
In Depth What nobody tells engineers about becoming a CEO | Jay Kreps (Co-founder and CEO, Confluent)
15 snips
Mar 26, 2026 Jay Kreps, co-founder and CEO of Confluent who built Apache Kafka at LinkedIn, reflects on moving from engineer to CEO and the new skillset that requires. He recounts Kafka’s rocky open-source launch and the long blog post that sparked adoption. He also explains why Confluent doubled down on a cloud-first path and the strategic choices that made that survival bet.
AI Snips
Chapters
Transcript
Episode notes
CEO Work Is A Fog Of Partial Understanding
- CEO work lives in a fog of partial understanding where you must be roughly directionally right rather than perfectly informed.
- Jay Kreps says it's impossible to know everything as orgs scale, so learn the most important things deeply and enough about others to judge them.
Aim To Know 80 Percent Of Each Function
- CEOs should know about ~80% of what each executive knows so they can judge performance and set expectations.
- Jay Kreps frames this as an aspirational practical threshold: know good work versus being the deep practitioner.
Match Go-To-Market To Customer Journey
- Go-to-market motions must match the customer's decision journey; copying another company's playbook often fails.
- Jay Kreps warns founders who will a product-led motion into existence when the product and buyer journey don't align.
