arcanus

72 ms·

721.
by arcanus 10y ago
The United States has traditionally had rather poor governance, and that has not materially impacted our tech sector. Furthermore, I don't think California is the most business friendly state in the union, but that has clearly not supp
722.
by arcanus 10y ago
Great intro! I went on coffee tastings and tours in Seattle last time I visited and the head to head comparison between brews was a great way to train my pallet on the different flavor profiles. Treating it much like wine tasting has been a
723.
by arcanus 10y ago
> extracting a parisomonious mathematical model from a neural net is nontrivial Has this ever been done? It strikes me as impossible. > Of course, were the model really 100% accurate, then converting it to a the most parsimonious perf
724.
by arcanus 10y ago
I see more geopolitical impediments than technological...
725.
by arcanus 10y ago
Isn't that the magic of OSS? Linux::Windows, matplotlib::mathematica, android::iPhone, etc. In each case, the free variety quickly catches up to the proprietary version, and in doing so, cuts into the profitability of the parent. Furt
726.
by arcanus 10y ago
Excellent post, I rather agree. > As long as AI is an open technology there will always be some criminals who just want to see the world burn. Thankfully we have yet to see terror attacks with nuclear weapons, but the lower barriers of e
727.
by arcanus 10y ago
In a variety of engineering fields, including but not limited to software, we have wonderful tools to track down and eliminate 'bugs'. While high standards are often not upheld, the concepts are largely sound. In particular, I&#x2
728.
by arcanus 10y ago
> The Hessian is too expensive and too big. Not necessarily, often derivatives are analytically known in ML. > MPI is not great at even high hundreds of cores ? You realize that Sequia, which I have run on, has codes that scale to all
729.
by arcanus 10y ago
Nocedal's text, 'Numerical Optimization' is the standard for that field. As he notes, I've always been surprised that more techniques in ML do not leverage the Hessian to get quadratic convergence rates. Nevertheless, th
730.
by arcanus 10y ago
> it is a niche technology because it works against human physiology I'm not sure I agree. Many complained of motion sickness playing early FPS games, such as DooM or Descent. I wouldn't be surprised if it just takes an equili
731.
by arcanus 10y ago
> Buy an abandoned roller skating rink and transform it into a VR game space I also love this idea, but it still has many barriers. Principal technical issue is that these headsets are not wireless, which means you would need anot overhe
732.
America's upper middle class is growing(wlwt.com)
46 points by arcanus 10y ago|105 comments
733.
by arcanus 10y ago
The MOOC on coursera is an excellent place to pick up some basic terminology and simple algorithms. It is taught at an undergraduate level, so it is rather accessible.
734.
by arcanus 10y ago
> resorts to cheap tricks like regularization, Regularization is not a 'cheap trick'. It is an essential technique for mathematically ill-posed problems, where no unique solution exists. This is extensively studied and well-und
735.
Job-Seeking Ph.D. Holders Look to Life Outside School(wsj.com)
2 points by arcanus 10y ago|0 comments
736.
by arcanus 10y ago
Behind a Paywall.
737.
Russian Top Secret Hypersonic Glider Can Penetrate Any Missile Defense(sputniknews.com)
1 points by arcanus 10y ago|1 comments
738.
by arcanus 10y ago
I'm getting a negative tone from this post and I don't really see why. Training is the point of graduate school. That is why we still call them graduate students . > I thought a PhD was about making an original contribution to
739.
by arcanus 10y ago
I used to schedule time after updating to fix my broken X11 server
740.
by arcanus 10y ago
Toyota's market cap is 200B VW is at 90B Ford is 53B GM is at 45 B Hyundai is 30B Roughly speaking he is predicting that not only will Tesla become the largest automotive company by valuation, but that it will be as large as all the ot
741.
by arcanus 10y ago
> Is it not that way in other sciences? I can speak to the fact that this is absolutely the case in the physical sciences.
742.
Fraud, bureaucracy and an obsession with quantity holds Chinese science back(economist.com)
1 points by arcanus 10y ago|0 comments
743.
by arcanus 10y ago
> You'd then have them simultaneously overrepresented in top positions AND scoring lower on IQ, verbal scores, etc. Yep... you can have distributions with identical means, and fatter tails.
744.
by arcanus 10y ago
> Economic expansions don't die of old age. I don't disagree, my comment was purely empirical. According to the historically observed frequency of recessions in the USA, we have gone longer than is typical.
745.
by arcanus 10y ago
Along those lines, I've heard less press about statistical learning start ups than some of the recent deep learning work. Anyone know of some recent ventures in this space? Anyone to keep an eye on?
746.
by arcanus 10y ago
Good stuff! Nice proof of concept. I would have like to see some scalability plots, to evaluate the parallel speedup you observed.
747.
by arcanus 10y ago
We have been out of recession (in the USA) since '08, which historically is a long time. Furthermore, we are at full employment, with unemployment at 4.7%. So we certainly could see the end of this business cycle soon. That alone is no
748.
by arcanus 10y ago
I'm a southpaw myself, and therefore quite biased, but I've always been struck by the prevalence of CEOs, Presidents, notable scientists, etc. that I have met who are fellow lefties. Don't have a p-value, but it seems much hi
749.
by arcanus 10y ago
"As of 2014, the annual cost of a UBI would have been about $200 billion cheaper than the current system. By 2020, it would be nearly a trillion dollars cheaper." Remarkable how much entitlement spending is booked in the near futu
750.
by arcanus 10y ago
Most of my stuff is in numpy/scipy, so that sort of array would be great (they are C arrays under the hood, so pretty fast). As you can tell, most of my work is in python. But R is absolutely huge, so providing an interface to that wou

More ›