Hacker News Reader

JesperRavnon Russia bans all GMO food production

For those mocking Russia for this, note that they are joining Germany and France in enacting these laws[0]. While I believe in democracy and freedom, I also think that anti-Russian sentiment is being stirred up for the wrong reasons, by people who see Russia as an enemy of the US.

[0] https://en.wikipedia.org/wiki/Regulation_of_genetically_modi...

JesperRavnon How Bernie Sanders Tech Volunteers Code the Bern

Thanks a lot. Yes, Trump is very wealthy, but on the other hand there a lot of big wealthy companies who would fire their employees for saying the things that Trump does. Universities even more so. The middle classes are basically forced away from the positions that Trump holds. And so really the big money is against him.

JesperRavnon Surprisingly Turing-Complete

Just some more on dependently typed languages: The article actually mentions them in footnote 1. So as it stands, the CSS part really doesn't make sense, because everything written about CSS applies to these dependently typed languages.

In particular, you could write some Agda/Coq/Idris program that implements rule 110, and has some trivial user interaction to repeat this until a stable pattern is reached. This simulates a Turing machine.

JesperRavnon UX Designer Salaries Around the World

I think there is a huge opportunity to move more software development to developing countries. It could become the next version of sweatshops, which have done an immense amount to improve the wellbeing of people in the third world. The winners of using cheap labor from overseas are the companies that hire them and the people they hire. The losers are the workers in the first world who are outsourced. But it's crucial to understand that both nations as a whole benefit, and the first world nations are more than capable of redistributing wealth.

When you cut through all the self-interest and xenophobia surrounding outsourcing, there still remain many quality and communications issues. But I believe these will be reduced as more people learn English and recognize the value of English as a market skill.

JesperRavnon How Bernie Sanders Tech Volunteers Code the Bern

I would be very interested in volunteering for Donald Trump. Does anyone know how I could get involved?

I like the way he is the only candidate with the guts to oppose illegal immigration, even though it should be obvious to everyone that illegal immigration is wrong and must be stopped. It's especially ridiculous given that almost all liberal candidates support much worse in Israel, while calling Trump a racist for proposing relatively mild measures in the US.

JesperRavnon How Bernie Sanders Tech Volunteers Code the Bern

This thread is about how great it is that volunteers are coming together for a good cause. Let's keep politics out of it.

JesperRavnon Surprisingly Turing-Complete

While I question the tone of the original post, it is correct.

The difference between being Turing complete and implementing rule 110 is crucial, and any discussion on the subject needs to make clear that these are not the same.

You are obfuscating the role of the driver. In the Turing machine, the driver is part of the Turing machine. In CSS, it is not, and has to be done by the user (or presumably a JS script).

As another post by Grue3 points out, almost any language can implement rule 110, including languages that are not Turing complete, and are explicitly said not to be Turing complete in the CS literature/community.

For example, Coq, Agda, Idris (with totality checking) and other languages based on Martin Lof type theory, are not Turing complete. That it why they are able to prove totality of their functions. But they can all implement rule 110 with no problem.

JesperRavnon The Hidden Wealth of Nations: Tax Havens

No, increasing competition for labor is good for everybody in the long run.

JesperRavnon Study links U.S. polarization to TV news deregulation

I think there is also a political dynamic in the US, where the inconsistencies in moderate politics have to be resolved one way or another, and people resolve them by choosing extreme views.

By inconsistencies, I mean that anyone who is moderately left-wing with regard to Israel, will be called a crazy anti-semite (as Rand Paul is) unless they are to the far left on domestic politics. Similarly anyone who is on the moderate right on domestic politics is expected to be far right with regard to Israel and the Middle East.

Anyone with moderate, consistent views would be called anti-Semitic by the right, and racist by the left.

JesperRavnon American Innovation Lies on Weak Foundation?

It's just not right. Given the existence of options, shorting is no more bound by liquidity than being long on a stock.

Imagine saying "Stock X is great, but the market can undervalue X longer than I can stay solvent"?

JesperRavnon American Innovation Lies on Weak Foundation?

American corporate labs are the stuff of legend. Researchers dream of the halcyon days of Bell Labs and its eight Nobel Prize winners, who brought us the transistor and Unix. Others reminisce about Xerox PARC, which came up with the graphical user interface that propelled the personal computer into just about every home and office.

Those days are long gone...

What about commodity clusters (i.e. all cloud computing), MapReduce/Hadoop, Spark, or the deep learning renaissance? Yes, I am super biased, but I think Google and Facebook have done as much to advance technology as their predecessors.

But apart from that they have a point. Raw numbers matter, and even though R&D might have a fuzzy boundary with ordinary business, I think that R&D expenditure is a useful signal.

JesperRavnon Epistemic learned helplessness (2013)

Well that's a good example of what the original argument was talking about since the comedian's argument was terrible (e.g. by the same logic we should not say that slavery is immoral, because the ancient Greeks practiced slavery).

JesperRavnon A deeper theory of testing

There is also a correlation between having an academic background, and having specialized knowledge that is inherently hard to communicate with non-academics. Just because the format was a blog post, doesn't mean that the author was aiming at a general audience. Even if it was possible and desirable, the author might not have had time to flesh out all the concepts that are taken for granted in their field.

I liked the general ideas in the blog. People think about these things informally all the time, but it's nice to see the ideas laid out systematically.

Maybe you can wait for someone to write a blog expanding on these ideas for a more general audience, or write one yourself.

JesperRavnon Epistemic learned helplessness (2013)

I think the author makes some good points, but I think the term "learned helplessness" together with the arguments made, imply that if reasoned arguments can't be trusted then nothing can. But I think that a better characterization is that people use other ways of thinking, such as intuition, emotion, etc. in addition to reason.

Any expert in a field other than pure math, will tell you that in order to evaluate an argument or evidence, you need experience and intuition, not just argumentation. Arguments are not a formal process that can definitively arrive at conclusions. The same applies to mathematical models. Economists often complain in private about the need to dress up their ideas as mathematical models, even when the model adds nothing to the discussion. This is especially annoying in empirical work, when no one is actually interested in the model, just the empirical results.

JesperRavnon For San Francisco Cab Drivers, Once-Treasured Medallions Now a Burden

I see a lot of comments suggesting that medallion owners should be compensated. But the government is not morally obligated to compensate people whose investments lose money due to changes in regulation. And it would be prohibitively expensive to do so, because it is a one-sided affair. No one who made money investing in medallions when they went up in price was arguing "when we bought medallions, we thought that the number of available medallions would increase with demand, resulting in a stable price. Instead, prices shot up, so we feel we should pay back the government some money because their unexpected policy change benefitted us".

The government might sometimes compensate people, but that would depend on many factors such as 1) To what extent the person was unable to avoid the risk, and to what extent they willingly bet on the price of the medallion, and 2) Whether the government did all in its power to keep policies predictable, given the unpredictable nature of technological change.

JesperRavnon Chris Poole

4chan (and it's successor 8chan) mix two tendencies. One is to post horrible or strange things for the sake of it. The other is to post things that have some validity or purpose, but would be censored in other places. Now clearly what things one puts in what categories are very subjective, but there are certainly people who claim to belong largely to one or the other group (as well as neither). But so far there hasn't been an approach to allow the second category but not the first, so people who have views that don't fit into the mainstream (mostly on the political right) tend to congregate at this sort of place.

JesperRavnon Most Americans will get a wrong or late diagnosis at least once in their lives

The flipside of that is that patients might bias doctor's decision away from what was a good decision in the first place. I've had many American doctors be somewhat apologetic when they recommended no treatment, or a less invasive treatment. My response was that if they didn't recommend a treatment then I didn't want it either. There are real risks from any treatment.

The only time I ignored advice was when I went to a doctor for a cough, and because I mentioned chest pain (from coughing), they decided to get an EKG, which showed some abnormality. Apparently this abnormality is not uncommon in healthy people. They recommended more tests with a specialist, but I figured that a not uncommon result on a test, when doing that test wasn't justified by my original symptoms, didn't amount to enough evidence to go any further.

JesperRavnon The Bitcoin Community Disagrees on What Happens Next

No that you mention it, I think that would be possible.

The main problem is that by its competitive nature, mining rewards are actually deadweight loss. Every dollar value of bitcoin that is given to a miner is value that is destroyed. To see this, imagine that a Chinese bitcoin mining farm is making a profit mining bitcoins, gaining $10 of bitcoins for every $8 of operating expenses. Then another farm could start up with an even lower profit margin. This would continue until the economic profit was zero, i.e. every dollar's worth of bitcoin mined cost a dollar in operating expenses, which are deadweight loss to the economy.

However, none of this implies that a constant inflation rate is not possible. As long as a dollar's worth of bitcoin generates more than $1 of value to the economy, the system should be stable, and the higher this number, the less wasteful bitcoin would be. Given the various figures for velocity[0][1] I've seen, it's not clear whether this system would be prohibitively wasteful or not.

[0] https://en.wikipedia.org/wiki/Velocity_of_money

[1] Note that velocity is the ratio of total money to nominal GDP but since one is a stock and the other is a flow, this doesn't measure how much economic activity a single dollar generates/supports.

JesperRavnon The Bitcoin Community Disagrees on What Happens Next

The equilibrium price would be set by the demand for money-as-currency. But in the limit of instantaneous (electronic) settling of payments, this demand premium converges on zero, with the only residual demand being for money-as-store-of-value.

Let's start with the empirical evidence. Velocity of money has not grown much at all, over the last 50 years[0]. Surely if the issue was mainly technological, then technological changes up till now would have resulted in a bigger change in velocity.

The alternate explanation is that the transaction value of money is driven by other kinds of economic frictions, in particular agency problems. Although it's not very popular, Holmstrom and Tirole have some theoretical work on the relationship between collateral, liquidity and agency problems. But the general idea is quite intuitive: any contract, whether debt or equity, changes the incentives of the counterparty and results in moral hazard. This places a limit on how much we can rely on credit instead of money.

That this logic is applied to hard money and not to smartphones is most parsimoniously explained by the political incentives that led to the abandonment of hard money in the first place,

Given the above, do you still think that the you know the simple obvious truth and that mainstream economics has ignored it out of willful blindness? Does your parsimonious explanation apply to me?

[0] https://en.wikipedia.org/wiki/Velocity_of_money#/media/File:...

JesperRavnon The Bitcoin Community Disagrees on What Happens Next

The economics professor stated As money, Bitcoin is terrible – a deeply deflationary currency that’s within a bubble. I'd like to expand on this a bit.

The deflation issue is complex. It's hard to imagine a stable situation where money has a fixed supply. The simplest steady state one can imagine is constant economic growth r, which is also the real interest rate. If we further assume a constant velocity of money (so the total value of all money is a fixed percentage of GDP), then any currency where the total supply of currency was fixed, would have a deflation rate equal to the real interest rate. Therefore holding currency or zero interest bonds would be equivalent. But money has some transaction value that makes it more valuable than zero interest bonds, so this situation could not be an equilibrium.

Most economists think that in practice there is no way to have a stable currency except by forcing inflation by printing money (which in the current implementation is injected into the economy by buying bonds). I don't think it is possible to create an inflationary currency except through central banks. So I really think that bitcoin faces long term problems.

JesperRavnon Why Bitwise Reproducibility Matters

Yes, I agree that the issue is in the toolchain. And more reproducibility in this area is a much more general and important problem. Even without aiming for bitwise reproducibility, it's still very hard to properly package your software and its dependencies. That is one reason why reproduction is so rare: simply installing the dependencies is too hard.

So a sufficient improvement in toolchains/package management etc. would probably enable bitwise reproducibility, but I think a lot of people in this thread, and the original article, were implying that current tools are sufficient and scientists just need to stop being lazy.

JesperRavnon The Hottest Restaurant of 2081

That is a really good point. Perhaps one could compare it to creating and driving an car from the 1920's, and trying to satisfy all the regulations around cars nowadays.

On the other hand, I do think that combination of extreme scarcity leading to government regulation is so out of line with how the world is moving, that it reflects flawed thinking. It's almost as if people want some kind of crises to escape the decadence of modern capitalism (and there are as many on the right who think this way). The only thing that could create this kind of scarcity is global warming. But maybe the reference to a farmer on the arctic circle was a reference to this?

JesperRavnon The Hottest Restaurant of 2081

Of course, but Robotness would probably be an improvement over humaneness given our tendency to ignore animal suffering as long as it's out of sight.

JesperRavnon Apple meets California officials to discuss self-driving car

TrueNorth chips would probably be a competitive disadvantage. I found Yann LeCun's critique[0] compelling. In short, simulating the brain via spiking neurons is less efficient than just doing the floating point math like regular neural nets do.

[0] https://www.facebook.com/yann.lecun/posts/10152184295832143

JesperRavnon Why Bitwise Reproducibility Matters

Given that floats are already bitwise reproducible (provided the compiler is appropriately shackled), what does this idea add to reproducibility?

JesperRavnon The Hottest Restaurant of 2081

Interesting idea, although it has the same failings as classic sci fi and much of the modern left when it comes to economic illiteracy (in particular agricultural efficiency being mandated by law instead of left to the market).

Also projections of population don't suggest any more than doubling by 2080 at the maximum. And raising and killing animals humanely should become easier with improvement in robotics technology.

JesperRavnon Why Bitwise Reproducibility Matters

I don't think the effort needed for this is worth it. Most scientists even ones who code, don't even know what bootstrap binaries are.

My original point is that the complexity of dealing with differing environments and code versions, combined with the sensitivity of floating point results on the exact source code and environment, practically kills the possibility for exact reproducibility.

That's why I'm claiming that a better goal is approximate reproducibility.

JesperRavnon Why Bitwise Reproducibility Matters

In principal there is nothing wrong with what you say, it's just that you need to now keep track of every possible change to the code and the environment.

For what I am proposing, keeping track of the major versions of dependencies should be enough.