Hacker News new | past | comments | ask | show | jobs | submit login

I love the Feynman reference. Thank you. And this may be right.

However, I also am reminded of Nassim Taleb's Fat Tony, who when asked 'what are the odds of flipping a coin heads 10 times in a row?' responds with 'fuggehdabowdit. it's a hustle.' There's a scientific response, which can be very naive in some ways, and there's the Fat Tony street analysis. As I've gotten older, I tend to value the latter.




As I've gotten older, I tend to value the latter.

Definitely a good perspective to keep in mind. But at Facebook scale, if a billion people flip an actually fair coin 10 times, around a million of them will get 10 heads. It would be understandable for those people to conclude that the coin is biased, but they'd be wrong.


Right. And 0.1% of them, 1000, will post here on HN how this can't possibly be a coincidence.


Oh I’m not saying they’re not trying. I’m more saying that they aren’t as good as it sometimes looks.

Humans are very good at finding patterns in happy little accidents. Much better than the ad networks are in making those accidents happen. If the tech was as good as they say, we’d never ever see an ad that wasn’t a hyper relevant instant click.

And it seems unlikely that Apple is selling my iMessage conversations to Facebook. The “we searched for it, created an explicit signal, and forgot because it’s such a habit” explanation seems more likely.

Plus if a simpler method like “hey your location is at a restaurant on most Fridays, why don’t we start pushing restaurant ads on Wednesday” works well enough, why wouldn’t they use it?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: