Sense and decide
Weight by Track Record
Weight views by track record on the topic
On a contested call, weighting views by who has been right in that specific area before, including about where AI is reliable, rather than by rank or confidence.
Borrowed from Ray Dalio's "believability weighting" at Bridgewater Associates, described in his book Principles. Read the original source.
Changes a rule or role · Minutes to start · Anyone on the team can start it
Try this first
Ask one question. Who here has been right about this kind of thing before?
Use it when
A contested decision where the loudest voice isn't the most credible.
Skip it when
You need a relevant record to weigh. In a new domain nobody has one, and this should never become a reason to silence a junior. It weights a view, it does not gate one.
How to introduce it
On a contested question, weight opinions by track record in that specific area, not by rank or by how sure someone sounds. Believability gets earned, and it is domain by domain. That includes a record of judging where AI holds up. The loudest enthusiast is not the most believable witness about AI.
How to show up
Weigh contested views by demonstrated judgment, which means saying out loud whose record here is strong. Watch for rank and confidence creeping back in as the real weights.
How long it takes
No setup. It's applied on contested calls.
What makes it hard
Rank and confidence pull hard, so this takes effort and some honesty about who has been right. Records are domain-specific too. Someone strong on one question can be weak on the next, including on AI.
What it looks like when it's working
Contested calls tilt toward people with a record, and sounding certain stops being enough. If the loudest or most senior voice still wins, the stated weighting is not affecting the decision.
How long until it sticks
Several contested decisions before weighting-by-record displaces weighting-by-rank.
How you know it stuck
The team asks who has been right about this before, and means it.
The idea behind it
Past judgment is useful evidence only when it comes from a similar domain. Seniority and confidence are not substitutes for that record.
Where it comes from
Ray Dalio's "idea meritocracy" model. The loudest AI enthusiast in the room is not automatically the most reliable judge of where AI works. Weight someone's view by their record in that specific area, not by rank or by how confident they sound. Standing changes by topic.
Why it matters for senior teams
Seniority can make this problem worse because people are less likely to challenge a high-status colleague
Also fits: Confident claims go unchecked
Home problem: Big calls are made by too few people, or for too short a time