Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
necubi
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
47 ms
·
481.
▲
by
necubi
12y ago
There's nothing wrong with going to college. College can be great! If you do it right, you'll learn a lot about a diverse set of things including history, literature, the arts, and the sciences. If you're in a good program, y
482.
▲
by
necubi
12y ago
I have. You have to click through to a separate screen to get the option to disable it. If you try that, it will warn you on the next screen that you are not using the recommended settings. Then variously using IE10 you will get prompted to
483.
▲
by
necubi
12y ago
Do Not Track has been effectively dead for 1.5 years. Very few advertising companies ever supported it due to Microsoft's decision to buck the spec and enable it by default for all IE10 users (a transparent attack on Google by a compan
484.
▲
by
necubi
12y ago
You get a sense of increasing exasperation from the Supreme Court with the incredibly pro-patent antics of the Federal Circuit. See the good discussion at SCOTUSblog here: http://www.scotusblog.com/2014/04/opinion-
485.
▲
by
necubi
12y ago
It's definitely an Apple keyboard (look at https://www.youtube.com/watch?v=Y3dUeGNIX4M&feature=youtu.be... — the command key is clearly visible), albeit an old one.
486.
▲
by
necubi
12y ago
Netflix always has another recourse in this situation: they could use a transit provider (like L3) that Comcast already peers with. Evidently they've decided that paying Comcast is cheaper than paying L3. If Comcast were deliberately s
487.
▲
by
necubi
12y ago
Note that these deals, while potentially objectionable, don't violate net neutrality as it's commonly understood. The key difference is that they involve the transit between Netflix and Verizon/Comcast, rather than within
488.
▲
by
necubi
12y ago
Hey, Peter. Fellow Wes'12 here. As a student at the most-active school on CollegeACB (CollegeACB was originally developed as a replacement for a LiveJournal-based board active at my school), I saw the good and the bad. It was terrible
489.
▲
by
necubi
12y ago
How do you propose we get competition without regulation?
490.
▲
Tokaido 1.0 Released
(github.com)
26 points
by
necubi
12y ago
|
17 comments
491.
▲
by
necubi
13y ago
Ah, interesting. My personal experience with Mesos has included clusters that are basically all transient services (like map reduce jobs) or all long running services, but not both. I can see how that might lead to pathological scheduling d
492.
▲
by
necubi
13y ago
Thanks for writing this. We're also running a large HBase cluster on EC2 and frustrated with the poor suitability of instance offering (there's literally nothing cost effective for HBase in the current generation). I am curious wh
493.
▲
by
necubi
13y ago
I haven't read the Omega paper yet, but plenty of people are running long-running tasks in Mesos (Marathon [0] is a framework for doing just that). [0] https://github.com/mesosphere/marathon
494.
▲
by
necubi
13y ago
Can anybody discuss how Flynn compares to Mesos? Superficially, it seems to be solving the same sort of problems.
495.
▲
by
necubi
13y ago
Here's a challenge for somebody: create a code review/pull request tool that helps you really understand the code changes. Some IDEs do an ok job of this for static code (IntelliJ for Java, Emacs haskell-mode), but I've nev
496.
▲
by
necubi
13y ago
Lynching involves rounding somebody up, typically a black man, and murdering them. Let's not get carried away with our metaphors.
497.
▲
by
necubi
13y ago
I was confused what RFC 520 ("A Proposed File Access Protocol Specification") had to do with any of this. It looks like Theo meant to refer to RFC 6520: https://tools.ietf.org/html/rfc6520 .
498.
▲
by
necubi
13y ago
Wes peep, reporting in. It's always bizarre to see our little liberal arts school pop up on HN. I also finished my Wesleyan CS degree in 3 years, and it was totally doable (although that was by taking 1.5-2x normal credits most semeste
499.
▲
by
necubi
13y ago
This is not meant for reading paperbacks. A kindle paperwhite, at $120, is a far better choice for that. This is for reading and annotating documents, and generally replacing paper in an office or academic setting. Given then number of rese
500.
▲
by
necubi
13y ago
Or people hoping to capitalize on others falling for this...
501.
▲
by
necubi
13y ago
The samizdat refers to a film (after which Infinite Jest is named). This film is so entertaining that, once seen, a viewer will want to do nothing but continue watching it, endlessly. (Also, IJ is a great novel.)
502.
▲
by
necubi
13y ago
If you're dealing with a lot of marketplace fraud, you might take a look at Sift Science ( http://siftscience.com ). We're a YC company that uses machine learning to detect fraud, and we have a number of marketplaces usi
503.
▲
by
necubi
13y ago
I, and basically every language designer since the 80s has disagreed with this. It's needless repetition: all of the information contained in the .h file is duplicated in the .m file. If you really want a nice listing of the exported m
504.
▲
by
necubi
13y ago
Stereo rendering requires rendering the scene twice, once for each eye. Off the bat, this is almost twice as computationally intensive as rendering it once.
505.
▲
by
necubi
13y ago
Nice work, but unfortunately it doesn't seem to implement the card draw algorithm correctly, which leads to unplayable games pretty quickly (the card draw algorithm must pick the same number of 1s and 2s over a short time horizon for t
506.
▲
Threes in your terminal
(github.com)
2 points
by
necubi
13y ago
|
0 comments
507.
▲
by
necubi
13y ago
I run http://hnhiring.me , which scrapes the monthly posts and presents them in a nicer format. I haven't had much time to work on it lately, but the code is open at http://github.com/mwylde/hnhiring if
508.
▲
by
necubi
13y ago
If you find yourself in Silicon Valley, you can see the original Utah teapot at the Computer History Museum. They have an incredible collection of artifacts from the beginnings of the computing age.
509.
▲
by
necubi
13y ago
Funny, I've been doing the same thing. The hardest (well, most tedious) part has been figuring out the exact distributions used for the various random parameters.
510.
▲
by
necubi
13y ago
Or a bit cleaner using lift-json: https://gist.github.com/mwylde/8724700 .
More ›