arcanus

39 ms·

391.
by arcanus 9y ago
Anything with a pulse can provide decent competition for Xeon Phi. It is well know these 'accelerators' are barely providing speed-ups compared to existing CPU solutions. While Phi are easy to port, they aren't close to GPUs
392.
by arcanus 9y ago
Betteridge's law be damned, the business of scientific publishing is bad for science. Scientific publishing is a complete mess. We have a large scale business built around an industry funded by public money. This creates barriers to sc
393.
San Francisco’s Housing Slowdown Could Spread(wsj.com)
1 points by arcanus 9y ago|0 comments
394.
by arcanus 9y ago
> "First, let's assume the cow is a sphere". It is worse than that, because at least physics is validated by experimental data. Much of (macro in particular) Economics is a set of axioms that are seldom predictive.
395.
Amazon Said to Plan Cuts to Shed Whole Foods' Pricey Image(bloomberg.com)
64 points by arcanus 9y ago|76 comments
396.
by arcanus 9y ago
Certainly a risk, but this is Google, Amazon, Microsoft, Facebook we are talking about here. The largest public companies in the world. Some of the most technologically advanced as well. If they cannot bring pressure to Intel, no one can.
397.
US Dept. Of Energy Awards $258M in Contracts to Develop Exascale Supercomputer(anandtech.com)
2 points by arcanus 9y ago|0 comments
398.
by arcanus 9y ago
I've always been curious why none of the silicon companies have attempted to provide hardware (or at least ISA extensions) supported AD, backprop, etc.
399.
by arcanus 9y ago
Man or woman, I'd ask a female friend for help. In my experience women have much better eyes for dress than men, for either men or women actually. See if you can't convince someone to come along shopping with you, the second opini
400.
by arcanus 9y ago
> Does it make you second guess the company prospects? Far from it, honest is one of the absolutely critical metrics I use to evaluate any prospective employer. Bravo. I don't mind someone stating why they believe their start up w
401.
by arcanus 9y ago
So is this perceived as a lateral move? Or is this a strong negative indicator.
402.
by arcanus 9y ago
At this point isn't this basically a fab attached to a particle accelerator?
403.
by arcanus 9y ago
Speculation on what, though?
404.
by arcanus 9y ago
With BTC at all time highs and ETH up over 2000% this year, are we in a crypto-currency bubble? Is this all just from money laundering, and off-shoring money from countries that try to prevent moving assets across borders?
405.
by arcanus 9y ago
https://streamhpc.com/blog/2016-12-15/amd-gets-machine-intel... "AMD will be delivery HIP version of Caffe, Tensorflow, Torch7, MxNet, Theano, CNTK, Chainer, all supporting our new MIOpen – our new Deep Learn
406.
by arcanus 9y ago
Hard to see this making a dent against GPUs for training. But the article doesn't specify that these devices are for inference.
407.
by arcanus 9y ago
Great point. I'm inclined to agree, but we are all speculating. Right now, you are certainly right: we can craft algorithms that are super-human in certain respects, but they do not generalize well. And of course, it took 100+ top PhDs
408.
by arcanus 9y ago
Algorithms are the new 10x developer. I'm actually half serious: one can see how something that is slightly better can have a massive multiplier, e.g. a team of inferior intelligences cannot 'scale' to beat a single superio
409.
by arcanus 9y ago
We need to differentiate between neural networks, which do not have robust theoretical underpinnings, and practical considerations. DNN is fantastic from a computational standpoint. Its _GEMM all the way down. You get high flop counts and w
410.
by arcanus 9y ago
Pretty early on Denis tweeted that AlphaGo believed his opponent had been playing a perfect game, up to that point: https://twitter.com/demishassabis/status/867584056095002624
411.
by arcanus 9y ago
I'm a huge proponent of alpha Go and I think it is a revolutionary leap. The key I think is, > yet it clearly generalizes rather well to unseen board situations and still evaluates them sucessfully I'm not sure this has been pr
412.
by arcanus 9y ago
^ very much so. My comment was admittedly (unintentionally) ambiguous.
413.
by arcanus 9y ago
Interesting given that they own ARM.
414.
by arcanus 9y ago
Next-gen Intel part?
415.
by arcanus 9y ago
'After completing the paper, we read it carefully to ensure it didn’t say anything meaningful, and as neither one of us could determine what it is actually about, we deemed it a success.'
416.
by arcanus 9y ago
The labs don't employ many life scientists compared to physics and engineering doctorates. The article is focused on biology. Not that we don't generally need more staff scientists, but the national labs are probably the best gig
417.
Should taxpayers cover the light bills at university labs?(pbs.org)
1 points by arcanus 9y ago|0 comments
418.
by arcanus 9y ago
There is virtually no way this would be fp32. If it was, they could sell these devices for many more tasks than just machine learning.
419.
by arcanus 9y ago
Or even lower, companies are experimenting with ternary, fp8, etc. for inference tasks.
420.
by arcanus 9y ago
Are (sales > 0)?

More ›