raganesh
raganesh on Hacker News
Recurring revenue - launch the satellite constellation once and generate monthly subscription revenue potentially running into billions of dollars a month.
Manufacturing limitations & launch pad availability, once they are resolved (along with launcher reuse) will lead to a point where there is not enough demand to fill the launch roster. This internal demand will help them fill it up.
Satellite internet service should provide major operational cash flow for SpaceX. That too on recurring basis.
Rocket launches generate revenue once per launch and for the foreseeable future, there will only a limited number launches per year.
Very good move by SpaceX.
Good point.
I guess it is a fair assumption that the neural network is trained using a corpus of logos that already look "well made" to the human eye.
It is possible there may not have been any/many samples that would be considered bad.
It would be interesting to see how this neural net evaluates a bad logo, using a tool such as LIME.
Posted by
raganesh on Thursday January 30, 2014 @11:43PM
from the daringfireball.net dept.
This is cool. But should the fourth section (sub-seconds) be running from 0-99? It is currently running from 0-59, too.
Sorry about the nitpick; this really is a cool implementation :-)
Not sure if your location is covered by Google's Mapmaker.
For my location, I usually provide updates/corrections via Mapmaker; it goes through a community review process and shows up in the main Maps within a couple of months (at most).
Just curious. Which company would be apt to provide "security solutions", then?
Some unintended consequences of playing the video backwards - when I go in reverse from slide 9 (Fan) to slide 8 (Thermal Core), the arrows indicating air flow and thermal flow are animated in reverse :-)
My respect for Gabe Newell just went up:
Says Ellsworth: "Gabe was completely behind it... I talked to Gabe, and he talked to the lawyers, and he's like, 'It's theirs, make it happen,' because he could see we were passionate about it."
Posted by
raganesh on Wednesday May 01, 2013 @12:41AM
from the daringfireball.net dept.
Posted by
raganesh on Thursday February 21, 2013 @03:27AM
from the mediaite.com dept.
That is cool. And thanks Mike, for your incredible d3.js library.
Posted by
raganesh on Tuesday October 16, 2012 @06:26AM
from the what-if.xkcd.com dept.
Posted by
raganesh on Tuesday October 09, 2012 @04:44AM
from the what-if.xkcd.com dept.
Posted by
raganesh on Monday October 08, 2012 @11:36PM
from the fosspatents.com dept.
Posted by
raganesh on Monday October 08, 2012 @05:03AM
from the slate.com dept.
Posted by
raganesh on Tuesday September 18, 2012 @04:42AM
from the what-if.xkcd.com dept.
Posted by
raganesh on Monday September 17, 2012 @10:45AM
from the officialandroid.blogspot.com dept.
Posted by
raganesh on Monday September 17, 2012 @10:29AM
from the fosspatents.com dept.
Posted by
raganesh on Tuesday September 04, 2012 @04:10AM
from the what-if.xkcd.com dept.
Not sure what you mean by "rarely beneficial to compress to one single file".
But in multiple situations I have seen significant web load performance improvements when the number of discrete files being loaded is reduced - merge all JS into one file, all CSS into another and sprite images. This is due to HTTP 2-connection rule.
Posted by
raganesh on Tuesday August 21, 2012 @07:17AM
from the what-if.xkcd.com dept.
Posted by
raganesh on Tuesday August 21, 2012 @02:48AM
from the allthingsd.com dept.
Posted by
raganesh on Saturday August 18, 2012 @01:32AM
from the androidcentral.com dept.
Posted by
raganesh on Friday August 17, 2012 @10:43PM
from the bloomberg.com dept.
Posted by
raganesh on Tuesday August 14, 2012 @10:32PM
from the daringfireball.net dept.
Posted by
raganesh on Tuesday August 14, 2012 @08:11AM
from the what-if.xkcd.com dept.
Posted by
raganesh on Tuesday August 14, 2012 @12:42AM
from the fosspatents.com dept.
Posted by
raganesh on Thursday August 09, 2012 @01:26AM
from the gigaom.com dept.