The a16z Show AI Can Write the Code. Can It Run the System? | Grepr CEO Jad Naous
203 snips
Oct 16, 2026 
Jad Naous, CEO and co-founder of Grepr, is a systems engineer focused on observability and reliability. He explores what happens when AI moves beyond writing code to running production systems. Topics include AI-powered SREs, debugging distributed systems, coding agents, telemetry costs, security, query optimization, and why top engineers may benefit most from AI.
AI Snips
Chapters
Transcript
Episode notes
Observability Is An Expensive Proxy For Reliability
- Companies spend 5–20% of infrastructure budgets on observability because they are really purchasing reliability.
- As microservices multiply telemetry volume, each additional dollar produces less reliability, making upstream signal extraction more valuable.
AI Widens The Gap Between Engineers
- AI amplifies strong engineers most because they can decompose complex work, assign tasks, and manage system-level context.
- Jad Naous saw the gap between merely good and exceptional engineers widen tenfold on distributed-systems work.
Use Persistent Agents With Human Design Review
- Start with a reviewed product and design document, then parallelize implementation across persistent command-line agents.
- Keep three or four sessions manageable, and consolidate every change into one stack for human review.