Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Tossrock
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
25 ms
·
271.
▲
by
Tossrock
9y ago
Dr. Denneau, what an honor that you dropped by this comment thread! Congratulations on the publication, it's a truly groundbreaking discovery :) 10^16 is pretty mindbogglingly huge, but then again so is 4 light years...
272.
▲
by
Tossrock
9y ago
The authors estimate that there is ~1 such object within 1 AU of the sun at any given time. So, more common than we previously expected, but not exactly teeming. You're right that it was detected due to its nearness though - it was pic
273.
▲
by
Tossrock
9y ago
0.16 AU, according to the paper.
274.
▲
by
Tossrock
9y ago
I recommend anyone interested read the paper which is available as a preprint from Nature (for the moment): goo.gl/cSsHoj (link shortened due to having an embedded access token). Some tantalizing points from the paper: - They constrain
275.
▲
Ask HN: Best bidirectional Android to desktop pipe?
2 points
by
Tossrock
9y ago
|
0 comments
276.
▲
by
Tossrock
9y ago
I hadn't heard of / thought of the thermal storage idea (freezing ice at night - when it's more efficient both cost wise due to cheaper power, and thermodynamically due to lower outside temperature - and then using that to re
277.
▲
by
Tossrock
9y ago
What VR platform would you want this on? For the desktop powered versions, you can use VirtualDesktop to run a browser in VR which can visit eg nytimes.com or whatever.
278.
▲
by
Tossrock
9y ago
Well, if I were to play the Yudkowsky's Advocate here (which in general, I'm not), I would say that this is precisely what he's talking about in the article. Because Karpathy knows how hard he's had to work, and how flaw
279.
▲
by
Tossrock
9y ago
For me, the "smoke under the door" moment was Karpathy and Li's Deep Visual-Semantic Alignments for Generating Image Descriptions [1]. The almost perfectly grammatical machine-generated captions of photos were unnerving to
280.
▲
by
Tossrock
9y ago
Well, except they explicitly call the non-sentient ones female (also stupid, and property), and the sentient ones are both referred to as male.
281.
▲
by
Tossrock
9y ago
I remember after reading Ringworld on a friend's recommendation, saying that it had one too many alien species with non-sentient females for me. One, okay, interesting idea I guess, two, now you're doing a thing.
282.
▲
by
Tossrock
9y ago
Is that a broken link in a major Google product release blog post? Gasp! Screenshot for posterity: https://i.imgur.com/r7TpRQY.png
283.
▲
by
Tossrock
9y ago
Carbide is hardly exotic. You can get < $20 drill bits made from it at your local hardware store.
284.
▲
by
Tossrock
9y ago
Credibility isn't necessarily lost over failed doomsday predictions, at least among the faithful - "When Prophecy Fails" is the classic example of this. And these people tend not to have any credibility outside the faithful t
285.
▲
by
Tossrock
9y ago
Professor Kay, it's an honor to see you taking the time to provide context to this interview. For what it's worth, I was skimming the article until I saw your comments on human universals, and that made me sit up and take notice.
286.
▲
by
Tossrock
9y ago
Not sure why you were downvoted, this seems quite reasonable to me.
287.
▲
by
Tossrock
9y ago
This is part of what makes Signal exciting - great crypto backed by a lot of security folks, yes, but also mindshare and adoption, even among not traditionally security minded users. I've seen a wave of "X is now on Signal!"
288.
▲
by
Tossrock
9y ago
The math is actually quite simple to do yourself, your dilation factor is just 1 / sqrt(1-x^2), where x is a percentage of c from 0 to 1. So at 0c, your factor is 1 (obviously), at .5c it's 1.15, and at .9c it's 2.294. Divide
289.
▲
by
Tossrock
9y ago
Very similar to pg's "What you can't say" essay - http://www.paulgraham.com/say.html
290.
▲
by
Tossrock
9y ago
I know you were asking rhetorically, but there are still reasons to own a vehicle, even in a world with L5 autonomy. A pretty big thing you can do is keep stuff in there. This is perhaps a nice-to-have for many people, but for certain use-c
291.
▲
by
Tossrock
9y ago
Obligatory Dresden Codak: http://dresdencodak.com/2009/07/12/fabulous-prizes/ edit: Sheesh, tough crowd.
292.
▲
by
Tossrock
9y ago
spoilers I think he implies pretty heavily that the Rhetors actually CAN influence the past, through a mechanism very similar to that described in the article - "choosing" which way an event happened in the past. Their powers wo
293.
▲
by
Tossrock
9y ago
Very interesting - reminds me of superconductors. Some quantity which normally cannot be escaped (ie, everything has some amount of resistance, or the Q factor is a finite number) that under certain special conditions can actually hit its s
294.
▲
by
Tossrock
9y ago
Putting my old photos on Google Photos is a good idea and I think I'll do that - thanks. Although it doesn't fully solve the issue of photos that had me tagged but were uploaded by someone else, which FB doesn't give you a co
295.
▲
by
Tossrock
9y ago
I deactivated my Facebook account several months ago, and it's been about 90% great, 10% frustrating. It's great for all the obvious reasons (less timesuck, less compulsion to endlessly scroll your life away, no notification inter
296.
▲
by
Tossrock
9y ago
For those wondering, this was (around) the 44 minute mark.
297.
▲
by
Tossrock
9y ago
It's exciting to me that there are still this many naysayers, even here on HN. That suggests there's still time to enter the field before the inflection point hits.
298.
▲
by
Tossrock
9y ago
pg has a good take on this in his addiction essay: http://www.paulgraham.com/addiction.html "We'll have to worry not just about new things, but also about existing things becoming more addictive. That's what
299.
▲
by
Tossrock
9y ago
A webcomic called Toothpaste For Dinner has my favorite take on this idea: http://www.toothpastefordinner.com/index.php?date=051810
300.
▲
by
Tossrock
10y ago
Unprofitable for ETH mining maybe, but it seems like a natural fit to rent time on it to deep learning people with slow training models. Although that could still be unprofitable after the cost of electricity, I guess it's a question o
More ›