skj

57 ms·

1. 571.▲
   by skj 11y ago

   Not so. Priors/posteriors are only as good as the model they're based on. For instance, if you choose parental income as the feature it will can be a stronger signal than skin color, even though both may be good predictors. But th
2. 572.▲
   by skj 11y ago

   For every buyer there is a seller, and the seller is divesting themselves of that real estate. So, every time someone sells a house is a real world example.
3. 573.▲
   by skj 11y ago

   Don't tell Google.
4. 574.▲
   by skj 11y ago

   I don't think it compares. Investing in real estate has the bonus that it allows others to divest themselves of real estate.
5. 575.▲
   by skj 11y ago

   This is just you buying the ad slot instead of the advertiser. Why should it be any different for the website's income? If anything, more entities bidding for the ad means the prices can go higher and the website gets more money (thoug
6. 576.▲
   by skj 11y ago

   I'm pretty sure that "hush hush" audit trails for drug manufacturers are strictly illegal, would be found out, would bankrupt the company, etc.
7. 577.▲
   by skj 11y ago

   Because then you don't have an audit trail to use in your study documentation.
8. 578.▲
   by skj 11y ago

   Your conclusion is true only if the results of these tests are completely unrelated to each other. Over-testing is still a problem, but it's important to use correct statistical modeling.
9. 579.▲
   by skj 11y ago

   He/she doesn't. It's false.
10. 580.▲
    by skj 11y ago

    I think most people do care, but that many of them are disillusioned.
11. 581.▲
    by skj 11y ago

    Wouldn't this just be a general election with sort of weird rounding rules? edit: no, because of winner-takes-all nonsense. seems like a bad idea.
12. 582.▲
    by skj 11y ago

    seems irrelephant to the discussion at hand?
13. 583.▲
    by skj 11y ago

    > So your link to high level machine intelligence is wrong, I believe. Alternate explanations are not refutations. Both could be factors.
14. 584.▲
    by skj 11y ago

    > after which there is no support burden at all. Sounds like having a well-designed product can help you reach the $1m/year mark :)
15. 585.▲
    by skj 11y ago

    That they look at the other comments by the same person in the thread, and infer an overall message? :)
16. 586.▲
    by skj 11y ago

    The word "just" (at least with this meaning) trivializes what follows.
17. 587.▲
    by skj 11y ago

    > the bourbon scotch the what?
18. 588.▲
    by skj 11y ago

    I believe the post was not disputing the existence of the requirement, but the idea that it was to improve ad targeting.
19. 589.▲
    by skj 11y ago

    I did.
20. 590.▲
    by skj 11y ago

    Larry Page (my boss's boss's boss's boss's boss's boss's boss) seems like a pretty decent human being, to me.
21. 591.▲
    by skj 11y ago

    Given the list of things you need to do to ensure security, I'm going to go out on a limb and say that "configured and managed properly" is something that does not happen in reality. Until there is something more fundamental
22. 592.▲
    by skj 11y ago

    I'll have to take your word for it :) I'm not savvy enough about this topic - I just know the MVM architecture.
23. 593.▲
    by skj 11y ago

    Having signed images to allow trust does seem valuable, but easily verifiable images, without the need for crypto, seems like a more fundamental building block upon which more complex processes can be built.
24. 594.▲
    by skj 11y ago

    I don't understand why we don't make the container ID a hash of the image contents, have the docker CLI verify it, and let people use conventional means to trustedly pass around the correct container ID. This seems like a lot of e
25. 595.▲
    by skj 11y ago

    Classic app engine doesn't use docker, and Managed VMs app engine is run in a Compute Engine VM in the customer's own project, partly to avoid having to worry about running untrusted code since Compute Engine handles the security
26. 596.▲
    by skj 11y ago

    The GCS command line tool, gsutil, can talk to both S3 and GCS. That might be a nice place to start.
27. 597.▲
    by skj 11y ago

    It wasn't immediately obvious from scanning that link. Does it provide channel select? Because that's the hard part. Threadsafe queues are easy, select on them is hard.
28. 598.▲
    by skj 11y ago

    Don't trust. Use a client that adds an encryption layer on top of file.io.
29. 599.▲
    by skj 11y ago

    I assume that's because you don't get the building back? I think they get the kid back in these particular drills. If not, then yeah this is a terrible idea.
30. 600.▲
    by skj 11y ago

    http://www.theguardian.com/education/2001/aug/25/research.hi... apparently not, according to this guy.