Dev Interrupted

Many tokens make all bugs shallow & open source’s new maintainers | Chainguard's Dan Lorenc

19 snips
Mar 17, 2026
Dan Lorenc, co-founder and CEO of Chainguard, secures the software supply chain and researches agentic engineering. He discusses how autonomous agents are accelerating development and the security risks that follow. Topics include turning guardrails into reliable guide rails, agent-driven open source maintenance, many-tokens automated inspection, sandboxing teams, and which parts of the stack agents will replace.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

LLMs Have Inverted How Code Is Produced

  • AI tooling has flipped development from human-written lines to LLMs writing the majority of shipped code.
  • Dan Lorenc compares the shift to handing the industry circular saws where speed increases but risk of mistakes rises dramatically.
INSIGHT

Adoption Lag Turns Years Into Decades Against Attackers

  • Enterprises must narrow the gap to the AI bleeding edge or attackers will outpace defenders.
  • Lorenc warns that long adoption delays on an exponential curve can leave defenders decades behind rather than years.
ADVICE

Set Up Isolated Sandboxes For Agent Experiments

  • Do create sandbox environments and give developers isolated laptops to experiment with agentic tools safely.
  • Dan Lorenc recommends carving out time and constrained spaces so teams build intuition without exposing core codebases or production keys.
Get the Snipd Podcast app to discover more snips from this episode
Get the app