ghj

I think those were synchronous requests one after another. 250ms per page load sounds about right.

As a layman, is causality worth studying?

For example, can I answer questions like what are the main causes that are moving corona stats?

With intervention I suppose you can conduct an experiment where you (randomly) pick cities and make half of them wear masks and half of them not. But this is of course unethical! And some stuff you want to know the effects of (e.g., how did gatherings at protests affect the infection rate) are one time events that can't be replicated again.

So all we have left are lots of natural experiments. Different countries/states/communities are handling the situation differently with a wide range of outcomes. No two communities are directly comparable since they differ along many other dimensions other than corona policy. But as a human I am still drawing plenty of conclusions on what caused what and to what degree. So it seems like a solvable problem. How do I make it rigorous?

It's really hard to write correct code.

Sorting was broken in java and nobody noticed for a long time: http://envisage-project.eu/wp-content/uploads/2015/02/sortin...

The same was true for java's binary search: https://ai.googleblog.com/2006/06/extra-extra-read-all-about...

So I am not sure I will ever trust an ML algorithm trained on inputs/outputs only (which is what I think "neural program induction" means). The above bugs are only hit because the programmers who wrote it didn't think about overflow. What is this "neural programmer" assuming about their inputs? We'll never know.

OTOH the standard library sort can be improved if you know the distribution of the numbers you're sorting (e.g., small numbers can bucket sort, small lists can use a sorting network, etc). If this thing can efficiently almost-correctly-sort because it's better at these types of pattern matching, we can just run a final pass of insertion sort to make it useful!

OMG there are multiline lambdas!!! https://coconut.readthedocs.io/en/master/DOCS.html#statement...

But trying it out, it doesn't even look like python anymore. Especially if you want to nest them a lot like in other callback heavy languages. It requires a lot of semicolons and parens to get past the parsing ambiguity problem: https://stackoverflow.com/questions/1233448/no-multiline-lam...

    # Incorrect, the lambda returns (y, [1,2,3]) here!
    map(def (x) ->
          y=x+1;
          return y
    , [1,2,3])
    
    # Right
    map((def (x) ->
          y=x+1;
          return y
    ), [1,2,3])

I don't know anything about this field. Are these distributed systems? A billion 1024 dimensional vectors sounds like a whole lot of memory (8 TB if each dimension is a double?).

Reading through some of these talk titles feels like reading https://www.reddit.com/r/VXJunkies

I remember first reading this book around the same time I read "How to Solve It". They were both popular within highschool-level math competition circles. If you have a college math degree, you should be fine!

Why are some empty ocean squares colored dark blue and some light blue? Does light blue mean there is some land mass there but it's uninhabited?

Am I the only one who is reading this as Scott flexing on NYT and that there's nothing to be sad about?

> After considering my options, I decided on the one you see now. If there’s no blog, there’s no story. Or at least the story will have to include some discussion of NYT’s strategy of doxxing random bloggers for clicks.

So his blog will be offline for a bit but it is clearly not permanent (if everything goes as planned).

Great move and probably will gain him even more clout within the rationalist community.

Can someone point me to resources on how to re-implement the "vs" feature?

I feel like this can be done with just GPT-2 instead of using the google search api. For example if you type something like "Instead of using Tensorflow, use ..." into https://talktotransformer.com/ it can babble pretty coherently about related technologies so the data is there.

Some of the arrows are so randomly placed it reminds me of:

https://www.tylervigen.com/spurious-correlations

https://www.buzzfeednews.com/article/kjh2110/the-10-most-biz...

I guess we're past the point where codewords like "winnie the pooh" work but is NLP so advanced it can parse all the other ways human hint and indirectly communicate with each other?

Like can't I just send something like "hey coworker, lets do the thing that's opposite of ionized!". Or will the AI block my message (or more scarily, subtly rewrite it) and inform my manager?

This seems like tech with huge unintended consequences if perfected (even for their intended benign uses in workplaces) since it has the potential to be scaled up to censor society.

Random question: what's the word for propaganda in chinese? I am also chinese but can only speak with poor vocab. Google translating it yields 宣传, which is a word that I do know but it means something more like "publicity" which loses all the negative connotation of the word propaganda.

I just want to make sure it's not a common mistranslation and that people aren't talking past each other. One side is saying the government is actively lying in order to manipulate people. And the other side might be interpreting as the government working on their public image... which is uh a lot more reasonable when put that way.

Funny how language shape the way we think.

It requires a google account to download the archive. Can someone summarize or cherry pick some examples? Just wondering what counts as manipulation and whether I would be able to tell.

Would that be classified as a denial of service attack? (which is apparently a federal crime)

Any example of this being used in the wild? I feel like it could save with backend costs (especially for hobby projects) but the bandwidth/loading time might make it infeasible?

Also if you have any novel model it will be trivial to reverse engineer it. You gotta send the weights over to the client and they can just run tensorflow.js themselves right?

I stole the link from this comment: https://news.ycombinator.com/item?id=23451288 I have no idea where this is from!

I think a better way to sell a language is to name a few concrete niches that they are amazing at. For example: C for low level systems, javascript for browser apps, bash for shell scripts, java/swift for mobile, erlang for actor models, haskell for compilers, python for ml, go for distributed systems, etc etc.

To me, scala used to own the niche of "if you must use a jvm language but don't want to use java". But now kotlin is the first language that springs to mind. What would you say scala is best in class for now?