Conviva

🦨⚡ Another milestone from the Paramount Skunk Works proving grounds is now in market. On June 8, 2026, Paramount+ Lite for Samsung Tizen 2017 TVs officially launched, bringing a dramatically more efficient streaming experience to one of the most resource-constrained environments in the Connected TV ecosystem. Powered by our patent-pending, multi-tenant, multi-brand CONVERGED platform architecture and the BlackBird Playback SDK first unveiled at NAB Show, Paramount+ Lite represents a new approach to streaming application design—delivering premium experiences with a fraction of the resource footprint of traditional implementations. The same underlying codebase now serves both Paramount+ and Pluto TV, while providing a common foundation that can be deployed across Smart TVs, streaming devices, set-top boxes, and gaming consoles. By converging multiple streaming services onto a shared architecture, we're reducing complexity, accelerating innovation, and creating a scalable foundation for future Paramount streaming experiences. The results from Paramount+ Lite on Samsung Tizen 2017: 📈 Streaming Performance Index improved from 82.45 → 94.43 (+14.5% improvement) 🎬 Video Startup Time improved from 6.99s → 4.98s (29% faster playback startup) ⚡ High Startup Time Rate reduced from 12.84% → 2.29% (82% fewer slow-start experiences) Compared to the legacy Paramount+ implementation: 📉 89% Lower Memory Usage on Home screens (90 MB vs. 832 MB) 📉 58% Lower CPU Utilization on Home screens (11% vs. 26%) 📺 71% Lower Memory Usage During Live TV (200 MB vs. 691 MB) ⚡ 52% Lower CPU Utilization During Live TV (22% vs. 46%) At peak utilization: 🏆 76% Lower Peak Memory Footprint 🏆 52% Lower Peak CPU Consumption What makes Paramount+ Lite particularly exciting is that these improvements were not achieved by reducing functionality. Instead, they are the result of a fundamentally different architecture built around BlackBird, Convergence, and a multi-tenant platform strategy designed to scale across brands, devices, and operating systems. For customers, this means faster playback starts, dramatically fewer startup delays, smoother navigation, and a better overall experience on devices approaching a decade in age. For engineers, it demonstrates what becomes possible when playback, rendering, application architecture, and platform engineering are designed as a unified system. This launch validates the BlackBird + Convergence + Lite strategy we shared at NAB Show. Across Paramount+, Pluto TV, Smart TVs, streaming devices, set-top boxes, and gaming consoles, the architecture proves that performance, scale, and product velocity can coexist. Huge congratulations to the engineers, architects, product leaders, QA teams, and partners who helped bring Paramount+ Lite to market. And we're not done yet. More platforms, more optimizations, and more Skunk Works innovations are already making their way through the proving grounds and into production. 🦨⚡

Two badges. One week. One question worth sitting with. Microsoft Build and Snowflake Summit 26 — both running the same days this week in SF, close enough that I shuttled between the two. Different lanyards, same headline: the pilot era is over. Both are now selling AI in production. Both nailed two layers: → the runtime where agents execute → the governance perimeter around them (Snowflake’s Natoma acquisition was the most underrated move of the week) Here’s where I think the conversation goes next. Intelligence lives in two places — inside the model and outside it. Inside, the AI labs keep improving model capabilities. Outside is the context that makes it useful for businesses— user intents and sentiments, business data, processes, state — plus the layer that measures whether outcomes improve. That outside world is just as big an opportunity, and it belongs to the companies building on top: the ones we used to call #SaaS. Meanwhile the systems are taking on more. We’ve gone from single model calls to compound AI agentic systems — solving more sophisticated tasks, running longer, executing multi-step trajectories across many components. And that opens the gap I keep coming back to: The complexity of these systems is racing ahead of our ability to understand, learn from, and improve them. Today’s #observability tools are making real progress — but the systems are getting complex faster than the tools can keep up. The gap isn’t closing. It’s widening. We can run agents. We can govern their actions. What we still can’t do well: continuously measure whether a deployed fleet is learning and improving outcomes — or quietly drifting — at population scale, across long trajectories. That’s the wide-open space. Not the runtime. Not the perimeter. The continuous learning layer above the compound system — measure, learn, reduce drift, improve outcomes. A personal highlight from each: At Snowflake CTO Summit, Jon McNeill ’s talk on his book “The Algorithm” — his five steps (question, delete, simplify, accelerate, automate last) are really a discipline: you can’t improve what you do not understand. At Build, a sharp conversation with Jay Parikh on where Microsoft’s developer ecosystem goes next. Outside the model, the runtime is getting solved. The perimeter is getting solved. The learning layer is still up for grabs. #MicrosoftBuild #SnowflakeSummit #AgenticAI #ContinuousLearning #Context #Outcome

🎙️ 𝗘𝗽. 𝟳𝟰𝟭 𝗖𝗼𝗻𝘃𝗶𝘃𝗮 | 𝗖𝗼𝗻𝘀𝘂𝗺𝗲𝗿-𝗙𝗮𝗰𝗶𝗻𝗴 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 (𝗳𝗲𝗮𝘁. 𝗞𝗲𝗶𝘁𝗵 𝗭𝘂𝗯𝗰𝗵𝗲𝘃𝗶𝗰𝗵) — — — For episode 741 of the BlockHash Podcast, host Brandon Zemp is joined by Keith Zubchevich, President and CEO of Conviva. Keith Zubchevich is president and CEO of Conviva, where he helps digital businesses understand what their customers actually experience — not just what dashboards say is happening. He leads Conviva’s work at the intersection of agentic AI, real-time analytics, and customer experience, with a focus on measuring outcomes, friction, and risk once AI is deployed in production. Learn how Conviva gives AI agents context at conviva.ai 🔗 Spotify: https://lnkd.in/gNrVr4ya  🔗 Apple Podcasts: https://lnkd.in/gcaN_MCY  🔗 Amazon Music: https://lnkd.in/gUUZUdBq  🔗 YouTube: https://lnkd.in/giJrsC2X ⏳ Timestamps: (5:52) What is Conviva? (8:25) Agentic AI has a context issue (12:54) Conviva use-cases (15:07) Industries where Agentic AI will have the biggest impact (19:53) Agentic AI in 5 years (24:28) Likelihood of AGI (25:38) Conviva roadmap (30:29) Conviva website & socials #BlockHashPodcast #conviva #agenticai #ai #podcast #podcastshow #podcastepisode