zby

41 ms·

421.
by zby 7y ago
"Why Correlation Usually ≠ Causation: Causal Nets Cause Common Confounding" https://www.gwern.net/Causality - that should be a required reading.
422.
by zby 7y ago
My explanation: https://www.gwern.net/Littlewood - the surprising stuff is what spreads on the internet. Anecdotal information is becoming more and more useless - the only recourse is statistics.
423.
by zby 7y ago
My intuition is that you are wrong - but your argument is important one: https://slatestarcodex.com/2015/09/16/cardiologists-and-chin...
424.
2019: The year revolt went global(thefifthwave.wordpress.com)
53 points by zby 7y ago|63 comments
425.
by zby 7y ago
"One common theme in the chronic Lyme communities is that people who go into remission tend to attribute their response to whatever technique, supplement, or medication they were trying at the time." This is spot on. People probab
426.
by zby 7y ago
OK - a lot of answers here are about "doctors using online resources to fill up their knowledge gaps is totally normal" - which is a fine point. But the article is about a particular online resource not known for an a priori high
427.
by zby 7y ago
Printing money only works if the debt is denominated in the printed currency. That is why there are plenty of government defaults.
428.
Gwern's Most Important Writing(guzey.com)
4 points by zby 7y ago|0 comments
429.
by zby 7y ago
I was more thinking about making the licensing attractive to cloud providers. The idea is that cloud providers are the dominant players - so for Open Source to survive it needs to be adopted by them. But yeah - I don't know how to do t
430.
by zby 7y ago
The cloud is eating business relying on licensing. This is relevant for both open and closed licenses. By the way I have wrote a post on that when the MongoDB debate was raging: https://medium.com/hackernoon/aws-and-mon
431.
by zby 7y ago
There is more than just Matt Levine's opinion - there is his explanation. The SoftBank accounting is not exactly what the journalists suggest, I still don't think it is fine (I don't quite agree with Levine) - but the article
432.
by zby 7y ago
"We have talked about this strategy before, and actually reading the article mostly assuaged my concerns about SoftBank’s bookkeeping. In fact, SoftBank says that it doesn’t aggressively mark up its unicorn portfolio (and take accounti
433.
by zby 7y ago
It is shallow as it always was - but at least it is not mean.
434.
by zby 7y ago
What our brains do is curve fitting plus experiments . Causality is learned through experiments - and as children we do a lot of experiments (with moving our limbs etc). With just observations you can have only correlations. For example y
435.
by zby 7y ago
"Subscriptions are ultimately bad for the buyer because you pay more and get less in most situations." - in efficient markets anything that is good for the supplier will eventually be also good for the customer. Companies compete
436.
by zby 7y ago
The question is what do you call "great" and my hunch is that "great" means "impressive" or "surprising" - and when you see lots of small steps they don't seem impressive and they don't surp
437.
by zby 7y ago
"None of the truly great achievements of science were done through atomized conjectures and refutations." - that is kind of tautology - only the big steps are called truly great - the small steps are just small steps.
438.
by zby 7y ago
Elon has overpromised many times in the past - and probably it won't happen in 2020 but I still believe that Tesla is leading the driverless competition. They have the data.
439.
by zby 7y ago
I am all for interoperability, and personally I wait for a Facebook API to write your own news feed filters, that will open so much new possibilities. But I am afraid that, at least in the beginning, it will only make "hateful, extreme
440.
by zby 7y ago
Sure there is risk, Wordcom - at least the management went to prison - do you think the king would go to prison in this case? But on a more fundamental level it is not even a sham, a sham would mean that one side breaks the rules - but what
441.
by zby 7y ago
I have been reading about that deal on Matt Levine newsletter from time to time and I have always wondered who is going to trust the Saudis on this. The country is a kingdom, not a parliamentary kingdom but a traditional one where the king
442.
by zby 7y ago
Yeah - but that is kind of special case when talking about the problem with infections. It does not justify "All sickness".
443.
by zby 7y ago
The article does not really support the implicit thesis from the title - it lists some diseases that are caused by bad inflamation, some items are maybe surprising - but it does not mention diseases that obviously have different causes like
444.
by zby 7y ago
It was an opinion. I am now sorry for being snarky - but my opinion did not change - whole article about the surprising iformation that spies meet at restaurants.
445.
by zby 7y ago
Warning: I tried to read it - but after half of it I found no content.
446.
by zby 7y ago
Open Source is about licenses - Saas is making licenses obsolete. I think the story is more about universal cloud providers dominating specialized Library as a Service providers (AWS dominating MongoDB) - the fact that the smaller providers
447.
by zby 7y ago
Citing wikipedia: "Vitamin K is a group of structurally similar, fat-soluble vitamins found in foods and in dietary supplements", but of course drugs always contains also other substances - so different label drugs do differ and c
448.
by zby 7y ago
I have observed a case here in Poland - the parents refused a vitamin K shot to their baby because they found something in the leaflet. Later it turned out that this particular vitamin K was not approved in Poland for normal use, only for u
449.
by zby 7y ago
How about "Legacy of Ashes"? https://www.amazon.com/Legacy-Ashes-History-Tim-Weiner/dp/03...
450.
by zby 7y ago
We want our public figures to be predictable. When we lived in small bands we knew the peoples our whole life and we could build a really good model of their thinking. Later when we had bigger polities, we did not have that much information

More ›