The situation is very similar in India. You can't just convert INR to USD freely. There's a limit of $250,000 per person per year. I don't know if anyone mines bitcoins in India (because of horrendously expensive power), but if they did, they'll be subject to similar scrutiny.
Unless an order comes from the Supreme court, you can conveniently ignore all these messages. So far, the Supreme court has not mandated mobile/payments app linking with Aadhaar. Given the recent win in the privacy battle, Aadhaar proponents will find it hard to mandate linking Aadhaar with every goddamn financial activity.
This is my only gripe with Apple. It's an incredible PITA to share devices even temporarily. You have to log the primary user out of their phone, login with your credentials and we are in business. This whole process is so long and frustrating that it forbids any sharing experience. I do wish Apple or someone made sharing devices a lot easier.
All of that is guesswork if you don't understand what the CNN is doing at each stage.
Well, what the CNN is doing at each stage is very simple to understand. There is a forward pass which is a matrix multiply (and addition of there is a bias) and then the matrix weights are learned in the backward pass, which is just basic differentiation and chain rule application. Now I am not trivializing differentiation ( when you try take the differntial of a vector, you are tearing your hair apart) but it's fundamentally a simple concept to understand.
Even with this understanding, designing deep neural nets and tuning hyper-parameters is mostly guesswork. Yes, the frameworks have little or nothing to do with this.
What I've found is that TensorFlow is difficult to wrap your head around for programmers because it's more like a DSL. You declare a computational graph and then run it multiple times. So when you are declaring a computational graph, you have no way of debugging that graph, unless you run it. Also the conversion from numpy arrays to Tensors and back is an expensive operation.
PyTorch simplifies this to a great extent. You just create graphs and run like how you run a loop and declare variables in the loop. This is great for imperative programming. However, think about it - every time your graph is recreated. Now if it's just a variable re-initialization it's not a big deal, but we are dealing with Tensors so you give up efficiency for flexibility.
Again, all of this is immaterial for learning how to build deep neural nets. I would say, just stick to whatever framework you can wrap your head around. I am learning that my ability to tweak numpy arrays, visualize them in pyplot, load data from csvs using Pandas and the like will take me a lot further in learning deep learning.
I think the single largest obstacle for the older mind to learn is time. A lot more responsibilities curtails your ability to spend as much time as you want to learn a new skill.
Well, sarcasm aside, it's fairly common to dismiss a new technology that can grow exponentially, because as humans we are tuned to linear thinking. In 1985, McKinsey speculated that there would be 900,000 cellphone subscribers overall globally. Now we can just dismiss McKinsey as just consultants, but guess what, any one in their position would have probably speculated around the same number because that's what the models and the Math would show.
Quite similarly, almost everyone is speculating on SDCs based on existing dogma and numbers, not taking into account how rapidly deep neural nets are evolving. Just look at the Imagenet competition history. I mean today we have neural nets that are better than humans at image classification. I'm quite sure, if you had asked even the most eminent researchers to speculate in 2012, they wouldn't have predicted such an outcome.
It is also easy to take the position that I am taking, that SDCs will be a reality in the next 5 years, just based on hubris and some Elon Musk cool-aid. However my view is purely based on the growth that we are seeing in deep neural nets.
As an Indian, who has lived and worked in USA, UK, Singapore and LatAM, I can attest that produce in supermarkets in all these geographical regions are tasteless. Mostly because, as someone mentioned below they are now bred for ease of shipping.
Even in India, vegetables are slowly going the tasteless route, particularly chillies and garlic, but they are a lot more flavorful than their western world counterparts. I remember as a kid garlics were so pungent, you could smell them from a mile. Not the case anymore.
Yeah am feeling quite the same. Nevada Energy is another such suspect company in his portfolio. Oh well, I think it's time to take a bearish view on BRK. Their main cash cow - vehicle insurance is gonna get upended very soon, probably in the next 10 years.
What really baffles me is that the likes of Warren Buffett support WFC despite uncovering such wrong doing. I mean, what WFC did is completely unethical. Buffett, who practically epitomizes ethics, should have taken some action financially. He acknowledges that what WFC did was wrong, and he also acknowledges that more wrong doing might be uncovered. He still believes that WFC is a good long term investment.
One possible reason could be that this malpractice stemmed from one or two individuals who are now fired. So there could be reason to believe that the bank as a whole is still fine.
I am not as experienced as you are and I don't even know the term re-entrant functions but accidentally this is how I am now designing programs, mostly writing self contained functions that don't alter any global state. This approach has made my programming life hell because I am finding it incredibly hard to get rid of the old shared variables habit.
What I have experienced is that shared data access is more or less inevitable and I design to restrict the shared data access at the database.
Tech has made real interactions hard, not just parenting. How many times have we sat for dinners or lunches with friends staring at our phones. I say we because I am guilty of this behaviour.
It is not hard to fathom a future dystopian world in which most human interaction is through a tech medium.
Apple will not buy any US company using it's cash hoard because of tax implications. They can of course raise debt, similar to what Amazon is doing with Whole foods.
I think, more than the outcome, it's the window of uncertainty that burns through you. Gamblers get a kick because the uncertainty window is fairly short. I'd bet that they'll flip when faced with an extremely long uncertainty window, like closing a corporate sales deal, or waiting for a job offer.
This has been my experience as well. What I've found helpful is to just take a decision real fast and move on. Sure, it might be a bad decision but it's a done deal nonetheless. I tend to feel less burned out even if the decisions turned out to be bad, vs spending an inordinate amount of time debating decisions. And I have gotten myself stuck in some real holier than thou decision wormholes, for instance, which programming language to learn now.
Well doesn't this sort of support Damore's hypothesis ? Some of the smartest girls I know went into marketing, purely because they just loved that field. Somehow to them sitting in an office in front of a computer all day didn't seem that appealing.
Is it safe to infer that, in th developed world, given a career choice women have a propensity to not choose tech ?
I have to agree that there was a lot of noise in the nosql world and a lot of people got carried away by the ease of use metric. I guess people were lured by the prospect of not having to write SQL joins.
That said I found nosql databases extremely helpful for storing and querying large unstructured data. Mostly because it was really hard to build relations and to store this data in tables. Think Wikipedia for instance. Since then my way of choosing databases is to try and model data into a relational db as much as possible and if that doesn't work out choose a nosql equivalent.
Unlike more common smaller stars, such as the red dwarf stars Proxima Centauri and Trappist-1, they are not so faint that planets would be tidally locked, showing the same side to the star at all times.
In such planets, the most habitable zone is around an equator like region where the light and dark regions kind of merge to produce a reddish sunset like hue all through the day. I think one of the planets that Kepler discovered is like that. Life would evolve to absorb these light wavelengths. So for instance plants would all look black. Nova has a great episode on these exoplanets. https://www.youtube.com/watch?v=5HZsFMqqGJo&t=793s
I am rooting for cryptocurrencies just for cross border payments. What we have today is abysmal and pathetic. We pay a lot of overhead sometimes 3% + 2.5%. I think circle or some startup is about to offer near zero cost cross border payments. We might never need a payment gateway !, per my understanding of the protocol at least.
I heard of a study recently that showed that cities in Japan that ended up getting Shinkansen stations ended up almost always shrinking, because people got used to going to Tokyo all the time and ended up just moving there.
This is so counter intuitive. This is the exact opposite of suburbia. Faster and cheaper transport led to the growth of suburbs. Can the movement of people to Tokyo be attributed just to the faster trains ? Can you please share a link to the study if you have it handy.
The article has a taste of criticism but offers no points of criticism. Sort of a passive-aggressive play on words, which is ok just for a read, but not something that I expect from New Yorker.
Comparing Hyperloop to Concorde and it's eventual demise does seem valid at a cursory level. What's key to understand is that Concorde's failure can mostly be attributed to a combination of design, safety and maintenance factors (sonic boom being the foremost). In addition, what plagued the Concorde was the upkeep and maintenance of the aircrafts, and of course the spiraling fuel costs. Are there such equivalents in the Hyperloop world. If so, can they be resolved.
Why was this underwhelming ? Elon unveiled the most important aspects - speed, range, safety, price and the range anxiety question - how many superchargers.
I agree that his delivery style is underwhelming and there's a lot of debate about that. I am personally fine with an underwhelming presentation and an overwhelming product.
I bet if we start testing self driving cars in India, we'll get there in perhaps another 10-20 years. What's needed is an extremely good path planning algorithm. Rest of the self driving tech SLAM et al is already solved.
However, the bigger problem on Indian roads is congestion. I think you need smaller autonomous vehicles that can navigate through the crevices and cracks ( gallis ). Something smaller than the Tata Nano.
Ah, another ban. Welcome to the new age governance in India. Some government official in Karnataka wanted to ban ride sharing as well. I hear the government's sentiment about creating jobs, but hey have these guys ever sat in that shitty truck and driven all night ? Fucking armchair politicians.
Anyways, I used to think why hasn't any company invested in self driving tech in India - and now the answer becomes pretty clear. One can possibly take the government head on and challenge the ban in the Supreme court.
I'm not kidding, but I proposed doing something like this as an open source project for the Segment open source fellowship. I didn't qualify but I am keen on doing this. A few priorities have taken over lately, looking for a job primarily but I will get there. My goal is slightly different than what you have mentioned here. I want to start with a few tagged images and use clustering to get more images, tag them and increase the training set volume.
I've seen the same behavior in London tube and Singapore MRT as well, despite display boards that indicate removing backpacks in crowded trains.
Mumbai locals handle this in the best possible way, you simply cannot get on the train during peak traffic if you have a backpack. I mean you probably can wiggle in, your backpack might not :)
Yes, you mean well and I don't mean to imply that people are misinterpreting your words. People understand things differently. They read correlation and imply the converse. I have done it so many times. Another analogy - Alan Turing was a great computer scientist, he was also a long distance runner. The thought process in my head after reading that sentence, "To be a great computer scientist it's good to start long distance running". This is because I like running and my brain perceives it as the path of least resistance to achieve mastery over a subject.