Evidence receipt / commitment
Published · transcript-backedNate Silver: commitment
13 Aug 2025 Conversations with Tyler Nate Silver on Life's Mixed Strategies
“You can say we are going to go back to 1936 — I think was the first time Gallup issued a presidential election poll — and look empirically at how accurate those polls are at X number of days from the election.”
Source trail
Everything needed to verify it.
- Speaker
- Nate Silver
- Attribution
- Verified speaker
- Claim type
- commitment
- Recorded
- 13 Aug 2025
- Publisher
- Conversations with Tyler
Transcript context
…Do you think the major tech companies have proprietary information where they can predict elections better than anyone or anything? They may not even bother to do it, but is it there in principle? I tend to doubt it. One nice thing about polling, for all the flaws that polling has — and I think polling has, I wouldn’t call it a crisis, but on the route to a crisis — is you do have a good baseline to compare with. You can say we are going to go back to 1936 — I think was the first time Gallup issued a presidential election poll — and look empirically at how accurate those polls are at X number of days from the election. Whereas there is just nothing to calibrate AI-type models. You could probably predict how an individual voter will vote, if they’re gathering . . . The amount of things I tell — But if you spent a lot of money, and you have everyone’s, say, Facebook posts, and you just fed it into a sufficiently complicated model, wouldn’t that be the best predictive model?…
Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.