A single question, asked directly, tends to produce more useful expert commentary on uncertain AI questions than the standard framing most interviews use: rather than asking an expert what they think will happen, asking how they would actually wager money on their own prediction, and at what odds, if forced to put something real behind the answer.
Why this reframing works
Experts asked for a general prediction tend to state it with more apparent confidence than they'd actually stand behind if the stakes were concrete rather than purely reputational. Asking someone to specify actual odds they'd accept on their own claim forces a more honest reckoning with their true confidence level, because vague, hedged confidence doesn't translate cleanly into a specific betting line the way a genuinely well-grounded prediction does.
What this reveals in practice
An expert who states a confident prediction in ordinary conversation will sometimes, when pressed to specify real odds, reveal considerably more uncertainty than their initial framing suggested-a useful signal for readers trying to calibrate how much weight to actually give the opinion. Conversely, an expert willing to specify aggressive, confident odds on their prediction is demonstrating a different, more genuinely held level of conviction than one who hedges once pressed for specifics.
Why this matters more for AI commentary specifically
AI's genuine technical and business uncertainty makes it a beat where confidently stated opinion is common and well-calibrated opinion is comparatively rare. A framing that surfaces an expert's actual confidence level, rather than accepting confident-sounding language at face value, gives readers a more honest basis for weighing competing expert views against each other.
How we try to use this in practice
Where practical, we push experts we interview to be specific about how confident they actually are in a stated prediction, using concrete framing like implied odds, rather than accepting a general, confidently-worded claim without probing its actual certainty. This produces less polished-sounding quotes sometimes, but more honestly calibrated ones.