wsxcde

Godel's incompleteness is not that complicated. The main insight Godel had was that proofs as well as provable statements can be encoded in numbers. This is obvious to everyone reading HN -- all computing works by encoding things into numbers. And it's easy to see how you'd encode proofs, proof rules etc. into numbers. (hint: as flattened ASTs!)

The thing is, Godel didn't live in an era of pervasive computing so he came up with a wonky encoding based on products of powers of primes. This makes the proof technically challenging, but the fundamental idea is not that complex. What he was able to eventually do was encode a recursive statement of the form "this statement is not provable" where the "this" is kind of like a pointer back to the full statement. The rest is straightforward.

An attack is a reason to believe that you can solve the problem. I have no idea how'd I go about solving P=NP, but I did have some thoughts on provable security against transient execution attacks. Which is why I work on the latter but not the former.

You're trying to come up with a homemade ORAM. All I'll say is that this problem is harder than it seems.

Websockets are subject to the same origin policy. There's nothing you can do to violate the SOP via websockets that you wouldn't be able to do with regular HTTP or XHR.

Your point about the need to "popularize" zk-snarks etc. is well-taken and I upvoted you for it.

But that said, we seem to be talking past each other at this point. I don't see why my criticism of the OP was unconstructive. This is not actually a research paper, it has not undergone any sort of academic peer review, and if the goal is to present it as one example of the kinds of things that are possible to build with zkSNARKs, then it ought to be presented in that context.

I do want to note that it is possible to compose zkSNARKs in an application in ways that result in the composition not being zero-knowledge. Not saying that has happened here, but implying there's no danger with incorrect usage of zkSNARKs seems sketchy to me.

When I clicked the link, I figured this would be a research paper from MIT. It took me a bit of reading to figure out it was just a student project report. That's why I posted the GP.

I stand by my comment that there doesn't appear to be anything particularly novel about this work. Especially with crypto it is important to stick with well-studied and well-understood primitives. There's a danger that someone looks at this and assumes it is being endorsed by MIT and implements it in their system.

This is not a peer-reviewed research paper. It seems to be a project report likely done by undergraduates. The paper is full of typos and it is not clear what the specific novelty of the proposal is (if there is any).

In the real number field, additive inverse of a positive real number is indeed the negative of that number. The negative of that number is also obtained by the application of unary negation operator on the positive number.

This is a fact that follows from the definition of +. But + needs to be defined before you can start making assumptions about what the additive inverse is. The set over which the field is defined (Z or R) already contains -3, -2 etc. and -3 * -2 or -3 + -2 needs to be defined when you're constructing the field. It then turns out that -3 is the additive inverse of 3. You can't use this when arguing about the definition of why applying * on negative 2 and negative 3 gives you the result positive 6. Because you need to define * over all members of the field before you construct a field in the first place.

You are wrong in several ways.

1. You literally cannot prove this fact from the Peano axioms because Peano arithmetic operates on natural numbers, not integers.

2. As I said in my original post at the top, negative numbers are different from the unary subtraction operator (the additive inverse in the field). The number -2 is an entity that exists by itself regardless of whether you've defined an additive inverse. It turns out that the additive inverse of every positive integer is the corresponding negative integer, but this follows from the definition of +, not the other way around.

3. Even if you give OP the benefit of the doubt regarding his dodgy proof, it is saying something about the additive inverse and its relation to multiplication. It is not saying why the result of multiplying two negative values must be positive.

4. The multiplicative operator over the field must already be defined for you to be able to prove distributivity over addition. You can't assume distributivity over an operator that is only partially defined.

--

Think about how you'd define a field. First, you need a set (let's call it Z), then you need two total operators over the set (+ and ), and two elements of the set (0 and 1) and each of these must satisfy specific properties (aka the field axioms). In particular, + and must be defined for all members of Z, not just Z+ and further + and * must be distributive. These are all facts you need to prove about Z, *, +, and 1 and only then do you have a field. You cannot work backward by assuming the field axioms (which are unfortunately named because they are not axioms at all but properties) to derive the definition o the field operators.

My point is that you cannot prove something that is true by definition. The OP trying to prove that the product of two negative numbers is positive is like asking to prove that 0 + 1 = 1 in Peano arithmetic.

The OP thinks that his "proof" is showing why multiplying negative values yields a positive result. But the proof is a load of nonsense because it assumes facts like distributivity of multiplication over addition and subtraction. It is literally impossible to prove that $\forall a, b, c \in Z. (a - b) * c = (a * c - b * c)$ -- distributivity of multiplication over subtraction -- without having already defined the meaning of a * b for all integers! This leads to a circular reasoning loop that the OP's "proof" can't get out of.

The thing to realize is that multiplication is not some magic operation handed down to us by god. It is just a binary total function defined over the integers. What the OP is trying to confusedly get at is the following:

1. There is an intuitive definition of multiplication as repeated addition over natural numbers.

2. It is not clear what the corresponding definition of multiplication over negative numbers is.

3. If we want to define multiplication as a total function over the integers, we need to define what the result should be when multiplying negative integers.

4. Specifically, with (3), we are taught in school that the result of multiplying two negative numbers should be positive, but it is not clear why this seemingly arbitrary choice was made.

Unfortunately, the OP is going about this all backwards. One cannot prove what the OP wants to prove. What one can instead do is argue that the specific (but seemingly arbitrary) definition that one has chosen for multiplication is a "good" choice because it has the same properties (distributivity etc.) as multiplication over natural numbers. At its core, this is a stylistic appeal about the "naturalness" of the definition.

This is not a proof of why the product of negative numbers is positive. The reason why the product of negative numbers is positive is that we define multiplication to be that way.

Also, this post conflates the unary negation operator with negative numbers. The two are not the same. In so far as this post constitutes a proof (which IMO it does not), it is a proof about the behavior of the negation operator.

A good question to ask is why we made this specific choice of definition. Why should multiplication be defined such that -2*-3 = 6? This is a question that the post does shed some light on. If we'd chosen some other definition of multiplication, a lot of the "intuitive" properties of multiplication that hold over the natural numbers (such as the distributivity of multiplication over addition and subtraction) would no longer be true over the integers.

Yeah, many red states have made it illegal to strike. And of course, this eliminates any possible leverage that employees have over employers to force better treatment and that is precisely the reason why Republicans have done this.

The main argument is that NIMBYs are outnumbered at the state level. They "win" at the local level because older, richer homeowners are overrepresented at community meetings.

HN doesn't allow edits after a few hours.

Sorry I had a typo. I meant to say very little of the content is in English. But the page title is in English and that matters because I can google for "Jagran UP budget" to get to the second article.

Some people on the website are complaining that people are putting their video titles in English but the content is in some other language. This is not a lame attempt to get more clicks, but driven by the fact that: (i) Hindi typing works only somewhat well on mobile phones and does not work at all on most desktops/laptops in India, and (ii) search in languages like Hindi sucks even more. Given that most tech savvy Indians are bilingual anyway, it just makes sense to have titles in English so that people can search for them conveniently.

You'll see this is something even Hindi newspapers do. I pulled these two articles from the front pages of the Dainik Jagran and Dainik Bhaskar, two of the bigger Hindi newspapers: [1] and [2]. Very little of the content, except for the title is in Hindi. This is so that people can search for content using English searches.

[1] https://www.bhaskar.com/bihar/patna/news/prashant-kishor-pol... [2] https://www.jagran.com/politics/state-up-budget-2020-yogi-ad...

I actually think it's not as bad as some make it out to be, provided grants are funded at a reasonable rate. What's changed over the last couple of decades or so is that funding rates are single digit percentages. This is the real problem. My advisor said that when he was a younger researcher, it wasn't quite so hard to get funding so he had more time to pursue independent ideas.

Even as a much younger researcher, I feel that at any point in time, I usually have a couple of worthwhile new ideas that are quite likely to pan out and will get funded someplace or another. But if I have to spend a multiple months of effort into getting them funded, I have a lot less time available to generate the next couple of good ideas. This doesn't mean that people just stop doing research, most researchers are competitive workaholics. Instead, we'll just rehash the same old stuff in a new bottle: we target newer applications and/or chase the latest fads (e.g. adversarial ML) with the goal of getting some of our old ideas reused in a new domain.

This creates a negative feedback loop because the quality of research gets worse and the people think this stuff is not worth funding and that in turn reduces funding rates even lower which causes research to get worse.

Unfortunately, scientists working on "useless" ideas wasting public money is a much too convenient bogeyman for politicians to give up.

This is practical advice, but it also points to the reality of scientific funding today -- you can only get money to do things that you already know how to do. You must have an idea that is pretty much ready to go on day one. All that should remain are working out the details.

I can see why bureaucrats and government officials prefer it this way (accountability!) but I don't understand why we scientists put up with it.

Three points are worth making here.

Nobody lives in the Sahara, the population centers in Africa are in the tropical regions and the Mediterranean.

Africa will be the only continent experiencing population growth post-2050. Africa's largest economy, Nigeria, is expected to become the third most populous country in the world by 2050.

The effects of climate change on monsoon-dependent areas (the Mediterreanean coast of Africa and the semi-arid regions bordering the southern parts of the Sahara) are not well-understood and the effects are more complicated than everything will become a desert. For example, the Indian monsoon is expected to produce 10% more rainfall because higher temperatures allow the air to collect more moisture. This is going to be a huge boon for certain semi-arid regions of the country.

This will likely change once Africa becomes the most populous continent, which is expected to happen somewhere around 2100.

The answer is because the script evolved and was not designed so letters appear in somewhat random order.

Brahmi, which is the parent script for most Indian scripts as well as a bunch of other scripts in the region (Tibetan, Thai, Burmese, Javanese etc.), has a well-designed alphabet ordering. Simple vowels come first, followed by dipthongs, followed by constants. Consonants themselves are ordered as velar first, followed by palatals, retroflexes, dental and finally labial consonants -- notice the articulation point is moving from the back to the front of the mouth. Within each group the ordering is unvoiced, unvoiced aspirated, voiced, voiced aspirated and finally nasal consonant. It's a thing of beauty.

Swiggy is a scam! They stole 200 bucks from me (as food wastage charges) because I didn't look at my phone for 15 minutes after I placed an order with a full address that I had already fully paid for.

Not too long ago, my official title was postdoctoral scholar and I was at UC Berkeley. A little while before that, I was a PhD student working with a couple of "postdocs" whose title was postdoctoral research associate. I submit that I do exist as did my co-authors (who have since stopped being postdocs, but continue to exist as researchers.)

This was at another well-known US school. Perhaps thing are different in Britain, but I am not aware of any R1 university in the US where postdocs "do not exist."

I feel both sides are making reasonable points. On the one hand, telling people where they ought to buy their food from does seem like government overreach. On the other hand, there's considerable evidence that people's nutritional choices are terrible for their own and their children's health.

But there's an obvious way to resolve this debate! Run these laws as A/B trials across demographically similar neighborhoods. Track people's health outcomes over the course of the experiment. If the laws seem to be working in making people healthier and happier, keep them. And if they don't, just drop the law.

We're drowning in data! There's no need to go back to the world of Socrates/Plato to try and figure everything out via thought experiments in one's own head!

Thought-provoking article.

I got over my obsession with emulating others' success when I realized that a lot of supposedly successful people were successful primarily because they were also lucky. Once you accept that a big factor in whether you'll end up "successful" is just chance, it becomes easier to accept that it may not be entirely under your control and so there's no reason to obsess about it.

PG's advice, which in some ways is a restatement of the Bhagvad Gita's best-known verse, really speaks to me. Just focus on doing the best you can on the things that actually interest you. The rest will follow, and if it doesn't, that's okay. At least you'll have had fun doing your thing.

Speaking of NYC specifically, a lot of cities across the US, as well as worldwide, experienced a major decrease in crime at the same time that NYC did. Most of the cities were not practicing broken window policing. That's one reason to be skeptical.

From an academic perspective, it is true that there is some debate about the efficacy of broken windows policing, but even the most supportive academic studies find only a small correlation between violations of "order" (like jaywalking and graffiti) more serious crimes. There's just isn't any evidence at all that the way to reduce serious crimes is by going after jaywalkers.

The problems of appropriate policing in so-called "high crime" neighborhoods are well understood. Many academic studies, as well as popular non-fiction like Jill Leovy's Ghettoside and Chris Hayes' Colony in a Nation, have discussed the issue. To sum up the literature in a few sentences, the problem is that minority neighborhoods are both overpoliced and underpoliced. There are a lot of useless arrests for minor crimes like jaywalking which makes the residents of these neighborhoods hostile to the police. (These arrests are driven by the debunked theory of broken window policing.) Simultaneously, there's not enough effort put into solving serious crimes like murder. In this context, the actual effect of predictive policing is that it ends up doing more of the useless over-policing, which unfortunately makes these neighborhoods even worse.

So, how does this relate to your question? The point is that predictive policing is solving the wrong problem. What's needed are not more accurate neural nets predicting crime, but techniques for addressing the underly sociological factors that cause crime.

Taking a step back and speaking broadly, Cory's point is that the focus on data and quantitative analysis is causing problems in two ways: (i) people are using quantitative methods to solve the wrong problems, and (ii) they seem to be oblivious to (and in some cases actively hostile to acknowledging) the harms being perpetrated by their methods. Both of these problems seem to be driven by a lack of understanding of well understood (but non-quantitative) social science literature.

not sure I have noticed a pattern of US media talking it up

Don't you remember when the NYT "credulously" put out stories from Dick Cheney about supposed WMDs in Iraq, which btw were sourced as coming from a "senior administration official" and then Cheney and his cronies then used these reports to actually start a war in Iraq that has resulted in half a million deaths and millions of global refugees? Very little any other country has done could possibly compare with this.

There's no problem with directly calling sycalls from OpenBSD from golang or C or whatever else. The issue here is that OpenBSD is adding a new security feature where code addresses that call syscalls must be whitelisted at link time. golang could just create their own syscall interface and linker scripts appropriately whitelisting these locations and this problem wouldn't exist.

The author has rediscovered the need for software fault isolation (SFI). Bytecodes or IRs like WASM can also provide SFI but are overkill because they provide more than just SFI.

If I were him, I'd have used this as an excuse to play with NativeClient.

--

The original paper on SFI was by Robert Wahbe and colleagues: https://cs155.stanford.edu/papers/sfi.pdf.

Google's NativeClient is a modern take on SFI for x86: https://static.googleusercontent.com/media/research.google.c....