Loading video player...
Dash0 CEO Mirko Novakovic breaks down what actually works (and what catastrophically fails) when you try to scale AI observability from demo to production. The gap between "one trace" and "one million traces" is where most AI tools collapse, and vendors aren't telling you about it. In this segment: - Why the AI gave perfect answers on OpenTelemetry data (it had memorized the test) - The scale wall where LLMs work on 100 traces but break at 1 million - How open standards accidentally built the perfect AI foundation - What you need when AI demos stop working in production The demo-to-production gap is where most AI bets fall apart. Here's what survives contact with reality. Connect with us here: Mirko Novakovic: https://www.linkedin.com/in/mirkonovakovic/ Dash0: https://www.linkedin.com/company/dash0hq/ Guy Podjarny: https://www.linkedin.com/in/guypo/ Tessl: https://www.linkedin.com/company/tesslio/