AI should help researchers think deeper, not think lessagentbayes.com2 points·by guyzana·il y a 11 jours·0 comments
Token accounting and cancellation in durable agent workflowsagentbayes.com2 points·by guyzana·il y a 11 jours·0 comments
AI agent safety and alignment research, mappedagentbayes.com1 points·by guyzana·il y a 12 jours·0 comments
Show HN: Skillwave – agent orchestrator with async comms and goal-alignment loopgithub.com2 points·by guyzana·il y a 4 mois·0 comments
guyzana·il y a 3 jours·discussBtw, now that the repository is open-sourced, it poses an increased security risk for any customers who haven't migrated yet.
guyzana·il y a 3 jours·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.