emcq

15 ms·

91.
by emcq 10y ago
I was on opiods for a month. At the time it was prohibitive to get out of bed, let alone go to the hospital. I dont remember it much but sleeping on my own bed beats the heck out of the hospital.
92.
by emcq 10y ago
Back 5+ years ago Qt was my go-to framework for cross platform GUIs. You would get the kitchen sink, but the abstractions provided were quite sane. Signals and slots, nice OpenGL and painting libraries, reasonable asthetic and layout option
93.
by emcq 10y ago
I observed 3 things about women in boot camps that explain this phenomena outside of affirmative action: 1. They are self selecting to jump into the known gender gap in engineering. 2. Perhaps due to 1, they generally came from STEM backgro
94.
by emcq 10y ago
I don't think it's luck that this happened to you. All companies with HR departments that care about attrition should compensate their employees fairly. If you weren't, I suspect you would have left. Contrast this to a small
95.
by emcq 10y ago
With a roommate and no rent control monthly cost is 1500-1800. For a studio ~2500. 1br closer to 3000. 3 bedroom homes for a family are easily 6k+. If you are young and don't mind living with others or in small spaces you can save a lo
96.
by emcq 10y ago
One key predictor to patient outcomes is the experience of a surgeon. There are tons of papers pointing to how long the learning curve for a technique is. This isn't rolling the dice. I don't want an overly confident surgeon. I wa
97.
by emcq 10y ago
Based on this and subsequent posts, it seems like you're making an "ends justifies the means" argument. Does the ends justify the means here? Not really. Lyft is arguably similar to Uber but without so many alleged and eviden
98.
by emcq 10y ago
Unfortunately, I dont think this will happen. So much in society revolves around delicious food and relaxing beverages. How would a business make money from someone fasting?
99.
by emcq 10y ago
Apologies if I'm being too academic here but not all memory bottlenecks or communication bottlenecks are the "Von Neumann Bottleneck". The term was originally focused on both data and program memory being on the other side of
100.
by emcq 10y ago
I appreciate healthy discussion of technical topics. However, I'm not sure you're having this discussion in good faith. I wrote this response in case you are. LeCun never said anything about the Von Neumann Bottleneck. TrueNorth i
101.
by emcq 10y ago
I'm sure today we could create an _even better_ chip focused on the advancements in neural networks and chip design, but it's awesome that a chip from 2012 can still take us so far! AlexNet was just a baby then. I doubt any CPU, G
102.
by emcq 10y ago
The near future is not in putting learning inside a chip. We're a long ways off from the one shot learning needed to make a device actually interesting with localized learning. Instead, the future is recording and uploading your observ
103.
by emcq 10y ago
The reality is that TrueNorth can use CNNs and be trained with backprop to achieve good performance with networks like ImageNet despite LeCun's comments [0, 1] LeCun's primary critique is that binary wont work; "to get good r
104.
by emcq 10y ago
For training networks, designing en efficient chip is a scary proposition. However for inference tasks, low precision GEMM (as you said) goes a long way and better than what you often get. That's why chips like Movidius' Myriad ar
105.
by emcq 10y ago
Yes, this is traditionally a big effort in the neuromorphic community originating in Carver Meads lab at Caltech. Perhaps the best success of such analog computation efforts are with Neurogrid (full disclosure I worked with this group): ht
106.
by emcq 10y ago
What's wrong with training in full precision? There's a huge market for efficient inference in robotics, devices, and autonomous vehicles.
107.
by emcq 10y ago
The market as a whole is going up. Are your seeing above market average returns?
108.
by emcq 10y ago
What's new in transfer learning?
109.
by emcq 10y ago
I had no suggestion that people converse in first order predicate logic. I did suggest we are all smart enough here that we can have rich discussions without committing clear logical fallacies. HN is special in that the majority cares about
110.
by emcq 10y ago
Why is it that people like scratching the hypocrisy itch? It's not even a logical fallacy. In fact, pointing out hypocrisy as a way of negating the merits of an argument _is a logical fallacy_: https://en.wikipedia.org/
111.
by emcq 10y ago
We should be careful about reading too much into this election. There are many well off republicans. Based on the statistics I've seen, Trump had _fewer_ low income voters than Hillary. In addition Trump has similar percentages of male
112.
by emcq 10y ago
XNOR-Net is not currently part of TensorFlow, although there is a feature request to support it. [0] The main difference is that XNOR-Net uses 1 bit quantized weights and XNOR+popcount to approximate the dot product of convolution, which ca
113.
by emcq 10y ago
This is for optimizing inference. For consumer devices such as smart cameras, speech recognition on your phone, or robotics that's the part that matters. They have real speed improvements by using 1-bit quantitzed weights and then runn
114.
by emcq 10y ago
Yes it is used. But there is a whole wealth of more general additional techniques based on probabilistic graphical models. A good reference for those curious is probabilistic robotics: http://robots.stanford.edu/probabilisti
115.
by emcq 10y ago
And supported by prior lawsuits against Oracle: https://www.theregister.co.uk/2014/01/13/oracle_sued_over_pa...
116.
by emcq 10y ago
I think the reality is that some blocks will be lazily erased. I believe modern flash firmware periodically GCs in the background to prevent write amplification. If this is triggered you may overwrite data. NAND requires an erase before wri
117.
by emcq 10y ago
There was a time before computer science degrees existed. Guess who were some the first folks to use computers and pioneer our field? Physicists and mathematicians! Von Neumann, Feynman, Metropolis, Vapnik, Babagge, Lovelace, Turing, Church
118.
by emcq 10y ago
Craft your resume around the job you want. Add in the other details to fit. If you dont know what you want, do more internships, soul searching, or perhaps even explore career coaches. The best thing getting my food in the door in silicon v
119.
by emcq 10y ago
The average spending per individual in the US per year is $10000 [0]. For an all inclusive insurance plan, if you have say a $5000 deductible and are spending less than $416/month per person you are contributing less to the pool than i
120.
by emcq 10y ago
Sure there are plenty of undigested elements that inevitably become waste products, but if it's food that most people are eating, I'm very unconvinced that calories in minus calories out isn't the first order effect to model.

More ›