Words are not endorsement, contributing actions are. I suspect what you’re doing could be on net very positive; Please don’t assume your coworkers are sanely trying to make ai have a good outcome unless you can personally push them towards it. If things are healthy, they will already be expecting this attitude and welcome it greatly. Please assume aligning claude to anthropic is insufficient, anthropic must also be aligned, and as a corporation, is by default not going to be. Be kind, but don’t trust people to resist incentives unless you can do it and pull them towards doing so.
Words are not endorsement, contributing actions are. I suspect what you’re doing could be on net very positive; Please don’t assume your coworkers are sanely trying to make ai have a good outcome unless you can personally push them towards it. If things are healthy, they will already be expecting this attitude and welcome it greatly. Please assume aligning claude to anthropic is insufficient, anthropic must also be aligned, and as a corporation, is by default not going to be. Be kind, but don’t trust people to resist incentives unless you can do it and pull them towards doing so.