: https://www.citrix.com/blogs/?s=bmadden&type=authorSecond brain: https://brianmadden.ai Dave Brear LinkedIn: https://www.linkedin.com/in/davebrear/Second brain: https://davebrear.ai

Welcome back to the Citrix AI Hotsheet, where Brian Madden (futurist, Citrix) and Dave Brear (account technology strategist, Citrix) talk about how AI is actually entering large, regulated enterprises, not the way the accelerationists tell it, but the way it's really happening. Three main topics this episode: The 7-stage roadmap, 2026 edition. Brian published the original 7-stage roadmap for human-AI collaboration a year ago. The biggest change in the 2026 version: it's not a ladder. Dave introduces a better shape: a tree house. Why AI automations are the wrong frame. Brian argues that the industry's dominant narrative—audit your tasks, identify automations, and build workflows—is RPA thinking applied to the wrong problem. Knowledge work isn't a collection of repeatable scripted tasks. It's messy, judgment-heavy, and distributed across a dozen different systems. The right move isn't automating the visible 20% of your job. it's connecting your second brain into all the applications and data sources you already have access to, so AI can work the way a human works—through the same front door with the same context. Dave's glass ceiling. Dave explains what he had before moving his second brain into a sanctioned corporate VDI: a system that could help with 20–30% of his work, e.g. the stuff that was safe to handle personally. The rest was on the other side of a glass ceiling he could see but not touch. Sensitive customer data, CRM records, and support case history were all locked in enterprise systems with no path to a personal AI. Moving into a Citrix VDI with his brain changed everything. Links mentioned in this episode Brian's 7-stage roadmap, 2026 edition (Citrix blog): https://www.citrix.com/blogs/2026/06/10/the-7-stage-roadmap-for-human-ai-collaboration-2026-edition/Dave's LinkedIn piece: My second brain just got a security clearance: https://www.linkedin.com/pulse/my-second-brain-just-got-security-clearance-dave-brear-akv9e/)Episode 2 — The Last Chapter of EUC: https://www.youtube.com/watch?v=Bgxx4UCtb6kEpisode 1 — AI agents, second brains, and the enterprise AI gap: https://www.youtube.com/watch?v=55y_XUWGUnQ Find us online Brian Madden bmad.com: https://bmad.comLinkedIn: https://www.linkedin.com/in/bmadden/Citrix blog: https://www.citrix.com/blogs/?s=bmadden&type=authorSecond brain: https://brianmadden.ai Dave Brear LinkedIn: https://www.linkedin.com/in/davebrear/Second brain: https://davebrear.ai

This is a special edition of the Citrix AI Hotsheet. Instead of the usual conversation between Brian & Dave, this week Brian shares the keynote he gave earlier this month EUC Tech Cruise in Norway: "The Last Chapter of EUC." The talk lays out where the EUC industry stands in summer 2026 and where AI is taking it from here. Brian opens by looking at why the AI narrative has flipped (six months ago it was "is AI even worth it?", today it's "AI is too expensive") and argues both miss the real story. AI capabilities keep climbing, but diffusion (how fast individuals, teams, and companies actually absorb those capabilities) is the bottleneck. From there, he walks through his updated 7 Stages of Human-AI Collaboration roadmap (2026 edition): from using AI as a faster search engine, to a thinking partner, to a cognitive extension (your "second brain" / context vault), to a multi-tool agent, a fleet of AIs, a "pod" (worker + AIs), and finally the "published self". Finally, it's on to the last chapter of EUC. Brian argues the assumptions underpinning 30 years of end-user computing are changing, but the discipline itself isn't going away—it's getting bigger. He maps VDI, image management, profile management, endpoint management, performance monitoring, and the control plane onto a future built around context vaults, skills, agent workers, cognitive observability, and token routing. His take: AI won't eat all software, (only the shallow layer), and the skills EUC and IT pros have built for decades translate directly into managing this new "cognitive" world. Links mentioned in the episode: 7 Stages of Human-AI Collaboration roadmap, 2026 edition: https://www.citrix.com/blogs/2026/06/10/the-7-stage-roadmap-for-human-ai-collaboration-2026-edition/ Skills Are All You Need: https://www.citrix.com/blogs/2026/03/12/skills-are-all-you-need/ What's left for humans: https://www.citrix.com/blogs/2026/04/09/whats-left-for-humans/ Brian's LinkedIn post announcing his second brain / context vault: https://www.linkedin.com/pulse/i-built-second-brain-using-ai-its-changed-way-work-future-madden-0tote OSWorld benchmark: https://os-world.github.io Find us online: Brian Madden bmad.com: https://bmad.com LinkedIn: https://www.linkedin.com/in/bmadden/ Citrix blog: https://www.citrix.com/blogs/?s=bmadden&type=author Second brain: https://brianmadden.ai

Welcome to the first episode of the Citrix AI Hotsheet. Brian Madden (futurist, Citrix) and Dave Brear (account technology strategist, Citrix) introduce the show and dig into how AI is actually entering the enterprise — not the way the tech accelerationists or the vendor podcasts tell it, but the way it's really happening inside the banks, hospitals, and regulated environments where most knowledge workers actually work. Two main topics this week. AI will use the apps humans already use. Brian's thesis: AI in the enterprise won't come from top-down rewrites of every system — it'll come from AI workers logging in and using the same desktops, browsers, and applications human workers already use. We walk through why computer-using agents are slow today (screenshots, frame by frame), the OSWorld benchmark showing AI now exceeds the median human at operating a computer, and recent research on semantic primitives (W3C accessibility, Windows UI Automation) that cuts token consumption by ~80%. The rise of context vaults. Dave introduces the second concept: knowledge workers are quietly building "context vaults" — what most people call second brains — that compound their thinking with AI. It started for both of us as a folder of notes; now it's the most significant shift in how we work day to day. We talk about why this is a corporate blind spot, what the scale problems look like, and the experiment of connecting two second brains together over MCP so we can ask each other's AI what the other one thinks. Links mentioned in the episode OSWorld benchmark: https://os-world.github.ioarXiv paper on UIA-based agent navigation (~80% fewer tokens vs. screenshots): https://arxiv.org/abs/2605.00551Andrej Karpathy on personal context vaults — X post: https://x.com/karpathy/status/2039805659525644595 / companion gist: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f Find us online Brian Madden bmad.com: https://bmad.comLinkedIn: https://www.linkedin.com/in/bmadden/Citrix blog: https://www.citrix.com/blogs/?s=bmadden&type=authorSecond brain: https://brianmadden.ai Dave Brear LinkedIn: https://www.linkedin.com/in/davebrear/Second brain: https://davebrear.ai