1 episode appearancesAcross 1 podcast
Home/Guests/Ryan Lopopolo

Ryan Lopopolo

1 episodes Β· Page 1/1

β€œinstead of prompting it better or to β€œtry harder,” the team would look at β€œwhat capability, context, or structure is missing?””

β€” Ryan Lopopolo

Extreme Harness Engineering for Token Billionaires: 1M LOC, 1B toks/day, 0% human code, 0% human review β€” Ryan Lopopolo, OpenAI Frontier & Symphony

  • β€’

    Harness engineering is the new context engineering

    β€œMove over, context engineering. Now it’s time for Harness engineering and the age of the token billionaires.”

    β€” Host/Guest
  • β€’

    Deploy production code without human review

    β€œno human REVIEWED code before merge.”

    β€” Ryan Lopopolo
  • β€’

    Consume one billion tokens every day

    β€œborderline β€œnegligent” if you aren’t using >1B tokens a day”

    β€” Ryan Lopopolo
  • β€’

    Structural improvements beat better prompting

    β€œinstead of prompting it better or to β€œtry harder,” the team would look at β€œwhat capability, context, or structure is missing?””

    β€” Ryan Lopopolo
  • β€’

    Optimize codebases for agent legibility

    β€œoptimize an entire codebase, workflow, and organization around agent legibility instead of human habit.”

    β€” Host/Guest