Barraketh

I really love this! The pictures are beautiful, the facts are interesting - it's genuinely pleasurable to see. And if it reminds people to be a little less consumeristic, so much the better!

Some people that are complaining about C++ templates (and suggesting generics from other languages as alternatives) are missing the code generation aspect of templates.

Generics operate entirely at the Type level. For instance (using java)

  List<Integer> list = new ArrayList<Integer>();
  list.add(1);
  Integer i = list.get(0);

is under the hood equivalent to

  List<Object> list = new ArrayList<Object>();
  list.add(new Integer(1));
  Integer i = (Integer)list.get(0);

The differences between the two snippets are entirely at the type level - they compile to the exact same bytecode. Templates are strictly more powerful - they allow you to generate different code for each instantiation. The most obvious place where this is useful is when dealing with primitives that take up different sizes in memory. This is why you can't use generics in Java with primitives - you need Objects, because pointers to objects are always the same size. But if you want performance, so you want to be able to represent collections of primitives generically, then you need the power of templates because you need your code to compile to different things depending on the template parameters.

Yup, but we're all running these apps on laptops, which have a fixed amount of RAM that is sometimes impossible to increase (looking at you, Apple)

Actually I think that you're just looking at the situation from a different angle than many of us. The reason I have some distrust of unions is that they sometimes seem to promote inefficiency as a policy. The distribution of money problem is actually orthogonal to the problem of "how do you do task X most efficiently". If a machine can do the work of 10 men, then to have those 10 men doing the work of the machine is literally wasting their lives.

I wish that unions pursued more contracts of the form "we get X percent of revenue/profits" and then distributed those profits amongst their membership the way they saw fit. This way decreasing the amount of work per person does not decrease the amount of money per person and business and union interests are aligned. I'm sure there are practical reasons why that's difficult though.

That's not quite accurate. Willingness to Pay indicates the true preferences of people who have money. If you try to take it as a metric of the preferences of society you will get the preferences of individuals weighted by their net worth.

I have at various times wanted to start a non-profit that would provide benefits similar to veteran benefits (paid college tuition, paid health insurance, etc) to people who donate kidneys to strangers at the top of the transplant list.

The fact that the kidney goes to the person at the top of the list solves the potential issues of the rich getting better health outcomes because they can pay. Also, we as a society have already accepted these benefits as being acceptable reward for risking life in order to (ostensibly) save American lives in a military context - this simply extends the concept.

I wonder why there are no private non-profit prisons.

There are many states that allow private prisons. There have been multiple articles about how awful these prisons are. Why not try to open a non-profit prison where gov't funds are supplemented with donations and focusing on rehabilitation and humane living conditions. Then we both improve the lives of some inmates, and can more effectively compare outcomes.

Element Analytics | San Francisco, CA, US

-- Who we are --

We're a small startup (<20 people) focused on making heavy industry safer, cleaner and more efficient. To that end we're applying machine learning techniques to sensor data in industries from energy, transportation, and manufacturing, to mining and more. We're solving hard problems for our customers by putting their data to work for them.

-- Hiring process --

You will do a phone screens (about an hour) to make sure there's a good fit, then and 1/2 day onsite where we talk about engineering problems and make sure you know what kind of problems you’d be working on with us. Offers will be extended within 1 day of the onsite interview.

-- Our stack --

We run on linux in Azure, our APIs are primarily implemented using Scala (on Spray), the frontend is written with React.js, and we use python in some machine learning applications as well as for systems programming. Our data processing/machine learning is done in either Scala or python on Spark. Our service infrastructure is docker containers running on DCOS across multiple tenants (both our customers' and our own).

-- What you would do --

We are currently hiring tech leads in both systems, full-stack, and backend engineering. If you are at the top of your game and want to work on a small team solving hard problems, send an email to - sean at elementanalytics dot com

There's a bit more safety than that. When you enter the country on an H1b, you get an I94 that is valid until your visa expires. If you then lose your visa (because you've left the company or got fired), you can still legally stay in the country until your I94 expires. The caveat is, an immigration officer can take away your I94 at any time, at which point you would be forced to leave. However, as long as you avoid immigration officers, you are not breaking any laws by staying and looking for another job.

As someone currently working on fighting advertising fraud, the article premise is wrong. Specifically, there is a significant amount of wasted money that currently goes to fraudsters (usually through RTB exchanges). If we are able to eliminate all of them, then that wasted money will be left to split amongst the legitimate players, which is good for everyone. So yes, CPMs will go up, but the value of each impression will go up even more.

These are all incredibly lucky w.r.t the world at large. However, amongst HN readers and the tech community, there are millions of people who are U.S / Canadian citizens, with educated parents, and who know how to program. What were the 'outrageously lucky' factors that separated Musk from these millions?

I don't disagree (at least w.r.t electric cars), but to play devil's advocate: a factor of 50-100 is A LOT. We're talking about two orders of magnitude here, which would require a fundamental breakthrough in battery technology to achieve parity with gasoline. Even if we make batteries twice as efficient, we're still 25-50 worse than gasoline. And breakthroughs in battery technology are hard to come by. So if Toyota wants to hitch their wagon to fuel-cell technology instead, all the power to them! Batteries are not necessarily the future of portable energy.

Sure, the motor is more efficient, but energy density of gasoline is around 50-100 times that of lithium-ion battery packs[1]. Sure, you can get around that in electric cars because of the more efficient motor and simplified car design. You can remove a bunch of conventional car parts and replace the weight with batteries. Still, I don't think you can dismiss outright research into fuel cells, which promise to be much more energy dense than batteries.

[1] http://en.wikipedia.org/wiki/Energy_density

So then why doesn't New Jersey do this? It is smaller than SK and has comparable population density: 1288 vs 1171 per sq mi. Or how about NYC? Or LA? People keep bringing up USA size or density statistics when talking about government run utilities, but there are plenty of places where they could work given a will to invest in public infrastructure.

I've had mild IBS-D for as long as I can remember, and I've just had a very good experience with a 7 day water fast. It's been 2 weeks since I've finished my water fast, and I've had 0 incidents. Obviously this is anecdotal, since there is very little peer reviewed research on water-fasting in general, and specifically wrt to IBS.

This, if you would like to do it yourself. However, the display advertising space is not currently as turnkey as search advertising, so in most cases you're better off handing off the trafficking of the campaign to an ad network. Quantcast and RocketFuel are two that focus specifically on delivering conversions, and both consistently beat Google on CPA (cost per action). Disclaimer: I currently work at Quantcast.

You are unhappy with the results because you're optimizing towards clicks. You should always optimize towards whatever is important to you. So if you would like people to click through to a specific page, define that to be a conversion and then optimize towards conversions. Optimizing to clicks is not great even if you exclude mobile devices - it's just not what you want.

I could not agree with this more. If there was one law I would pass it would be this - any fines / fees / property seizures get passed on to the state, and ideally distributed evenly amongst the people of the state. The police should not be a funding mechanism.

Really liking Loyal3. The ability to buy into an IPO at the same price as institutional investors is actually pretty cool. As a data point, IPOs priced in the last couple months closed their first day of trading 34% higher[1] than underwriter pricing.

1. http://www.renaissancecapital.com/ipohome/pricings/priced.as...

Standard Silicon Valley business model - get as many people on the platform as possible, and then monetize later (probably as a marketing channel, but they also have some other monetization strategies in the works). They batch their transactions, and they are themselves a broker dealer, so their internal per-trade cost is extremely low.

Also slightly OT, but Loyal3 (http://www.loyal3.com) provides fee-free stock trading - both entry and exit. Unfortunately they don't carry TSLA on their platform yet.

A large problem with software patents is that feature definition and implementation have been rolled into one. Patents were designed to protect implementation - that is, you can't patent having an engine in the car, but you can patent a specific design (implementation) of an engine. In the software world I feel like that distinction has been blurred.

Even if we assume that software is patentable, the bounce back effect should not be patentable. The reason is that all the novelty is in coming up with the feature itself. I don't know whether I could come up with the bounce back effect by myself. I know for a fact that I could implement it, and so could any sufficiently strong software engineer. There is no novelty in the implementation that is worthy of patent protection.

This is trade dress all over again hiding under a different name. It should be given a different status (just like business patents), and different time frames of protection. This way we still protect truly novel research in algorithms (pagerank for example), while eliminating 95% of the patents we as software engineers find so objectionable.