4 Comments
User's avatar
Sterling's avatar

IDSS isn’t the problem.

The problem is incentives.

AI isn’t good or bad—it just optimizes.

Expand full comment
Chara's avatar

Spot on! 💯

Expand full comment
Bryant Woodward's avatar

Great breakdown of IDSS! The challenge of balancing AI safety with meaningful progress is one I’ve been thinking about a lot in my own work on intelligence structuring. If AI safety frameworks become rigid bureaucracies, they could stifle the very intelligence we need to guide AI toward alignment. Do you think IDSS has the potential to adapt dynamically, or is it another top-down system that risks missing the bigger picture?

Expand full comment
Chara's avatar

Thank you Bryant! To be honest, I think it entirely depends on the developers at the top and their own established policies.

Expand full comment