Post #1157323

@Nico There is also a different dimension of responsibility regarding security. Without sandboxing, I would call agentic coding completely reckless. But even with sandboxing, there are big (partially unsolvable) security problems. This talk gives a good introduction to the problems: https://media.ccc.de/v/39c3-agentic-probllms-exploiting-ai-computer-use-and-coding-agents