What ThinThread can tell us about early voting

I just left a Trump/Vance rally in Grand Rapids. As I left the lot, I immediately got a call from a local Republican Party rep asking me to commit to "vote absentee or vote by mail". What? Why?

I’ve been thinking about this “vote by mail, vote early” topic for a while now, and after today’s phone call, I think I finally understand why it’s now being pushed by the Trump campaign as well as by the Democrats.

I don’t like it; but I understand it.

For the record, my personal preference is to conduct voting using a secure digital public ledger, i.e., recording votes securely on a blockchain. It is the best method for conducting, counting, and securing the vote that I know of. I’ve thought it through in depth and have defensible ideas about how to make it work from a technical point of view.

I’ve written about this topic tangentially before, for example in this post about Cyberoperations from November of 2021. I’m not a lightweight when it comes to this topic, given my background in networking and technology.

When Brian Cates went to the Cyber Symposium in Sioux Falls, SD in late 2021, I spent a few weeks before that event teaching him the basics of digital packet capture technology so that he would know what to look for and ask about with regard to the claimed “PCAPS”. While the PCAPs were never produced and the event didn’t lead to much of value, Brian and I became friends from that point forward.

The ideas behind blockchain voting require a non-trivial amount of explanation to convey to the average person; to that end I’ve created a PowerPoint slide deck with the basic ideas behind blockchain voting that I share with friends and family.

I may cover this idea in a future post here on Substack, or maybe in a video. But it’s not something that very many people are ready to embrace, so let’s set that topic off to the side for now.

Many conservatives have been persuaded/inoculated over the past few years with a “get rid of the machines, vote by hand counted paper ballot only” mind virus, which I’m actually not in opposition to.

It’s not the best method, in my view, but it’s better than what we have now.

I would agree that paper ballots hand counted on election day is far preferential to black-box (closed source and insecure) voting machines which are involved at multiple layers in current US elections—everything from digital pollbooks to touch screen ballot entry to ballot scanners, ballot counters and aggregation nodes, as well as county and statewide aggregation and reporting servers.

I’m not naive about the flaws in this system, having created an explainer video for Catherine Engelbrecht of True the Vote back in 2020 and having also helped research and uncover the Konnech and Eugene Yu issues in 2022.

I’m also fully aware of the egregious security flaws found in the machines used in 2020 in Arizona, Colorado, and elsewhere. They are a huge vulnerability (but blockchain can solve almost all of their flaws.)

What I know, however, is that hand counted paper ballots don’t entirely eliminate the “machines” that are used in election processes to aggregate and report the results; they just eliminate machines at some of the intermediate layers. They also may not help if “counterfeit” ballots are injected by mail in voting, for example.

So, they don’t entirely eliminate the possibility of fraud, just reduce it somewhat.

When I got the call today asking me if I would commit to absentee ballots or vote-by-mail—by Republicans!—I was slightly annoyed; but I finally connected it to something—the NSA’s ThinThread program from the 1990’s.

Scratching your head at this point? Read on.

What was ThinThread? You can learn about in the documentary A Good American, which is the story about Bill Binney at the NSA (he’s a personal hero of mine.) Some of you who have read my posts for a while know that I used to design and build supercomputers, and in the 1990s I installed the first wave of our company’s hardware in the data centers at the NSA myself.

While I didn’t know about Bill Binney at the time, there is a small chance that I may have stood next to him in line once or twice when I was escorted to the cafeteria for lunch while working at the NSA. Makes me happy that I might have been near a legend and hero.

Binney designed ThinThread, and he makes an interesting point in the documentary: with increasingly sophisticated encryption over the years, it became harder and harder for the NSA (who are known as master code-breakers) to be able to decode every new encrypted message in a timely enough manner for it to matter in the real world.

What Binney realized was that the NSA didn’t need to decrypt the messages in order to deduce what operations might be planned or taking place; one only needed to look at the metadata (who was messaging whom, from where, at what times of day and how often—and then thereafter who, when and how often did the next layer of people message, etc.)

Without even knowing the content of the encrypted messages, they could learn a lot simply by mapping the message traffic flows in the command-and-control networks of our adversaries. Binney made the infuriating revelation that 9/11 might have been preventable, had the leadership of the NSA under the kakistocrat Michael Hayden not terminated his program prematurely.

By the way, if you haven’t figured out why Signal, WhatsApp, and other ‘end to end encryption’ communication tools (as well as VPNs) are being pushed “to keep your messages safe from prying eyes” —well, now you know that the metadata is perhaps just as valuable. That “VPN” is maybe not as useful as you are made to think it is…

In the same way as ThinThread was used to analyze intent from metadata alone, when people “vote early” by absentee ballot or by mail, it is almost not necessary to open the actual ballot and read the votes; if you have simply have a record that I “voted early”, and you have a database that shows my likely political preferences, all you need to know is that my ballot has been received—thus you have the metadata that I voted.

You can make a good guess at how I voted long before election day.

In fact, you don’t even need to know that it was my ballot; if all you knew is that 65% of the ballots in some precinct have already been mailed in (but remain unopened), you could make good predictions about the likely outcome in that precinct; you can then make some well-informed guesses and even pinpoint places with low turnout so that you can focus on “turn out the vote” efforts where it is more needed.

The Democrats have mastered the art of ‘ballot harvesting’ for a while now, and we have seen enough to conclude that besides letting them analyze the metadata to figure out where to focus turnout efforts (a legitimate use of this information, although again I don’t like this idea of early voting), they also use the information to figure out who hasn’t voted, so that those (unused) ballots can then be hijacked and injected into the system as “counterfeit votes” by fraudulent means.

So, while I don’t like this idea of vote-by-mail—if it is going to be used again (which it is) then I have to grudgingly admit that the Republicans need to be as good as analyzing the metadata as Democrats are, while staying on the correct side of the law and morality and not manufacturing fake votes, as Dems would do.

Our system of voting system is in drastic need of overhaul, but while it exists in the vulnerable state that it is, we have to be as good as our opponents at leveraging what metadata we can.

---

Thank you to all of my loyal subscribers, and as always—if you’re not a paying subscriber but you can spare $5 a month, your support is greatly appreciated.

---

CognitiveCarbon’s Content is a reader-supported publication. To support my writing and research work, please consider becoming a paid subscriber: at just $5 per month, it helps me support a family.

Freedom isn’t free, and I need to work harder to pay other people’s student loans.

You can also buy me a coffee here. Thank you for reading!