Cover photo
Verified name
Yonatan Zunger
Works at Google
Attended Stanford University
Lives in Mountain View, CA
137,363 followers|131,406,306 views


Yonatan Zunger

Shared publicly  - 
We have reached the ultimate evolution of the clickbait headline. Apparently its final form is a sort of phoenix-like death and rebirth, involving cockroaches.

In retrospect, that doesn't surprise me as much as I thought it would.

Via +Jennifer Ouellette​
The circle of life is gross. It’s even grosser when you throw in a cockroach, larvae and some violently ravenous fire ants.
Rennie Allen's profile photorare avis's profile photoAlexandra Pedro's profile photo
The video actually delivers what it says: a cockroach giving birth while being eaten by ants (I watched it. It's fascinating.)

What I have issue with is that the person filming keeps iterating that this is for educational purposes, gets philosophical, but makes sure everything is clear and visible and feeds live cockroaches for entertainment. The video also uses OMG in the title. Kinda giving mixed messages there.
Add a comment...

Yonatan Zunger

Shared publicly  - 
The national Fraternal Order of Police has endorsed Trump. I did not think I could become any more disappointed with American police, but they have found another way.

It is clear that they see Trump as deeply representative of their priorities and likely to stand behind them no matter what. Unfortunately, this makes clear what their priorities are. Police unions have decided that their first and foremost principle is to protect individual officers from any form of accountability, up to and including for rape and murder; they apparently have also decided to include white supremacy in their formal charter.

If you combine this with other police union statements in the past few days - like the Miami union's saying that they will not provide police protection to the Dolphins football team until and unless the team forces its members to stand during the anthem - it has become painfully clear that police unions across the country have converged on a belief that any opposition to them, any suggestion that their power should be less than unlimited, is "anti-cop."

I have always been suspicious of the notion of public sector unions, but police unions have gone so far beyond any prospective worst case of how such a union could behave that their very existence has become unconscionable. The armed forces of a state must always be subordinate to civilian oversight - and a police union which can demand exemption from this, and threaten violence or public disorder (as Miami's just did, and as many others have) if it is not granted, is an enemy of democracy itself. 
Rob Gordon's profile photoMiss Melony's profile photoRugger Ducky's profile photo
Earl appears to be from Oklahoma. From where he apparently spews violent threats and rhetoric (as well as his horrifying grammar and spelling), all over. But mostly only on NRA websites.
Add a comment...

Yonatan Zunger

Shared publicly  - 
I don't know what it is about this scene, but I loved it tremendously as a kid, and time has not taken away a bit of its shine.
Robert “Bob” Brewster's profile photoChristopher Tate's profile photoRoss Peter Nelson's profile photoBenjamim Silva's profile photo
"and time has not taken away" :-) unfortunately The Great Dictator's Speech (1940) still, perfectly applicable to our days.
Add a comment...

Yonatan Zunger

Shared publicly  - 
Psittacosaurus lived in Asia between 123 and 100 million years ago. They grew up to two meters long and weighed up to 20kg, (about as much as a bulldog), walked on two legs, had relatively large brains for dinosaurs (comparable to modern birds), and ate some combination of plants, nuts, and seeds.

And thanks to a bit of luck, a fossil Psittacosaurus was found in spectacular condition, with skin intact enough to both show off the structure of its cloaca (that all-purpose orifice it shares with modern birds) and to recover actual pigments from its skin.

Using this and other fossils, a team at the University of Bristol have performed the most detailed reconstruction of a dinosaur ever done: from figuring out its facial musculature (which you can determine from skull thickness) to the coloration of its body. This, in turn, has allowed more discoveries: for example, from its camouflage pattern, we know it lived in a relatively dark environment, like under a forest canopy.

The article below is full of details and pictures: everything from what this odd creature looked like, to how it behaved, to how we figured it all out.

h/t +Stefani Banerian
Alex Foster's profile photoSailen Naiya's profile photojiezel hambalos's profile photo
Wow 😀
Add a comment...

Yonatan Zunger

Shared publicly  - 
This is a very interesting technical challenge. Pixelation blurring is a traditional way to conceal people's identities in photos, and it works well because it interferes with the mechanisms our brains use to recognize faces -- spotting the points of the eyes, nose, mouth, and so on.[1]

Like any other recognition system, natural or artificial, this is based on the system seeing a bunch of these things and learning which features differentiate one thing from another. If you were to start a recognizer by showing it faces both blurred and unblurred, and telling it which were the same, then it would learn a very different set of features to look at than our brains normally use -- say, patterns of light and dark. And it turns out that these patterns are pretty good at identifying faces, too. They aren't as good as the ones our brains use for telling each other apart, but what they are good at is recognizing someone even when they've been blurred.

The practical upshot of this is that it's surprisingly easy to train a computer to recognize faces that have been blurred.

From a security perspective, this is very important, because we often rely on these kinds of obfuscation to conceal data -- even life-or-death data, when concealing the identities of people like confidential informants or political dissidents.

But we shouldn't focus too much on this one special case. Instead, we should look at this as a special case of a much broader phenomenon. In theory, things which look similar can be distinguished in all sorts of ways. I'll bet that the pattern of pores on the back of your hand is unique, for example. Whenever we use some kind of "blurring" -- that is, systematic information loss -- to eliminate information, the right way to analyze it is not whether people can still extract the original information, but by truly information-theoretic approaches: does enough information exist anywhere to recover that data?

For example, if you're trying to conceal identities: There are under seven billion humans on the planet, which means that in theory 33 bits of information is enough to uniquely identify anybody. In practice, a lot of information is noisy, but if you reduce an image of a person down to 64 bits, there's probably enough information there that somebody (either now or in the future) could reverse the process and figure out who it was.

This kind of validation -- of using mathematical approaches to validate that enough information has been destroyed to make something unrecoverable -- is similar to the ways in which cryptography is validated. There, you're trying to show that a certain minimum amount of computation would be needed to recover the plaintext; here, you're trying to show that no computation can recover the original.

Various practical versions of this have been coming up. Recently, a Russian company started releasing a program which took pictures of people and searched over the Internet to find other pictures of them and figure out who they are. While it was billed as helping people spot celebrities, its primary use seems to be in identifying sex workers and similar people, and then either outing or blackmailing them. Why does it work? Because it's suddenly possible to tie a picture taken today to a picture taken a decade ago, and show that it's the same person, something which wasn't easy to do a few years ago.

Today, I sometimes see news outlets cropping people's heads in an effort to anonymize them. (Especially for victims of crimes, etc.) This feels like an example of the same mistake: bodies are just as distinctive as faces, it's just that we don't normally look at them that way. But it would be almost surprising if nobody could figure out a way to identify you by a picture of your torso.

And there will always be subtleties that you didn't think of, unless you are literally attacking it from the perspective of "are there enough bits of information?" For example, if you were to completely remove a person from a picture and replace them with a region of perfect black, would the reflections off their skin on other objects in the picture be enough to figure out who they are? Certainly not to a human eye, but I wouldn't rule it out in general without some serious computation. Likewise, your camera captures a lot more information than you see -- by which I'm not talking about location metadata, but about different colors of light and pixel variation which is only in the RAW file, not in the jpg's that normally circulate.

The moral of this story isn't an easy one. It's not "ban people from recognizing things," because in general, people will. Instead, it's a combined moral about treating information destruction with the same sort of seriousness which we do cryptography, and about recognizing that just because something is secure today doesn't mean it will be secure forever.

More research needed, as they say.

[1] The human brain has an extraordinary fraction of its volume dedicated to nothing but face recognition, which is why we're generally so much better at it than computers. This system trains during our youth, which is why people tend to be significantly better at remembering and distinguishing faces similar to the ones they've seen in early childhood. (Thus the "All X look alike" effect) It's also helpful for distinguishing things that look sort of like human faces, such as dogs' faces, but we aren't nearly as good at that -- and we're downright terrible at telling jellyfish apart. Only they know the difference.
God Emperor Lionel Lauer's profile photo
Add a comment...

Yonatan Zunger

Shared publicly  - 
From the department of "the more things change:" +Gretchen McCulloch's question was answered with a "yes," and a passage by Bokenham from 1440 getting angry at those kids these days with their "corrupcioun of Englysshe men yn þer modre-tounge."

(That's "corruption of Englishmen in their mother tongue," for those of you who have so debauched their notion of English that they can't understand the plain and ordinary speech of the mid-15th century. Seriously! What's with education these days?)
Amanda Rachelle Warren's profile photoMatt Schofield's profile photoScott Watson's profile photoAndres Soolo's profile photo
+Chris Welty: That's such an authoritarian sentiment of child-raising that I can believe it of Plato, but not really of Socrates.
Add a comment...

Yonatan Zunger

Shared publicly  - 
Alas, this post is more of "a few days ago, I learned... something very unpleasant." But because many people who follow this collection may know Scott Lewis, it's important that I post it here.

Scott Lewis is someone I've known for quite a few years via G+, and he was part of my circle of friends. And ever since this article appeared, accusing him of running elaborate and very nasty confidence games on people, I've been having a lot of conversations where people are saying "wait... this wasn't a one-off?" and giving a lot of details.

A lot of pieces about why various interactions with him seemed odd over the years are suddenly clicking in to place.

At this point, I have very strong reason to believe that not only are all of the allegations in this post true, but they represent a substantial underestimate of what he's been up to. AFAICT, Lewis has been running long cons against people for many years, probably an average of 4-5 per year, taking an average of a few thousand dollars from each victim. And this is not one of those "he didn't know he was being manipulative" situations; it's clearly extremely deliberate and systematic, the work of an experienced con artist who then uses a nasty combination of emotional manipulation, social isolation, and blackmail to keep people from talking about it to either one another or the police.

People I know have been victims of his; other people I know have been attempted targets of his, who are now figuring out just what that weird shit going on was. He may have even recently gotten the idea of trying to target me, using how hurt he was by these allegations and how he wanted to "open some sort of dialogue" as the opening. Given that most of his stories begin with telling victim #N how badly hurt he was by victim #N-1, I'm going to file this under not fucking likely, buddy.

I don't know which, if any, of his cons are likely to be police matters; this sort of thing is notoriously hard to prosecute. But I can offer advice to those who know him: if he's spending a lot of time talking to you, reminding you of how close your relationship is to him and how much you've been helping him deal with how hard everything is in his life, watch out. This guy is not kosher.

Pamela L. Gay – Verified account @starstryder. Another day, another #astroSH long con revealed. I've confirmed this story across multiple sources. I believe her. …
Rhonda  McAlister 's profile photoDeen Abiola's profile photoAmber Peall's profile photo
How did I miss this?! I can only imagine something was up with the algorithms and it got buried for me to miss it from so many of my favourite plussers.

Echoing sentiments here AND the story elsewhere. 
Add a comment...

Yonatan Zunger

Shared publicly  - 
This article is both a great and a terrible example of writing about technical issues for the public.

The body of the article is the great part. What it's explaining is an idea called "formal verification," where you can construct a mathematical correctness proof that verifies that a piece of code does exactly what it's supposed to, and nothing else. Formal verification has been an idea for decades, but until recently it was nearly useless, because the specification for "exactly what it's supposed to do" was nothing more or less than the entire program itself, which was just as prone to subtle errors. But the past decade has brought major advances, and there are now meaningful ways to specify what a small piece of code is supposed to do in a concise fashion, and then formally verify that the code does so.

If you apply this technique to small but critical pieces of code, like the network protocols which underpin the Internet, or the central control logic of an armed drone, or the control code of a cardiac pacemaker, this can give you a tremendous improvement in system reliability and security, especially for systems where it really matters.

The terrible part is what was stapled onto this article by some overly-excited editor: a headline reading "Hacker-Proof Code Confirmed." To anyone who works in computer science, this headline is approximately as reassuring as "Iceberg-Proof Ships Confirmed" would have been in 1913. It is, in fact, palpable nonsense, and misrepresents what happened here.

The experiment done was a very important one: a drone was designed to be controlled by code, where its central logic was verified using these methods, and then a very skilled attacking team was given network access to the system and were free to use any number of methods to compromise it. Their failure is of great practical significance, because it means that the security of the system has indeed proved robust so far.

But the headline might lead you to believe that this system is somehow "hacker-proof," or worse yet, that this technique might make the Internet as a whole "hacker-proof."

The first rule of security is that a chain is only as strong as its weakest link, and a wall is only as strong as its weakest point. You can have a perfectly unpickable lock, but if someone can simply kick in the door or climb in the window, that doesn't make your system secure. Formally verified code is a component of a secure system, but it isn't a secure system in its own right.

Here are some examples of things which could nonetheless make an armed drone with this code vulnerable:

* The part of the code which isn't formally verified could be vulnerable, and someone could use it to compromise the behavior of the vehicle in some other important way -- say, to cause a gradual oil leak which makes the system seize up. This especially includes the connections between the formally verified pieces and the rest of the system.

* The hardware itself could be compromised, with backdoors in the behavior of the chips. And before you start asking whether chip designs could also be formally verified, take a look at this attack that I posted about a few months ago:

In that attack, every single thing about the chip's behavior was unchanged -- unless you caused a particular wire in it to flip on and off rapidly enough, at which point for physics reasons which have nothing at all to do with the way the circuit is wired, another wire would magically flip on. You would have to model not the circuit, but the actual physics of the device, down to atomic resolution, to detect this -- and to do that, you would need to input not the physical model you started with, but the actual as-built physics of the thing.

* The signals leading into the system could be compromised. This is how chaff works to distract a heat-seeking missile, but it could be either more or less subtle than that. This could range from compromising its control channels over the radio (which means breaking some crypto), to confusing its IFF (friend/foe) detection by emitting false signals yourself, to simply hiding under a tarp where it can't see you.

* The people could be compromised. Nine times out of ten, this is the biggest vulnerability in any system, and it can range from infiltration by hostile actors, to "social engineering" that fools people into doing the wrong thing (look up "CEO email scams" for an example), to the highly effective cryptographic technique known in the field as "rubber-hose cryptanalysis."

(This technique works by finding the person who has the decryption key, and beating them with a rubber hose until they tell it to you. Surprisingly effective, and it has variants which are even more so.)

So whenever you read an article like this, read it with a careful eye. In this case, the meat of the article is quite accurate and interesting -- but the headline is meant to lead you astray. In general, if someone is promising you something that sounds too good to be true, it probably is: stop, think, and talk to people who might know when you hear that.

h/t +Don McArthur

Nathan Hourt's profile photoTerry Bixler's profile photo
+Nathan Hourt
 Maybe on government projects or certified secure projects:)
Add a comment...

Yonatan Zunger

Shared publicly  - 
There is nothing I can add to this.

Except possibly "Lyanna Mormont, Age Four."
‫آدم نمیدونه چ اسمی انتخاب کنه‬‎'s profile photo

Yonatan Zunger

Shared publicly  - 
Software updates often leave users of older hardware in the cold. Jason Scott has worked hard to remedy this, and thanks to his efforts, you can now run ProDOS 2.4 on mainline-series Apple ]['s, not just the IIgs.

This fixes most of the remaining compatibility issues across the series, and should be a welcome relief for those who have been put off by the IIgs' rather high cost ($999, or about $2,150 in current dollars) or otherwise unable to upgrade.

However, even with the system upgrade, you'll still need a IIgs to support the 320x200 and 640x200 graphics modes. Them's the hardware constraints...

Via +Emmanuel Florac
In September of 2016, a talented programmer released his own cooked update to a major company’s legacy operating system, purely because it needed to be done. A raft of new features, wrap-in p…
Kee Hinckley's profile photoTodd Vierling's profile photoGod Emperor Lionel Lauer's profile photo
+Todd Vierling I wouldn't have thought that the H8 would be that expensive, even fully-loaded.
Add a comment...

Yonatan Zunger

Shared publicly  - 
I support this idea.
Clapham Common Tube station in London is looking very feline today. After a very successful tongue-in-cheek Kickstarter campaign, the so-called Citizens Ad
6 comments on original post
Nathaniel Watkins's profile photoDavid Washington's profile photoJames Mason's profile photorare avis's profile photo
As much as I love the felid variety, I was kinda hoping for an altogether different kind of cat... more along the lines of Fred Astaire or Audrey Hepburn in Funny Face.
Add a comment...

Yonatan Zunger

Shared publicly  - 
Last week, I posted a question about whether there's a meaningful region of America that stretches as a sort of "ring around Appalachia:" containing eastern Oklahoma, up through parts of Arkansas, then through Missouri, Illinois (bordering on but not including Chicago), and curving around into Indiana and Ohio. Eastern Texas, western Pennsylvania, and northern Kentucky are other areas which came up as maybe belonging.

[ for that conversation]

A great conversation followed, with the tentative conclusion being that this may be an area which is becoming a meaningful region of the country: you don't see it as very distinct on many measures of the instantaneous state of the world (like distribution of accents, rates of heart disease, or economics), but you do see it on measures of the rate of change of such variables (like the declining life expectancy for white men and women, or nascent political affiliation).

One of the really interesting questions which came up a few times is whether this is tied to the ancestral history of the places. In particular, this area is roughly the area with a large Appalachian population, minus the heart of Appalachia proper: that is, it's the area to which a lot of people from that area moved over time. This is related to a hypothesis proposed by Colin Woodard in his book American Nations, which argued that various seed populations of Americans, which came in with very different cultures and expectations, left behind meaningful social and political differences which last to this day.

I'm personally a bit leery of Woodard's hypotheses, especially when it comes to anything west of the Mississippi. (He seems to group together a lot of culturally very different groups there, and miss a lot of similarities) But there are some good arguments that he has captured an important idea about the eastern part of the country.

On that thread, +Matt MacMahon shared a link to a fascinating article which uses a lot of data to argue for Woodard's hypothesis. It's full of maps of different things, and is arguing that these cultural lines, more than things like population density, account for modern political divisions in the country. You can read it here:

One of the best maps from that article is one that I'm blowing up below. It was drawn by Chris Howard, and it shows the 2012 Presidential election results. Many maps of this sort suffer from a serious visual presentation problem: they make the country look overwhelmingly red, even though the majority was blue. This is because of two simple problems. First, red is perceived by the human eye as "brighter" than other colors: people shown identical areas of bright red and bright blue (or green, or anything else) will see the red area as bigger. Second, the traditional maps ignore population density: what they really show is that the majority of the land area was red, while what we really are asking is about the distribution of the people.

Howard solves this problem very elegantly by using the hue of each county (from red to blue) to show how it voted, and its saturation (from pale to bright) to show its population. The resulting map is not only easy to read, it shows up the real distribution of both people and votes much more clearly than any other map I've seen before.

Here you can see an outline of the rough area discussed in that original post. It is purplish but leans distinctly towards red, unlike the purple/blue areas further north in Iowa and Minnesota. It is adjacent to the bright blue dot in Iowa / Chicago / Wisconsin (discussed in that blog post), passes near the bright red of Appalachia, is bounded on the East by the dense blue/red patchwork which describes the South, and on the West by the sudden and sharp drop in population density as you enter the Agricultural Midwest.

The thing I like about this map is that the visually distinct areas in it seem to match well with my intuitions about the politically and socially distinct regions of the country. That's something very few maps manage to capture well, so major kudos to Howard for finding the right distribution.
Chris Pugrud's profile photoKevyneShandris's profile photoHolly Mularski's profile photoBilali M'boni's profile photo
Add a comment...
Yonatan's Collections
  • Google
    Chief Architect, Social, present
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Mountain View, CA
Boulder, CO - Rehovot, IL
Head of Infrastructure for the Google Assistant
The obligatory (very important!) disclaimer: I'm not on this system as an official representative. While I'm listening to user feedback and interacting about the system, I'm also here for perfectly ordinary social networking purposes. If I am saying something official on behalf of Google, I will make that explicitly clear; anything else that I say here is not the position of Google, or of anyone other than myself.
  • Stanford University
    Ph. D., Physics, 2003
  • University of Colorado, Boulder
    B. A., Mathematics, Physics, 1997
Basic Information