Episodes (Page 2)
✨
API governance, not agent instructions, prevents misuse.
✨
Vector search is insufficient for complex coding context.
Peter Werry
✨
Serverless AI agents face cold starts and regional limits.
✨
Custom agents often remain siloed on individual laptops.
✨
Coding agents pose a 'lethal trifecta' risk on laptops.
✨
Gil Feig, co-founder of Merge, explained the need for company context graphs.
✨
Pedro Lopez from Airbyte detailed designing CLIs for agents, not humans.
✨
Gabriel Martinez from G2i defined 'AI slop' as work outpacing judgment.
✨
Anders Swanson from Oracle discussed agent memory beyond context windows.
✨
Dominik Kundel from OpenAI presented the Codex App Server.
Dominik Kundel
✨
Greg Pstrucha from Sentry advocated codifying rules over re-prompting agents.
Greg Pstrucha
✨
Ali-Reza Adl-Tabatabai from Sonar discussed validating AI-generated code.
Ali-Reza Adl-Tabatabai
✨
Felipe Blanes from Amazon AGI Lab highlighted the 'benchmark illusion'.
Felipe Blanes
✨
Context windows have limits; models forget middle information.
Elizabeth Fuentes Leone
✨
Viren Baraiya explained running AI agents safely in production.
Viren Baraiya
✨
Kamalakannan Nandagopal from Postman mapped microservices for coding agents.
Kamalakannan Nandagopal
✨
Em Shreve from Automattic shared strategies for building AI fluency in remote teams.
Em Shreve
✨
Will Lyon from Neo4j discussed turning agent memory into actionable skills.
✨
Matt Gibiec from Dynatrace argued AI agents need a 'nervous system'.
Matt Gibiec
✨
Most companies struggle to see ROI from AI agents.
Ali-Reza Adl-Tabatabai
Tariq Shaukat