Hey, journos are waiting by their phones, just ready for Uber et al to provide access to verifiable data to disprove the allegations.

Laura Wagner: I had heard a funny story/rumor about that person (or so I thought) and so I shared the story with the table. Only when I felt a foot slam into my shin and got a text telling me that I was an ABSOLUTE IDIOT did I realize my mistake: The funny story/rumor that I had gleefully shared with the table was not…

(Sarcasm aside, this was a very informative article...but “keep an old pair of glasses on the bedside table” is such an obvious fix that I’m surprised he didn’t suggest that)

On the other hand, someone with very poor eyesight who opts not to wear contact lenses for fear of acanthamoeba may be more likely to slip break his or her neck, Krauss says.

Sure, but as a practical matter, those fees are going to be there no matter how you pay - Best Buy or Macy’s or whatever has already assumed they’re paying X% to the credit card company and have worked that into their price.

It’s actually both.

I get the feeling we’re going to look back on vaping and wonder why people thought it was ok, like smoking or putting lead in everything.

Yeah definitely.

I meant “absolutely fantastic” in terms of prediction accuracy, not morality. If you’re talking about predicting outcomes that rely on human behavior before they happen, 90% accuracy is unheard of, a level of correctness straight out of science fiction.

I mean, just between Bidwell and Kroenke you’ve already got enough material for a lifetime, and I’d venture to say they’re not in the top half of high-profile owners. Once you get to the Snyders and the Jerrahs and the Maras you’ve got a full froth going.

I think we’re disagreeing on how the model is going to be used. You’re talking about the model as a last-step refinement after you’ve already come up with your list of “30,000 potential problems” through other means, then the model tries to narrow 30,000 down to 3,000.

Well, yeah, that’s the point: A model that’s 90% accurate (absolutely fantastic by all reasonable standards of predicting human behavior) is by definition wrong 10% of the time...which are the false positives.

Nope, I did not miss a decimal point, but maybe I wasn’t entirely clear.

Almost as importantly, the models have an enormous false positive problem. Let’s say you could create a predictive model that’s 90% accurate. That would be by all measures, absolutely incredible, but let’s assume we could.

Yeah, having a chart is extremely common.

Oh, and by the way - most libraries nowadays have an app which let you borrow e-books from the library to your device. So even if you don’t want to regularly travel to the actual building, you can still get all of these books from your local public library!

I think you’d be surprised.

Yeah, the thing about Atlanta is it 100% relies on which terminal you’re going to. The new International Terminal is excellent and some of the domestic terminals are perfectly nice...but a couple of them are friggin zoos. So much so that when I fly out of there, if I have to arrive 2+ hours early (e.g., international…

It’s always nice when someone produces a list like this and (as far as I can tell), doesn’t even bother to show their work.

Yeah, I’m pretty sure that’s what’ll end up happening.