jmalicki

82 ms·

1081.
by jmalicki 11y ago
To the long-term investor, the correlation changing sign very frequently is very similar to "uncorrelated"
1082.
by jmalicki 11y ago
Public companies have already rethought this. They give equity compensation in RSUs (i.e. options with a strike price of 0), rather than options with a strike price at the current valuation. With RSUs, you are rewarded for meeting high exp
1083.
by jmalicki 11y ago
As stated above in the thread, it has a 50% accuracy rate on an analogy dataset.
1084.
by jmalicki 11y ago
If you want an equivalent you'd have to think in similar units. Like Obama + 2017 = 1981 + Carter or Obama + 2017 = Bush + 2009
1085.
by jmalicki 11y ago
context to understand wtf you're asking?
1086.
by jmalicki 12y ago
Or you could just store that in a cookie.
1087.
by jmalicki 12y ago
That link is horrible. They state that the population is (in millions), when it's actually in thousands. And they can't be bothered to include the actual 2010 census.
1088.
by jmalicki 12y ago
No, if you read their technical paper, it's frequentist sequential testing with false discovery rate control, which is a fairly recent development (I mean, 25 years old is pretty new in statistics). http://pages.optimizely.c
1089.
by jmalicki 12y ago
RandomForest and other decision tree methods can actually handle missing data very well by treating an individual cell in the data matrix as missing, rather than discarding the entire row. So the assertion that "All algorithms should o
1090.
by jmalicki 12y ago
Net revenue isn't operating income - it does not include engineering or administrative expenses, where as net income does. So even with net revenue, they would not likely be profitable.
1091.
by jmalicki 12y ago
http://www.fedex.com/us/small-business/updates/February2011/...
1092.
by jmalicki 12y ago
If your database is ACID, you can pull the power on the server from the wall and have it instantly turn off w/o data loss. That's the entire point of transactions...
1093.
by jmalicki 12y ago
They mentioned LZW, LZSS, and LZH, if you had bothered to read it.
1094.
by jmalicki 12y ago
Actually, you can. Sites like topcoder and Project Euler are actually very good at building skills... they're much like playing arpeggios or similar exercises... they're small, may take 30 minutes to two hours to complete, and the
1095.
by jmalicki 12y ago
Correct... if there is only one minimum, it is unimodal, but not necessarily convex. For instance, if there are inflection points, it will not be convex.
1096.
by jmalicki 12y ago
Actually, although the direct method is often faster than gradient descent, it is not faster than conjugate gradient. And conjugate gradient is more numerically stable. So the question is, why would anyone competent use the direct method f
1097.
by jmalicki 12y ago
GPGPU designs from NVidia are an AI targeted processor. It's the application they're designed for, and used everywhere for.
1098.
by jmalicki 13y ago
Nokia HERE in Berlin hires pretty globally: https://www.linkedin.com/job/c-here,-a-nokia-business-l-berl...
1099.
by jmalicki 13y ago
Michael Porter's "Competitive Strategy" is oddly missing from the list, as it's one of the most influential corporate strategy books out there... if you like the spirit of that list, you should definitely add that one. (
1100.
by jmalicki 13y ago
The most interesting example is Royal Dutch vs. Shell Transport. Both companies had equal shares in Shell Oil - Shell Transport was in the S&P 500, Royal Dutch was not, and sold for much less, despite having identical economic interest
1101.
by jmalicki 13y ago
In the United States, at least, the poor often have marginal tax rates nearing or beyond 100%, once you include loss of means-tested benefits. If high tax rates affect work ethic, it is most pronounced among the working poor in the United S
1102.
by jmalicki 13y ago
> I'm not clear on why layer-7 level MTU discovery is preferable to regular ICMP-based pMTUd, which actually does work with UDP (though you have to use recvmsg in a very complicated way) One big reason is that it's really, real
1103.
by jmalicki 13y ago
>Curing cancer? There's no Federal Department of Stopping >Cancer, either you work for a big pharmaceutical company or >you get a PhD and scramble for grant money. Actually, there is a Federal Department of Stopping Cancer - t
1104.
by jmalicki 13y ago
A lot of the best people a company wants to hire are the people who are happily employed, whose employer wants to keep them, and need to be talked into applying - especially if a candidate might be in a different city and would need to relo