arcanus

38 ms·

361.
by arcanus 9y ago
I thought firefly systems in Austin (started by a SpaceX alumni) had already demonstrated this concept.
362.
AMD Stuffs a Petaflops of Machine Intelligence into 20-Node Rack(hpcwire.com)
3 points by arcanus 9y ago|0 comments
363.
by arcanus 9y ago
I've worked with non-technical writers for some of my research. The problem is not that these people are not smart, but they tend to be very superficial. Reporting seldom rewards (or attracts) individuals who are interested in deep div
364.
by arcanus 9y ago
> better deep learning support They released the cuDNN equivalent for AMD devices a few weeks ago: http://gpuopen.com/developer-quick-start-miopen-1-0/ So AMD devices should now be capable of accelerating tensorflow
365.
by arcanus 9y ago
I'm not convinced that SGD is the answer for large scale distributed problems. It is certainly the best answer for single node applications, but it might not be the best for distributed training. The search for effective optimization a
366.
by arcanus 9y ago
And they are correct. We encouraged quddafi to give up his program, lifted sanctions, and then pushed him into a corner at the first opportunity. He had no viable way to abdicate without being tried for war crimes. The same policy has occur
367.
by arcanus 9y ago
I'm always surprised more libraries are not coupled with good coffee shops. Seems a perfect match. There is a bookstore in Austin that has a great coffee shop inside, and it certainly attracts folks to come in and browse.
368.
Elon Musk has better things to do than run Tesla(wired.com)
4 points by arcanus 9y ago|1 comments
369.
by arcanus 9y ago
> plus I'm not sure how, but it seems that the main foundries (TSMC, Global Foundries, Samsung) have all almost caught-up on Intel This is in part because of Physics getting in the way. Moores law is dying, because the Physics of sh
370.
by arcanus 9y ago
> Big U.S. players like Intel Corp. and Micron Technology Inc. find themselves in a bind—eager to expand in China but wary of losing out to state-sponsored rivals. China has done this to many companies, Google, Uber, Facebook, General El
371.
by arcanus 9y ago
> When we talk about female CEOs she is rarely mentioned but here she is in probably one of the most technology intensive industries with a company that was clearly floundering and she has led AMD confidently out of the woods into a posi
372.
by arcanus 9y ago
Good point. Another noted expert in "AI" is Andrew Ng (formerly Google and Baidu) who stated that, "Fearing a rise of killer robots is like worrying about overpopulation on Mars." ...For some reason Elon is really hung u
373.
by arcanus 9y ago
I suspect it will have more in common with mobile GPUs than the big iron ASICs that come from NVIDIA or AMD.
374.
by arcanus 9y ago
It is not well documented by anyone. However, the expectation is that it is a matrix or convolution coprocessor, as this is a common operation in deep neural networks (for both inference and training). For instance, NVIDIA says they are sup
375.
China Plans to Use Artificial Intelligence to Gain Global Economic Dominance(technologyreview.com)
3 points by arcanus 9y ago|0 comments
376.
by arcanus 9y ago
The ever-timely: 'Fearing a rise of killer robots is like worrying about overpopulation on Mars.'
377.
by arcanus 9y ago
No, productivity growth _as measured by economists_ is becoming irrelevant.
378.
by arcanus 9y ago
What is the average boot camp instructor salary?
379.
by arcanus 9y ago
Anyone have a guess as to what company this was?
380.
by arcanus 9y ago
$250k, not pounds. $125k base, 100% bonus. I know two fellow Phds who had this first year. This is a top-10 program, but not more.
381.
by arcanus 9y ago
It may have more to teach: it was clearly a Post-scarcity society with some form of UBI,etc. Simultaneously it postulated that the humans of the future would not be cyborgs, which appears unlikely. Furthermore, few tasks were automated in a
382.
Intel’s Upcoming 12-Core CPU May Clock Well Below AMD’s Threadripper(extremetech.com)
8 points by arcanus 9y ago|0 comments
383.
Building a 50 Teraflops Deep Learning Box for Under $3K(medium.com)
2 points by arcanus 9y ago|0 comments
384.
by arcanus 9y ago
Is McKinsey really a good source for predictions? Do they have any in house expertise on machine learning (I doubt it)? And do they even have a track record of predictions for strategic, technological trends?
385.
by arcanus 9y ago
I think 400 is likely attainable for the hedge funds, no idea about other fields. The HFT guys pay about 250 (total comp, e.g, salary and bonuses) for a newly minted PhD.
386.
by arcanus 9y ago
And it is not competitive with GPUs for truly parallel workloads
387.
We ported CAFFE from CUDA to HIP – and here’s what happened…(gpuopen.com)
15 points by arcanus 9y ago|4 comments
388.
by arcanus 9y ago
If the world actually equitably shared wealth, the western would have an absolutely astounding drop in our incomes and quality of life. Few would accept this. The average person in the first world has a massively higher quality of life than
389.
A Post-Modern Ideology Delivery System on Cyberpunk(medium.com)
2 points by arcanus 9y ago|0 comments
390.
by arcanus 9y ago
I'll answer your subtle comment quite explicitly: never. Intel clearly rushed this out to prevent AMD from having the perception of leading the space, at least in terms of the largest core count. The article even references this. I jus

More ›