The Exchanges

Every argument clarity score on this site is built from rows on this page. Each question and answer was assessed with names hidden, the host's own answers included, on four things from 1 to 5: directness (does it answer the question asked), coherence (do the ideas follow), precision (concrete details and clear references), compression (says a lot per word). The weighted mix (30/30/25/15) is the exchange score. A person's published score averages their exchange scores on raw tape only, at least 8 of them, shrunk toward the cohort mean. Full method →

Bradford Cross no published score: only 1 usable exchange on raw tape, and a fair score needs 8+ record → ← everyone

Every exchange below was scored with names hidden, four dimensions each from 1 to 5. An exchange's score is 0.30·directness + 0.30·coherence + 0.25·precision + 0.15·compression. The published score averages the raw tape exchange scores and shrinks small samples toward the cohort mean, so five great answers can't beat twenty good ones. Produced feed rows count only toward coarse estimates, never toward a full score.

clear all ✕
1exchanges match
1on raw tape
0redirected or not addressed
Answered raw tape D 5 · C 5 · P 4 · Cm 4 4.60

Q Great. Um, Maybe one or two more from me, and then we'll, we'll, we'll give people a chance to ask questions. Uh, how do you think about the defensibility of those vertical startups? Um, I think, uh, Omar in particular mentioned, uh, the concept of what I call data network effects, but you were saying compounding effects. Is that, is that how you used it?

A Yeah, I think compounding effects is, is a terrific term. That, that's actually how I talk about it as well. I think what you want is this, This notion of kind of a flywheel between the data and the model, you know, you, and this is another reason for the full stack application. Um, you know, if I, if I, if I can come in there in the, again, in the Merlon example, where you have no, no data from any of your army of analysts that are saying yes, no, yes, no, yes, no, this is risky, this is not risky all day long. If you're not capturing any of that analyst interaction data, um, then you have no feedback loop to turn the problem into a learning to rank problem. Um, so first of all, I want to come in with an application. I want the analyst to use my application. I want to instrument that and get that data, because now I can turn it into a learning to rank problem. Um, but secondly, if I can do that now, add a number of different banks, and none of them had any of that data, and I can learn from the data in all the different banks, in a, maybe in a federated or distributed fashion, and kind of export the, the learned parameters outside the bank, you know, so it's safer, then I can, I can now blend them together, and I can have a model, um, that I've learned across the entire banking system. So now I have, A data compounding effect. I also have a data network effect, um, and I think…

AI assessment note: “So now I have, A data compounding effect. I also have a data network effect”

page 1
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 400 conversations transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.