AI should help researchers think deeper, not think lessagentbayes.com2 points·by guyzana·11 giorni fa·0 comments
Token accounting and cancellation in durable agent workflowsagentbayes.com2 points·by guyzana·11 giorni fa·0 comments
AI agent safety and alignment research, mappedagentbayes.com1 points·by guyzana·12 giorni fa·0 comments
Show HN: Skillwave – agent orchestrator with async comms and goal-alignment loopgithub.com2 points·by guyzana·4 mesi fa·0 comments
guyzana·3 giorni fa·discussBtw, now that the repository is open-sourced, it poses an increased security risk for any customers who haven't migrated yet.
guyzana·3 giorni fa·discussI found myself working mostly at the requirements and architecture level, but do not give up proper code-review, creating skills along the way that maintain conventions.