Ruben's EUC Industry Bento Box - May 2025
Welcome to the May edition of the End User Computing (EUC) Bento Box—your monthly dose of curated insights, tools, and updates from across the EUC universe.

WRITTEN BY
TABLE OF CONTENT
Whether you’re deep into VDI, exploring Cloud PC, or curious about AI’s role in modern workspace delivery, this Bento Box has something tasty for every tech appetite.
Let’s dig in.
For the Geeks
This section of the EUC Bento box is written for geeks, nerds, IT Professionals, and technology enthusiasts in general.
He Took a Chance—and Transformed VDI Forever
In this episode, Brad Peterson sits down with Harry Labana, whose career has driven innovations at Citrix, VMware, and several startups. From inventing “Farm in a Box” to co-creating App Volumes’ kernel-level mini-filter driver that spins up applications in seconds, Harry has a knack for turning customer pain points into breakthrough products. Did you know that Harry started his journey by bluffing his way into an H-1B on a New York vacation? Proof that “the more you learn, the luckier you get.” He even challenged Citrix to build a “Titanium Edition” to one-up its Platinum offering, reshaping both companies’ roadmaps. Harry also shares his simple prioritization test, “is it a headache or heart surgery?”, and a razor-sharp AI forecast.
“We’ll overestimate what AI can do in two years and underestimate what it will in ten.”
This is one of the most insightful interviews of the year. Worth your time.

OBUX: Benchmarking User Experience
At E2EMVC, Ryan, Daniel and I did announced! OBUX is a brand new community-driven framework that quantifies and compares modern compute environments' performance and user experience. OBUX delivers:
- UX Scoring: This tool generates both system-level and user-interaction scores, with fully documented formulas for transparent “apples-to-apples” comparisons across PCs, VDI, DaaS, and Cloud PC deployments.
- Flexible Deployment: It runs as a standalone performance engine or integrates into existing load-testing platforms (e.g., LoginVSI, LoadGen, EUC Score) while keeping data local or shareable for community benchmarking.
- Extensibility & Compatibility: OBUX will support plugins for AI-driven screen analytics, unified communications metrics, and more, plus multi-protocol compatibility (AWS, Citrix, Microsoft, Dizzion, Omnissa, etc.) to align with any digital workspace stack.
By establishing an open standard for UX benchmarking, OBUX supports organizations, ISVs, MSPs, and consultants to build baselines, validate improvements, and drive data-informed decisions around desktop performance and end-user satisfaction. OBUX just got started, stay tuned for more news. If you want to try OBUX now

Speaking of benchmarking, LoginVSI has just unveiled Login Enterprise powered by NVIDIA nVector. Kudos to Blair and the entire LoginVSI team. Be sure to check out their blog for all the details!

(Login Enterprise reporting example for non-GPU vs GPU test using NVIDIA nVector)
AI Workflow vs. AI Agent: Know the Difference
Are you confused about the AI terms flooding your feed? Everyone’s talking about AI agents. Most don’t get it. Here’s how it works:
Most people confuse 3 things: AI workflows, AI agents, and MCP. They sound similar, but they solve very different problems. Knowing the difference matters if you’re trying to build or use AI daily. Here is a clear breakdown
Created by Manthan Patel

- AI Workflow = Task automation
- AI Agent = Contextual decision-maker
- MCP = Model Context Protocol that glues everything together
Do you want to level Up Your AI Skills with Model Context Protocol (MCP)? Microsoft’s MCP for Beginners course is a great place to start leveling up your skills. Access here.

Building with AI: highlights for developers at Google I/O
At Google I/O 2025, Google rolled out several practical AI tools for developers:
- Gemini 2.5 Flash Preview: optimized for complex reasoning with cost controls
- Gemma 3n: an open, multimodal model that runs on phones, tablets, and laptops
- Jules: an autonomous coding agent that works through your bug backlog
- Agentic Colab: automatically fixes errors and refactors code on command
Standout update: SignGemma, an open model that translates American Sign Language into English in real time—opening the door to truly accessible applications. Read the full blog to see how these additions can streamline your AI development.

Microsoft Open Sources Windows Subsystem for Linux (WSL)
Microsoft has open-sourced the Windows Subsystem for Linux (WSL), making the entire codebase available at Microsoft/WSL on GitHub for anyone to build, fix, and extend. Read more about it here: Windows Blog. This follows WSL’s 2021 decoupling from the Windows codebase—now you’ll get faster updates and more community-driven innovation.
Key highlights:
- Complete WSL components: wsl.exe, wslconfig.exe, wslg.exe, the WSL service (wslservice.exe), Linux init/daemon binaries, and Plan9 file-sharing support
- Pre-existing open modules: WSLg (GUI/Wayland support) and the WSL2 Linux kernel have already been open sourced
- Remaining proprietary bits: The WSL 1 driver (lxcore.sys) and filesystem redirectors (P9rdr.sys, p9np.dll) stay closed to maintain core Windows integration
Special Note: This release closes out the first issue ever logged on the Microsoft/WSL repo—“Will this be Open Source?”—raised in 2016 at WSL’s debut, marking a full-circle moment in its open-source journey Windows Blog. Dive into the details and find out how to contribute at wsl.dev.
Industry News & Insights
The Desktop Has Dissolved. Where does work live in 2025?
In his latest Citrix blog, Brian Madden argues that desktops have “melted” into abstract control planes made of apps, identity, and context.
“Workspace + Worker = Work.” The real unit of productivity isn’t the device—it’s the dynamic experience.
Key takeaways:
- Workspace as abstraction: No longer tied to a single OS or machine, workspaces dynamically instantiate in VDI/DaaS/ Cloud PC, browsers, mobile apps, or on-premises instances whenever a user connects.
- Shift to context-centric management: IT must treat the workspace itself—not the device—as the primary unit of control, unifying policy at the identity source, adapting privileges in real time, and ensuring seamless cross-device experiences Citrix.com.
Special Insight: “Workspace + Worker = Work”—a workspace only truly exists when apps, identity, security, and context converge around a user.
Read the full post to rethink where work lives in 2025. Talking about Workspace and Workers, this LinkedIn post is a great read as well.

Prevent Burnout: 7 Types of Rest
You can’t grow a business when you’re running on fumes. Burnout doesn’t announce itself—it seeps in. First, it’s a creeping brain fog. Then, your replies get sharper, more irritable. Before you know it, it’s 9 p.m., you’re glued to your screen, going through the motions.
The wake-up call? Realize that rest goes beyond sleep. You need breaks for your mind, too.
Credits to Eric Partaker

The $600B AI Question
The AI bubble is reaching a tipping point. Navigating what comes next will be essential.
David Cahn’s “AI’s $600B Question” expands on his September 2023 analysis, updating the AI infrastructure vs. end-user revenue gap from $200 billion to $600 billion annually, Sequoia Capital. He shows that GPU shortages have all but vanished, hyperscalers are stockpiling hardware at historic levels, and OpenAI’s revenue has more than doubled to $3.4 billion.
Special Highlight: Microsoft alone accounted for roughly 22 percent of Nvidia’s data-center revenue in Q4—underscoring how a handful of hyperscalers are driving the AI CapEx juggernaut.
Read on to explore why closing this massive revenue shortfall will define AI’s next phase—and what it means for startups, enterprises, and investors alike.
Ever wished you could tap 34 different expert minds at once? Meet the Multidisciplinary AI Panel that turns any business plan into a 359° analysis.
Note: Why 359 degrees? There should always be a human in the loop! 😉

Europe vs. US Hyperscalers
'Close to impossible' for Europe to escape clutches of US hyperscalers?
European organizations aiming to repatriate workloads from AWS, Azure, and Google Cloud are running into steep hurdles—everything from datacenter capacity shortages and hefty egress fees to a talent pool steeped in US-hyperscaler expertise makes a mass exit “close to impossible.” The Register.
Key takeaways:
- Market concentration: AWS, Microsoft, and Google command over 70 percent of Europe’s public-cloud market share.
- Capacity crunch: At current build rates, Europe would need roughly 20 years to provision enough local cloud infrastructure to match today’s demand.
- Skill lock-in: Engineers trained on US platforms face a steep learning curve—“most would rather get a root canal than start over” on a new stack.
Special Highlight: The European Commission’s soon-to-be-published International Digital Strategy reportedly concludes that full “decoupling” from US hyperscalers is unrealistic—cementing cooperation, not replacement, as the practical path forward. The Register
Read the full analysis to explore pragmatic approaches—like hybrid sovereign clouds and multi-supplier strategies—that balance data sovereignty with operational reality.
Broadcom’s VMware Licensing Overhaul
Broadcom’s VMware licensing overhaul is imposing 800–1,500% price hikes in Europe—earning a RED rating from ECCO and sparking formal EU complaints, source.
- Mandatory three-year bundles replace pay-as-you-go, forcing customers to pay for unused software.
- Public services and hospitals report licensing costs soaring tenfold, slashing margins.
Broadcom has begun issuing cease-and-desist letters to major customers—an unprecedented legal tactic to enforce new terms.
Due to this surge, many public services are filing formal complaints and seeking multi-vendor alternatives. Now many are asking is it time to diversify?

I’m seeing growing interest from both customers and analysts in hybrid multi-cloud strategies—Dizzion DaaS included. The VMware→Broadcom transition, coupled with shifting geopolitical pressures, is pushing organizations to reassess their workspace and data-center architectures and explore alternative solutions
Inside the TU/e Cyberattack
The TU/e cyberattack of January 2025 from Fox-IT’s incident response perspective
This report from Fox-IT details the January 2025 cyberattack on Eindhoven University of Technology (TU/e) and the rapid incident response that prevented a far worse outcome .
Key points:
- Detection & escalation: SURFsoc analysts and a vigilant TU/e IT team spotted suspicious activity late on January 11 and engaged FoxCERT within an hour.
- Decisive containment: TU/e disconnected its entire network by 1:17 AM on January 12, halting the attacker’s access before ransomware could be deployed.
- Forensic findings: Investigation of over 300 systems revealed 91 touched by the attacker, but no large-scale data theft—only about 2 GB of technical files was exfiltrated.
Read on to discover the lessons learned and how rapid, coordinated action can turn a near-catastrophic breach into a case study in effective crisis management.

Bipul Sinha, CEO and co-founder of Rubrik, warns that today’s IT complexity has reached a tipping point: Microsoft alone will invest more than $72 billion in AI this year—double the inflation-adjusted cost of the Manhattan Project—while over 1,100 hyperscale data centers and billions of smartphones generate a relentless onslaught of data link. As cyberattacks surge—90 percent of IT and security leaders reported breaches last year—Sinha argues that decision-makers must either master this accelerating velocity or be crushed.
Growing IT complexity means more exposure to cyberattacks, but don’t view it as the enemy—embrace it
Bonus: AI Video Goes Mainstream
Have you seen this? AI video creation is going nuts! Google DeepMind’s Veo 3 now generates multi-second video clips with fully synchronized audio—ambient noise, dialogue, and sound effects—straight from text prompts, letting anyone spin up polished scenes in seconds. Google’s new Flow tool unifies Veo, Imagen, and Gemini models into a single AI filmmaking workflow, where you use natural language to manage cast, locations, camera moves, and scene composition within one dashboard blog.google. Meanwhile, Google is extending Gemini into a “universal AI assistant” that simulates real-world contexts, plans tasks, and interfaces across devices—integrating video understanding, screen sharing, and memory features into products like Gemini Live and Search blog.google. And in Search, AI Mode leverages a query fan-out technique plus Project Astra’s live and agentic capabilities to transform Google from a static info source into an intelligent partner—delivering deep research reports, real-time visual Q&A, and even automating tasks like ticket booking blog.google.
Check out Google DeepMind
Thanks for reading!
I am surprised that none of the Dizzion competitors are doing this …. want to take Dizzion DaaS (Frame) for a spin? Try now for instant access using Dizzion DaaS on Azure, AWS, GCP, IBM and Nutanix. Try Now
“This tough mudder travels the world spreading tokens of knowledge hidden in stroopwafel from the land of nether. Everywhere he travels, he shares information and sprouts understanding. He frames his experience in End User Computing so that others can learn the root of the technology, and what is most important in life.”
Ruben Spruijt, Field CTO & VP Product
Subscribe to our newsletter
Register for our newsletter now to unlock the full potential of Dizzion's Resource Library. Don't miss out on the latest industry insights – sign up today!

Dizzion values your privacy. By completing this form, you agree to the processing of your personal data in the manner indicated in the Dizzion Privacy Policy and consent to receive communications from Dizzion about our products, services, and events.