Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
kough
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
21 ms
·
151.
▲
AP Calls Donald Trump Next President Elect
(mobile.twitter.com)
6 points
by
kough
10y ago
|
0 comments
152.
▲
by
kough
10y ago
> implying there's a difference
153.
▲
by
kough
10y ago
Seems comparable to my undergrad experience /s
154.
▲
by
kough
10y ago
Super interesting. I'm on mobile and haven't had time to read the whole paper yet - would it be feasible to continuously compute these perturbation vectors during training and include them as part of a larger heuristic? For instan
155.
▲
by
kough
10y ago
Exactly what I think -- here's the "deus ex machina" that he needed to wash his hands of the whole thing, after having talked himself into a corner where the only other way out was to deliver a real, working, product. The let
156.
▲
by
kough
10y ago
I bought the Custom Art Music One that I linked to in the above post. Most of the reviews I'd read agreed that it was an insanely good deal considering the quality-per-dollar ratio. And I don't regret it. Maintenance: there's
157.
▲
by
kough
10y ago
I had the same issue with noise-cancelling headphones (the bose qc 20 in-ears) where I couldn't wear them without sound pumping through them, so I ended up buying a pair of custom-molded in ear monitors (IEM is the search keyword to le
158.
▲
by
kough
10y ago
Any recommendations for which receivers to look at? My search-fu has failed me repeatedly.
159.
▲
by
kough
10y ago
Does anyone know any open source implementations of Material Design? I've been using https://getmdl.io/ but I'd be interested if there are any other projects of similar or better quality.
160.
▲
by
kough
10y ago
I think I agree, although I'd like to point out a few things. 1) Just like any other authentication/security system, probably not a good idea to write your own for production. As a learning experience, definitely. 2) The spec is p
161.
▲
by
kough
10y ago
I tried this – on a visit to Iceland, my (older) family friend / impromptu tour guide made me and my traveling companions taste it. The young Icelandic locals in the rest area laughed their asses off and snapchatted the whole thing, be
162.
▲
by
kough
10y ago
Had the pleasure to work with Olaf during an internship at Coinbase. Great guy, super fun – if you're reading this, good luck with your new startup!
163.
▲
by
kough
10y ago
Don't even need to use humans, I tried something easier -- http://github.com/peterldowns/clickbait-classifier
164.
▲
by
kough
10y ago
-- Douglas Adams
165.
▲
by
kough
10y ago
This is sort of an insane suggestion. Why use a Python distribution to install Node?
166.
▲
by
kough
10y ago
Yeah, Tensorflow seems cool, but then you benchmark and realize that (with CUDA/CUDNN installed) it's 50% slower than Caffe.
167.
▲
by
kough
10y ago
But the kids don't! Think of the kids!
168.
▲
by
kough
10y ago
As long as the language specification is updated to specify ordered keys in dicts then this is good. The big complaint about key ordering I remember hearing is that it was basically a side effect of the CPython implementation (back in the 2
169.
▲
by
kough
10y ago
caught me, I hadn't read through the article, wasn't sure whether or not WORDS was a set or a dict. Like desdiv points out, this isn't a big deal, you can maintain an equivalent WORDS structure that is a set.
170.
▲
by
kough
10y ago
You can also use the set-comprehension syntax, which functions equivalently but looks slightly nicer: return {w for w in words if w in WORDS} could have also directly used the intersection operator on the sets: return words &
171.
▲
by
kough
10y ago
... Should be the official Nautilus tagline. It certainly seems to be their governing principle. Well, actually, that's forgetting "needlessly overwritten"
172.
▲
by
kough
10y ago
The Wiki page has a nice writeup of a few downsides: https://en.wikipedia.org/wiki/DBSCAN#Disadvantages
173.
▲
by
kough
10y ago
Cool writeup. This is part of sklearn, which has a great visualization of how it performs on some different clustering tasks, including comparisons to other algorithms: http://scikit-learn.org/stable/auto_examples/
174.
▲
by
kough
10y ago
In case you were looking for confirmation – yes, "under the bleachers" is exactly right, although it's a somewhat dated term at this point. Not sure what the modern equivalent is but at least when I was in highschool (~2010)
175.
▲
by
kough
10y ago
Yeah, I totally bet you would have turned down MIT because of a PE requirement /s
176.
▲
by
kough
10y ago
Strange to see no one has mentioned this – just down the river, MIT has mandatory PE classes. They're not all "running around a track" though: sailing, pistols, archery, yoga, weightlifting, sailing, all sorts of really cool
177.
▲
by
kough
10y ago
BlaBlaCar is quite prominent in Europe, or at least when I was there two years ago.
178.
▲
by
kough
10y ago
Can we amend the title rule to allow replacing clickbait headlines with concise, detailed titles instead? Fuck these clickbait headlines. "Founder Sold Plenty Of Fish to Match for $575m Cash, No VCs Involved [2015]" Needs a 2015,
179.
▲
by
kough
10y ago
Yet another incredible technical achievement due to someone's quest for more porn ( https://github.com/fake-name/AutoTriever/blob/master/setting... ).
180.
▲
by
kough
10y ago
Readability ( https://github.com/luin/readability ) is a classic, and included as part of Firefox (I think, maybe that's been discontinued). It's essentially a bag of hand-written heuristics but they're pr
More ›