Two announcements
Before my next main course comes out of the oven, I bring you two palate-cleansing appetizers:
- My childhood best friend Alex Halderman, whose heroic exploits helping to secure the world’s voting systems have often been featured on this blog, now has a beautifully produced video for the New York Times, entitled “I Hacked An Election. So Can The Russians.” Here Alex lays out the case for an audited paper trail—i.e., for what the world’s cybersecurity experts have been unanimously flailing their arms about for two decades—in terms so simple and vivid that even Congresspeople should be able to understand them. Please consider sharing the video if you support this important cause.
- Jakob Nordstrom asked me to advertise the 5th Swedish Summer School in Computer Science, to be held August 5-11, 2018, in the beautiful Stockholm archipelago at Djuronaset. This year the focus is on quantum computing, and the lecturers are two of my favorite people in the entire field: Ronald de Wolf (giving a broad intro to QC) and Oded Regev (lecturing on post-quantum cryptography). The school is mainly for PhD students, but is also open to masters students, postdocs, and faculty. If you wanted to spend one week getting up to speed on quantum, it’s hard for me to imagine that you’d find any opportunity more excellent. The application deadline is April 20, so apply now if you’re interested!
Follow
Comment #1 April 7th, 2018 at 4:23 pm
On 1.,
I wonder if the liberal left in the United States is so worried about potential fraud in US elections, why do they oppose so vehemently the requirement that proof of US citizenship be show at the ballot box?
One of the most glaring contradictions I have seen in American liberals -specially those in academia- is their irrational worshiping of everything European Union themed -particularly their welfare systems- but then at the same time their opposition to something as logical and rational as the common practice over there of requiring a valid ID at the ballot box.
I see efforts like 1. in the same light as the studies Joshua mentioned in https://scottaaronson-production.mystagingwebsite.com/?p=3654 : a way for the losers of the 2016 election to feel good about themselves.
To reiterate: it wasn’t a Russian troll or Putin that made me vote for Trump in 2016. The voice of people like me wasn’t captured by polls like this https://www.scribd.com/document/330170543/9amET-NBCWSJ-release that excluded from their polling universe those that had not voted in 2014 or 2012.
At which point will the losers of the 2016 election apply Ocham’s razor to their analysis of what happened then? Is the Rasmussen poll, which shows Trump’s approval level at this time of his presidency to be similar to Obama’s in April 2010, also lying http://www.rasmussenreports.com/public_content/current_events/politics/prez_track_apr06 ?
Was CCN’s Tal Kopan also a Russian agent in October 2016 when she wrote https://www.cnn.com/2016/10/19/politics/election-day-russia-hacking-explained/index.html “No, the presidential election can’t be hacked”. The whole article is worth reading. One of the experts quoted for example said:
“Nobody is going to be able to change the outcome of the presidential vote by hacking voting machines. The system is too distributed, too decentralized, too many implementations for any individual actor or group to make substantial change,” said Nicholas Weaver, a computer scientist and cybersecurity expert at the International Computer Science Institute at the University of California, Berkeley.
Move on people, move on. You want to have a Democrat win the 2020 election? Make sure your party picks a candidate that speaks to the issues that middle America careas about. Hint: it is not climate change, increased China favoring globalization or increased welfare checks in lieu of well paid jobs that don’t require a PhD in Computer Science.
Comment #2 April 7th, 2018 at 7:32 pm
Gatekeepers #1: I know from experience that I won’t change your mind about anything. But for the benefit of other readers, I thought it would be useful to answer your points briefly, since they have clear, decisive answers that are far from universally known.
I wonder if the liberal left in the United States is so worried about potential fraud in US elections, why do they oppose so vehemently the requirement that proof of US citizenship be show at the ballot box?
Because we know that voting by non-citizens is not a problem that exists in reality, at least to any non-negligible extent. If it were, then right-wing groups’ many “voter fraud investigations” would surely have turned up some serious evidence for it by now. For in contrast to, let’s say, quietly hacking the code shared by most or all electronic voting machines in a given state, shepherding non-citizens to the polls is an absurdly dangerous and expensive way to rig an election. Each fraudulent voter would provide yet another opportunity to get caught and for your whole scheme to unravel.
The only purpose of “proof of citizenship” requirements, is to try to suppress the votes of eligible voters (like college students and the poor) who tend to vote Democratic, and who are also likelier than average to have no idea where their records like birth certificates are.
One of the most glaring contradictions I have seen in American liberals -specially those in academia- is their irrational worshiping of everything European Union themed -particularly their welfare systems- but then at the same time their opposition to something as logical and rational as the common practice over there of requiring a valid ID at the ballot box.
The obvious difference here is that most EU countries, unlike the US, issue national identity cards. If the US were to do that, instead of its current crazy patchwork of repurposed forms of ID that not all citizens have (birth certificates, Social Security cards, passports, driver’s licenses…), I’d have no objection to making the presentation of the ID card a condition for voting.
“Nobody is going to be able to change the outcome of the presidential vote by hacking voting machines. The system is too distributed, too decentralized, too many implementations for any individual actor or group to make substantial change,” said Nicholas Weaver, a computer scientist and cybersecurity expert at the International Computer Science Institute at the University of California, Berkeley.
This is precisely the claim that Alex Halderman’s research has decisively refuted. He’s actually demonstrated how a hacker could insert malicious code into a software patch, which would then get installed in a large fraction of the electronic voting machines in a given state, in such a way that current security practices would completely fail to notice anything unusual. Not speculated about it, demonstrated it. Read his Senate testimony.
And crucially, even if Alex is unusually good at these exploits, he’s just one guy with a few grad students. He lacks the resources that, let’s say, the Russian or Chinese intelligence agencies could throw at the problem.
And you don’t need to think it’s likely that voting-machine hacking changed the outcome of the 2016 election—or for that matter, be a liberal or a Democrat—to be worried about how easy such hacking is now known to be. Worried enough that the simple fix of a randomly audited paper trail, for every vote cast in the country, would strike you as a total no-brainer, something to get done by 20 years ago.
Comment #3 April 7th, 2018 at 8:03 pm
Scott #2
The notion that the US doesn’t have a uniform ID system is not very persuasive because under the United States’ federal system, it is the individual states that organize elections, not the federal government. Every time we have a federal election what we have is 50 different states organizing elections. I could be wrong about this, but every American state I am familiar with has the ability to issue state certified IDs to any American citizen who lives inside its borders, be it a driving license or an ID without driving privileges. If you know enough Europeans -which I would imagine you do- you’d know how much this issue puzzles them irrespective of their politics. The notion of not requiring proof of citizenship as a precondition to registering to vote -let alone voting- is antithetical to the idea of ensuring clean and fair elections. I am not the first to point this out. Here is an article from 2012 at Foreign Policy on the topic: http://foreignpolicy.com/2012/11/06/foreign-election-officials-amazed-by-trust-based-u-s-voting-system-2/ .
Note that this argument,
“The only purpose of “proof of citizenship” requirements, is to try to suppress the votes of eligible voters (like college students and the poor) who tend to vote Democratic, and who are also likelier than average to have no idea where their records like birth certificates are.”
can be used both for and against the issue of requiring IDs at the ballot box and both carry equal weight. So I don’t find it particularly persuasive as to why the US should be unlike every other civilized country in the world when it comes to the issue of requiring an ID at the ballot box. I would add that it is a well known thing that dead people tend to vote Democratic too.
With respect to,
“This is precisely the claim that Alex Halderman’s research has decisively refuted. ”
Actually he hasn’t. I have followed his efforts for quite some time and I remain unpersuaded by his arguments. Note that even the election he talks about in the NY Times video he is talking about a mock election about a non issue when it comes to presidential politics (whether the University of Michigan is better than Ohio State University).
That’s the main problem with his arguments. That he has been able to hack some machine on a laboratory controlled environment and that he was able to apply that hacking technique to a mock election about some irrelevant topic does not prove anything when it comes to hacking elections across 100000 + voting precincts there are in the United States on any presidential election with the real issue of who will control the executive branch of the US government for the next 4 years at stake. His arguments were/are best received among the losers of the 2016 election, which include a large portion of American academics and Jill Stein who used his arguments as a powerful fund raising mechanism. Those who make a real living in politics don’t think much about them. The US Senate hearing on the topic was, like most US senate hearings are, political theater.
Comment #4 April 7th, 2018 at 8:21 pm
Scott #2 again.
Let me elaborate on this,
“That he has been able to hack some machine on a laboratory controlled environment and that he was able to apply that hacking technique to a mock election about some irrelevant topic …”
Hacking on a controlled environment and hacking in the real world are two very different endeavors. To make an extreme analogy, it’s like the capability of developing nuclear weapons. The theoretical principles behind the operation of nuclear weapons have been well known since the 1950s. They are available to anybody who purchases a standard college physics textbook like this one https://www.amazon.com/gp/product/0131495089 . Now, making working nuclear weapons is a totally different endeavor.
With hacking is kind of a similar situation. There is a reason why the NSA recruits, for its cyber operations, people with unusual backgrounds like Edward Snowden. They are the people who have the practical expertise to pull off actual sophisticated hacking of the kind that would be required to flip a US election. There is little evidence that anything of this sort happened in the 2016 election. Alex Halderman’s research at most shows that it is theoretically possible to patch these boxes, but then again, that was already known well before he decided to become a tool for some Democrats and Jill Stein in their grievances about the 2016 election.
Comment #5 April 7th, 2018 at 10:21 pm
@ “The Problem With Gatekeepers:”
The problem with state-issued IDs is that they certify identity, but do not certify citizenship. It is completely legal for legal residents and green-card holders to get drivers’ licenses. True, you could ask people to show a state-issued ID, but anyone can get one, even without being a citizen. All a driver’s license means is that you are certified to drive.
Further, motor vehicle bureaus have no way of determining whether an applicant is a citizen. Neither does anybody else.
Even a valid birth certificate for an individual born within the United States is not valid evidence of citizenship. The 14th Amendment defines U.S. citizenship, and it says the individuals born here are automatically citizens only if they are subject to U.S. law. The families of foreign diplomats are not subject to American law, so, say, the son of the French Ambassador to the U.S., born in America, would not be a U. S. citizen. But there is no mechanism to assert on a Birth Certificate that the holder, though born here, is not a U. S. citizen. To that imaginary son of the French ambassador could show his birth certificate and pass as a citizen rather easily.
If we wanted to get serious about this, we would require anyone registering to vote to prove citizenship. Having a valid driver’s license proves nothing, but for a naturalized citizen, naturalization papers, together with proof of identity, should work. For someone native born, we’d probably have to accept a birth certificate and just live with the fact that a few offspring of diplomatic families would squeak through. In either situation, a U.S. passport should suffice, with the same caveat.
I would contend, however, that only a small minority of bona-fide U.S. citizens possess one of the three documents cited: birth certificate, naturalization papers, or passport.
@Scott, you contend that there is no reason to believe that large numbers of non-citizens are voting. I’m not sure that is true, but it is certainly possible. As to your objection that right-wing groups would have brought forth thousands of examples, bear in mind that it is as hard to disprove citizenship as it is to prove it.
One more thing. One does not necessarily require U.S. citizenship to vote in local and state elections, as far as I know. I could be wrong on this, but a state could possibly allow anyone who works there and pays taxes to vote. If my guess is correct, then, the proof of U.S. citizenship for federal elections has to be deferred to the feds, which would have to set up a separate mechanism to certify U.S. citizenship.
Comment #6 April 7th, 2018 at 10:59 pm
S3CS 2018 looks very interesting! I probably can’t make it (been putting off lining up a grad degree because I’m too busy teaching), but I’m grateful for both lecturers making lecture notes available.
Comment #7 April 8th, 2018 at 9:48 am
Peter #5
That’s a fair analysis.
Ultimately, as the article http://foreignpolicy.com/2012/11/06/foreign-election-officials-amazed-by-trust-based-u-s-voting-system-2/ indicates so much of the US electoral system relies on trust and good faith that if one is going to try to find holes in it, our politics tend to color the particulars we have a problem with. Take the issue of hacking. Up until November 8th 2016, the left was almost unanimous that it was impossible to hack the system in any meaningful way for the reason outlined by Nicholas Weaver above. I contend that had Hillary Clinton won the election, that consensus would still be with them today even if actual hacking had happened. It’s only the end result that bothers people like Alex Halderman . It’s a similar situation to the misuse of Facebook data. When Obama affiliated forces did it to help him win the 2012 election the left was OK with it, now that it’s been shown that Trump affiliated forces did it -albeit apparently with little success https://www.theverge.com/2018/3/20/17138854/cambridge-analytica-facebook-data-trump-campaign-psychographic-microtargeting – it’s time to demand Marc Zuckerberg’s head. That’s just silly.
Similarly, the main reason the right takes issue with the possibility of non citizens voting (both legal and illegal) is because new immigrants tend to vote Democratic.
My own take is that unless we are talking about a very close election, these things cancel out. The 2016 election delivered Trump with a decisive victory (anybody doubting this should take uber blue California out of the popular vote equation to convince himself/herself). The recount effort in Wisconsin confirmed the result. And the recount effort in Michigan actually showed that http://insider.foxnews.com/2016/12/14/steins-recount-turns-more-votes-voters-detroit “in heavily-Democratic Detroit, more votes than voters were found across the city and now state Republicans are calling for an investigation”. So if there was any voting fraud in Michigan, it would seem it was on the Democratic side. Trump still won the state.
As I said in my first comment, these obsessions won’t help the left win any future elections. The policies put in place during the Obama years on the economic side only helped a minority of Americans: those highly educated with jobs at large American corporations and their subsidiaries/contractors. No society -large or small- can be sustained long term with such an economy. Universal basic income won’t cut it. Most people want to have middle class lives working meaningful jobs, not receiving government paychecks. The problem of the Obama economy was best captured by this http://money.cnn.com/2016/09/08/news/economy/us-startups-near-40-year-low/index.html . Even something as American as domestic migration had slowed down during those years. With the Trump economy it has resumed https://www.usnews.com/news/best-states/articles/2018-03-28/america-is-on-the-move-again-and-this-is-where-theyre-going .
The Republicans lost in 2008 and 2012 -and were headed for a defeat in 2016 as well had it not been for Trump- because they continued to live in the past and had lost touch with reality. To continue to push things like TPP in 2016 after the fiasco of giving free ride to China is as tone deaf as it comes. Both Trump and Bernie Sanders understood it. Neither establishment (Democrat or Republican) did.
Comment #8 April 8th, 2018 at 11:59 am
LOL. I suppose white bread (probably the most common palate-cleansing foodstuff) could also be used to clean a palette, but I doubt it would be very appetizing afterwards.
Comment #9 April 8th, 2018 at 12:57 pm
Aula #8: Thanks, fixed.
Comment #10 April 8th, 2018 at 1:28 pm
Scott, you are incorrect that the video is simple enough for all or even most members of congress to understand.
Comment #11 April 8th, 2018 at 2:33 pm
Somehow related to the discussion on 1.,
https://www.usatoday.com/story/news/politics/2018/04/04/facebook-gave-most-contributions-house-committee-question-zuckerberg-also-got-most-contributions-fac/486313002/
“Members of the House and Senate committees that will question Facebook CEO Mark Zuckerberg about user privacy protection next week are also some of the biggest recipients of campaign contributions from Facebook employees directly and the political action committee funded by employees.
The congressional panel that got the most Facebook contributions is the House Energy and Commerce Committee, which announced Wednesday morning it would question Zuckerberg on April 11.”
“Of the 55 members on the Energy and Commerce Committee this year, all but nine have received Facebook contributions in the past decade. The average Republican got $6,800, while the average Democrat got $6,750.”
Expect to see a lot of theatrics when Mark Zuckerberg testifies on Wednesday, a lot of “fake outrage” from both sides -the Democrats hammering him with the Cambridge Analytica scandal, the Republicans with the idea that the Obama campaign did the same thing in 2012-, but very little substantive action to follow.
In the United States, it is still the case that money talks. Big Tech has been “talking” to both sides of the aisle for quite some time now.
Comment #12 April 9th, 2018 at 1:37 pm
A new article by Aubrey D.N.J. de Grey claims to have partially solved the Hadwiger-Nelson problem, that is, the chromatic number of the plane with an edge between unit distance points: “https://arxiv.org/abs/1804.02385” . What shall we do? Wait for Gil’s comment?
Comment #13 April 9th, 2018 at 3:52 pm
Dear Scott,
I agree that we should use paper ballots. However, I thought the video was not very clear. How exactly was the election run, and who ran it? If Halderman was in charge of running the election, it shouldn’t be a surprise that he was able to hack it.
Comment #14 April 10th, 2018 at 6:48 am
jonas #12: Wow! I’ve had a soft spot for the Hadwiger-Nelson problem since I first learned about it at a math camp in 1996. I’ll be very interested to see whether this holds up, and wouldn’t venture a guess right now. Given the need to run a brute-force coloring algorithm on a 1567-vertex graph (and the lack of use of automated verification tools), I guess people will fully believe it when and only when it’s confirmed using an independent code base.
Incidentally, is this the same Aubrey de Grey who’s famous for life extension research?
Comment #15 April 10th, 2018 at 6:54 am
87987 #13: I thought the video made it reasonably clear: the election was run using standard electronic voting machines, identical to ones used in real elections, after Alex and his students had been given about 30 seconds to open the machines and load new software into them.
(While you obviously couldn’t see this part from the video, it’s software that silently alters the results, then deletes itself and leaves no trace that it was ever there, except for that provided by a paper trail.)
And of course running a “Michigan vs. Ohio State” election was just a publicity stunt (the good kind, I think); you can read Alex’s papers to learn about their actual research results.
Comment #16 April 10th, 2018 at 7:47 am
Scott: Tobias Kildetoft claims that it is the same biologist Aubrey de Grey: “https://chat.stackexchange.com/transcript/message/43907768#43907768”
Comment #17 April 10th, 2018 at 5:28 pm
Further Hadwiger-Nelson update: The 1567-vertex graph doesn’t work (apparently there was a bug in the code that produced it) but the revised code produced a 1585-vertex graph which has now been independently verified via SAT solver. https://gilkalai.wordpress.com/2018/04/10/aubrey-de-grey-the-chromatic-number-of-the-plane-is-at-least-5/#comment-40108
And yeah, seems to be the same Aubrey de Grey…
Comment #18 April 10th, 2018 at 7:06 pm
Also wow I had no idea so many people cared about infinitary combinatorics…
Comment #19 April 11th, 2018 at 2:50 am
Sniffnoy: Erdős cared about exactly this sort of stuff, colorings of infinite graphs. That makes mathematicians flock there, because we know Erdős usually asked the right questions.
Comment #20 April 11th, 2018 at 7:44 am
jonas #19,
I don’t think the Erdős connection is all that is going on here. The problem has the virtue of being very visual and also very easy to understand. The fact that the prior best bounds of 4 and 7 can be explained in a few minutes makes it even more appealing as a problem.
Comment #21 April 11th, 2018 at 6:10 pm
jonas #19: I know Erdős did a bunch of stuff regarding infinitary combinatorics, but I don’t heard it talked about much!