StarZero All shows Try the podcast clipper →

May 20, 2024 · 30m · big-technology

Inside the Tech Powering Your Streaming Obsession — With Conviva CEO Keith Zubchevich

Keith Zubchevich · 22m spoken Alex Kantrowitz · 6m spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

Host Alex Kantrowitz interviews Conviva CEO and co-founder Keith Zubchevich to explore the technological infrastructure, client-side telemetry, and explainable AI powering massive live streaming events and modern digital platforms. The discussion highlights how real-time operational observability resolves critical performance bottlenecks and shapes the future of internet-delivered media.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 21.2% of the talking time here. How this is scored →

Alex as informed peer 4.0 Guest teaching 5.5 Guest disagreement 1.0 Alex pushing back 0.8
05100:0010:0020:0030:004:06–11:57 · Alex as informed peer 4/10 Bi-Directional Telemetry and Real-Time Issue Resolution Alex asks how networks triage streaming issues across millions of unique devices on the fly. Keith explains the contrast between unidirectional TV broadcasts and bi-directional IP telemetry handling trillions of potential failure combinations.11:57–18:48 · Alex as informed peer 4/10 Conviva's Origins and Expansion Across the Digital Economy Alex prompts Keith on Conviva's sixteen-year arc, drawing a comparison to long-term tech plays like Nvidia. Keith lays out their academic origins and expansion into measuring every modern digital application interface.18:49–22:53 · Alex as informed peer 3/10 Addressing Micro-Incidents and Supporting Digital Publishers Keith corrects Alex's assumption about server crashes, pointing out that total outages are rare and that micro-incidents like transient congestion cause seven times more damage to user experience.22:54–27:25 · Alex as informed peer 5/10 Applying Explainable AI to Massive Telemetry Datasets Alex asks how Conviva handles massive data volumes without falling into the trap of opaque AI black boxes. Keith details how anomaly detection AI must be paired with granular drill-down telemetry to remain explainable.4:06–11:57 · Guest teaching 6/10 Bi-Directional Telemetry and Real-Time Issue Resolution Alex asks how networks triage streaming issues across millions of unique devices on the fly. Keith explains the contrast between unidirectional TV broadcasts and bi-directional IP telemetry handling trillions of potential failure combinations.11:57–18:48 · Guest teaching 5/10 Conviva's Origins and Expansion Across the Digital Economy Alex prompts Keith on Conviva's sixteen-year arc, drawing a comparison to long-term tech plays like Nvidia. Keith lays out their academic origins and expansion into measuring every modern digital application interface.18:49–22:53 · Guest teaching 6/10 Addressing Micro-Incidents and Supporting Digital Publishers Keith corrects Alex's assumption about server crashes, pointing out that total outages are rare and that micro-incidents like transient congestion cause seven times more damage to user experience.22:54–27:25 · Guest teaching 5/10 Applying Explainable AI to Massive Telemetry Datasets Alex asks how Conviva handles massive data volumes without falling into the trap of opaque AI black boxes. Keith details how anomaly detection AI must be paired with granular drill-down telemetry to remain explainable.4:06–11:57 · Guest disagreement 1/10 Bi-Directional Telemetry and Real-Time Issue Resolution Alex asks how networks triage streaming issues across millions of unique devices on the fly. Keith explains the contrast between unidirectional TV broadcasts and bi-directional IP telemetry handling trillions of potential failure combinations.11:57–18:48 · Guest disagreement 0/10 Conviva's Origins and Expansion Across the Digital Economy Alex prompts Keith on Conviva's sixteen-year arc, drawing a comparison to long-term tech plays like Nvidia. Keith lays out their academic origins and expansion into measuring every modern digital application interface.18:49–22:53 · Guest disagreement 2/10 Addressing Micro-Incidents and Supporting Digital Publishers Keith corrects Alex's assumption about server crashes, pointing out that total outages are rare and that micro-incidents like transient congestion cause seven times more damage to user experience.22:54–27:25 · Guest disagreement 1/10 Applying Explainable AI to Massive Telemetry Datasets Alex asks how Conviva handles massive data volumes without falling into the trap of opaque AI black boxes. Keith details how anomaly detection AI must be paired with granular drill-down telemetry to remain explainable.4:06–11:57 · Alex pushing back 1/10 Bi-Directional Telemetry and Real-Time Issue Resolution Alex asks how networks triage streaming issues across millions of unique devices on the fly. Keith explains the contrast between unidirectional TV broadcasts and bi-directional IP telemetry handling trillions of potential failure combinations.11:57–18:48 · Alex pushing back 0/10 Conviva's Origins and Expansion Across the Digital Economy Alex prompts Keith on Conviva's sixteen-year arc, drawing a comparison to long-term tech plays like Nvidia. Keith lays out their academic origins and expansion into measuring every modern digital application interface.18:49–22:53 · Alex pushing back 1/10 Addressing Micro-Incidents and Supporting Digital Publishers Keith corrects Alex's assumption about server crashes, pointing out that total outages are rare and that micro-incidents like transient congestion cause seven times more damage to user experience.22:54–27:25 · Alex pushing back 1/10 Applying Explainable AI to Massive Telemetry Datasets Alex asks how Conviva handles massive data volumes without falling into the trap of opaque AI black boxes. Keith details how anomaly detection AI must be paired with granular drill-down telemetry to remain explainable.

speaking balance: gold is Alex, purple is the guest (3 minute bins)

0:00 · Alex 45.1% · guest 54.9%0:00 · Alex 45.1% · guest 54.9%3:00 · Alex 9.7% · guest 90.3%3:00 · Alex 9.7% · guest 90.3%6:00 · Alex 27.1% · guest 72.9%6:00 · Alex 27.1% · guest 72.9%9:00 · Alex 17.7% · guest 82.3%9:00 · Alex 17.7% · guest 82.3%12:00 · Alex 31% · guest 69%12:00 · Alex 31% · guest 69%15:00 · Alex 0% · guest 100%15:00 · Alex 0% · guest 100%18:00 · Alex 23.9% · guest 76.1%18:00 · Alex 23.9% · guest 76.1%21:00 · Alex 28.8% · guest 71.2%21:00 · Alex 28.8% · guest 71.2%24:00 · Alex 0% · guest 100%24:00 · Alex 0% · guest 100%27:00 · Alex 25.8% · guest 74.2%27:00 · Alex 25.8% · guest 74.2%30:00 · Alex 42.8% · guest 57.2%30:00 · Alex 42.8% · guest 57.2%
Sharpest disagreement ▶ 18:49 Keith rejects the 'crash' premise

Keith pushes back on Alex's description of site failures as crashes, clarifying that catastrophic outages are rare compared to pervasive, invisible transient bottlenecks.

 Hardest push from Alex ▶ 20:58 Alex presses on broken digital news infrastructure

Alex challenges whether this monitoring technology actually reaches struggling digital publishers whose local news pages routinely lag and break.

 Biggest teaching moment ▶ 7:25 Keith explains bi-directional streaming telemetry

Keith educates Alex on the architectural leap from passive broadcast TV decoders to millions of unique bi-directional client streams.

 Alex holds their own ▶ 22:54 Alex challenges the utility of black-box AI in telemetry

Alex demonstrates strong technical fluency by drilling into the risks of unexplainable AI models attempting to parse massive real-time data streams.
the scores for every segment, with the reasoning behind each
ChapterTopicAlex as informed peerGuest teachingGuest disagreementAlex pushing backWhy
Bi-Directional Telemetry and Real-Time Issue Resolution 4611 Alex asks how networks triage streaming issues across millions of unique devices on the fly. Keith explains the contrast between unidirectional TV broadcasts and bi-directional IP telemetry handling trillions of potential failure combinations.
Conviva's Origins and Expansion Across the Digital Economy 4500 Alex prompts Keith on Conviva's sixteen-year arc, drawing a comparison to long-term tech plays like Nvidia. Keith lays out their academic origins and expansion into measuring every modern digital application interface.
Addressing Micro-Incidents and Supporting Digital Publishers 3621 Keith corrects Alex's assumption about server crashes, pointing out that total outages are rare and that micro-incidents like transient congestion cause seven times more damage to user experience.
Applying Explainable AI to Massive Telemetry Datasets 5511 Alex asks how Conviva handles massive data volumes without falling into the trap of opaque AI black boxes. Keith details how anomaly detection AI must be paired with granular drill-down telemetry to remain explainable.

Statements from this episode (5)

Assertion Partly supported
NBC built 20 million unique viewing sessions for Peacock NFL stream
“Every one session watching that NFL game is a unique session. So NBC has to build out twenty million unique viewing sessions as opposed to broadcasting a single signal that everyone on the other side has to watch at the same, same time and in the same way.”
Keith Zubchevich May 20, 2024 ▶ 3:27
Assertion Not checkable as stated
220 trillion combinations of variables affect internet streaming quality
“There's 220 trillion different combinations of things that affect you as a consumer when you do anything on the internet. Everything from the device side, software code level, something that's going on with your battery, all the way through to connections thro…”
Keith Zubchevich May 20, 2024 ▶ 10:38
Assertion Not checkable as stated
Conviva Operates Seven Billion Sensors Measuring Global Broadcast Quality
“Now, we looked up after 16 years of measuring premium quality video and have seven billion sensors in the world today collecting quality for a myriad of premium broadcasters in real time right now.”
Keith Zubchevich May 20, 2024 ▶ 15:14
Assertion Not checkable as stated
Transient incidents impact streaming consumers seven times more than full outages
“They're looking for the little things that outages don't cause that are actually seven times more impactful to a consumer experience. Outages happening once, For every one issue caused by an outage, there's seven times the incidents, what I call them, because …”
Keith Zubchevich May 20, 2024 ▶ 19:52
Assertion Not checkable as stated
Conviva ingests four trillion telemetry events every single day
“Over the top, if you can imagine, four, we collect four trillion events a day. Every day, four trillion events come into Conviva.”
Keith Zubchevich May 20, 2024 ▶ 25:27
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 300 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more