emcq

68 ms·

211.
by emcq 11y ago
Assuming a Titan X with single precision, those 600 GPUs are 4 PFlops! Deep Blue extrapolated to today with Moore's law would only be ~72TFlops. While DNN+RL+Tree search is cool, the hardware requirements for AlphaGo to play at this le
212.
by emcq 11y ago
This acquisition fits exactly with GM's plans outlined by investing $500 million in Lyft to create self-driving car infrastructure [0]. It seems like companies who build products that have clear and significant business value with exce
213.
by emcq 11y ago
It's both. They still have a long way to go for autonomous driving, but they have a young ambitious team and their leadership seems very strong. Relative to a company like Google, their weakness may be on the hardware side, but GM like
214.
by emcq 11y ago
Seems like somewhere in-between. They created a novel approach that is scalable and improves as you throw more hardware at it. And Google is throwing a lot of hardware at it based on their past matches with hundreds of CPUs and GPUs. I thin
215.
by emcq 11y ago
It has less to do with complexity and more go do with REPL/Jupyter notebook support. Rust is a compiled language and you won't get some of the ease of exploratory data analysis you do with something like ipython. I can use somethi
216.
by emcq 11y ago
Maybe their exists experts that can get this right every time but there are significant risks to damaging a chip desoldering and resoldering. It's not just removing a through hole capacitor.
217.
by emcq 11y ago
That's some very wishful thinking. You have rich people like Robin Williams or Michelle Pfeiffer living here unlike Detroit, not to name countless CEOs and retired VCs who wont disappear. As for the rest of us plebs, SF has survived ma
218.
by emcq 11y ago
Yes, on their site they post Torch CuDNN v4 as faster than Leaf [0]. Seems exciting for an early release. Can it get much faster than something like Torch? I would think if CuDNN is doing most of the computation time it would be hard to see
219.
by emcq 11y ago
Makes sense! You need to minimize tracking error on the underlying basket of assets, and the faster you are the less error you have.
220.
by emcq 11y ago
Isn't the big allure of ETFs that you buy and hold for long time periods betting on the market rather than individual stocks and prices?
221.
by emcq 11y ago
Ah I read his comment as a compliment to Rust, but upon re-reading it seems like he wasn't referring to the "complexity" of Rust as a good thing - what he views as complexity perhaps I view as good design :)
222.
by emcq 11y ago
Rust doesnt have objects, but otherwise is very compelling. I still use C++ but I'm also looking forward to the day (soon) where I can use Rust instead for many projects. With more languages coming to LLVM I expect there will be other
223.
by emcq 11y ago
Perhaps for a naive recommendation system, but quality ones don't do this. For example if you employ 1-nearest neighbors, you will actually return a sample that is guaranteed to reproduce another user's preferences. You could of c
224.
by emcq 11y ago
Drought is an objective measure. Some of us prefer green grass and trees to arid climates.
225.
by emcq 11y ago
What are you measuring in terms of productivity? The employee certainly improves their value in terms of skills acquired during a sabbatical. The employee also gains significantly in terms of value of life experiences when they take time to
226.
by emcq 11y ago
Kind of a weird swipe at bootcamps comparing them to devry. The top bootcamps produce junior engineers better than the average undergrad with more work experience (although in a different field such as mechanical engineering or accounting).
227.
by emcq 11y ago
I think what is being highlighted here is the rather strong statement "you cannot outrun a bad diet" followed by "it isn't realistic to achieve substantial weight loss through exercise" and "base metabolic rate
228.
by emcq 11y ago
If you like this book you will probably also like the gold standard for cellular biology: Molecular Biology of the Cell ( http://www.amazon.com/Molecular-Biology-Cell-5th-Edition/dp/... ). It costs more but it'
229.
by emcq 11y ago
I really like this. It would be nice to see some benchmarks versus a quad based solution where the quad is tessellated in a geometry shader with a fragment shader drawing the particle or using a texture. My guess would be that the bottlenec
230.
by emcq 11y ago
This is not strictly true, and happens when complexity is needed for some of the simple lower level parts or due to the libraries and tools available changing over time. For example, lets say the 10x engineer wrote fast inverse square root
231.
by emcq 11y ago
The no shame in changing positions is key; it describes the difference between humble and egotistical or science and religion. I think you're right that many 10xrs like to debate, and that value can come out of debate. However the best
232.
by emcq 11y ago
This describes me in the past. Don't let him date someone from HR, it opens up Pandora's box about relative value and standard practices with salary negotiation. The largest factor for salary at most companies is how much you ask
233.
by emcq 11y ago
I can think of two people who might fit this bill. Both have a propensity to jump to conclusions and not fully understand various perspectives before passing judgements and weighting their judgements significantly more than others who are m
234.
by emcq 11y ago
Mathematicians, scientists, and doctors absolutely! Catching and dissecting frogs, joining math clubs, debate teams, etc. It's not a requirement but when I think of the best folks in their respective fields they did not start in colleg
235.
by emcq 11y ago
Ones man's anecdotes are anothers statistics. I haven't seen any data driven evidence for this, but let's use a thought experiment: would you expect it is more likely for professional athletes to come from a background of pl
236.
by emcq 11y ago
I completely agree, but there is some historic precedence for this with QuickTime VR.
237.
by emcq 11y ago
Please don't compare the complexity of the brain to artificial neural networks. The complexity of a single neuron is far more sophisticated than we use for machine learning. Many people are starting to call the machine learning "n
238.
by emcq 11y ago
I'm not sure CISC beat RISC, except for complicating the instruction set enough to keep costs high for competitors. Underneath, Intel processors translate x86 to simpler RISC like microcode. They could make a more efficient chip withou
239.
by emcq 11y ago
Makes sense, thanks for the explanation!
240.
by emcq 11y ago
I love your posts, and completely agree that Bayesian Bandits are not replacements for A/B tests. To be fair though, the realworld issues like nonstationarity and delayed feedback are also concerns for A/B tests (which you also br

More ›