Saturday, November 21, 2020

Tracking down the Dominion team

 As you may know, Dominion is suddenly shutting down its offices and its employees and contractors are desperately trying to wipe any sign of their connection to the company from their online profiles. This archived tweet thread explains why. 

I was researching companies to see how many IT people they employ in US vs overseas, and I found a feature which lets me search LinkedIn by region, so I thought I would apply the same to DVS and see if we can contact any developers in Serbia to get them to flip on DVS. 

I found about 10 people working in Serbia for DVS, and I started to look into the developers, I started to notice that the developers were linked to the company, but more than a few did NOT have the DVS experience listed on their profile 

How does that happen? It happens when you used to have the experience linked but it's since been removed. LinkedIn Caches those employee employer relationships so you stay listed under their company even though you've removed the experience. 

That got me wondering WHY would someone want to remove their relationship with the company... only answer is to cover something up. So I started to research each one.

Here is the screenshot of the DVS Employees on LinkedIn: 

You can see on this screen shot that I am hovering over Aleksandar Lazarevic and that the link to his profile is showing in the bottom left of the screen. The url being: (… )

Note the handle ending in 777, that will come into play later.

In researching this developer I discover this person wrote the book on how to do algorithmic boosting of datasets over time. The Official name for it is SMOTE (Synthetic Minority Over-sampling Technique), I've attached the paper to this email and it's still currently available online here.

Read the whole thing before Twitter deletes it. She appears to have identified the specific individual responsible for writing the code that enabled Dominion to modify the electoral returns of the U.S. presidential election.

Also, Biden didn't win.

Labels: ,


Blogger boogeyman November 21, 2020 7:04 AM  

His name was Seth Rich
Epstein didn't kill himself
JFK only won because of the dead vote
Bill Clinton did actually sleep with that woman
Cuomo killed thousands of old people with the coof
US intelligence agencies spy on the American people
Journalists are really just propagandists
Going to university makes you less able to think clearly

And Biden did not win.

Blogger Rek. November 21, 2020 7:10 AM  

It is time for these devils to be thrown into a lake of fire. I mean when you call yourself Gucifer you are absolutely begging for it.

Blogger vinifera November 21, 2020 7:22 AM  

I believe everyone working for Dominion is a CIA / intelligence operative, knowingly.
Why does Gina Haspel still have a job?

Blogger Gregory the Tall November 21, 2020 7:34 AM  

The link already disappeared.
The same guy seems to know the remedy against voting machine fraud: "Applications of Neural Networks in Network Intrusion Detection"

Blogger Matthew T November 21, 2020 7:38 AM  

@Slippin Jimmy - agreed, it's a shame that the term "Dominion" carries such a negative connotation. Not too many people will be aware that the Dominion of Canada was so named in honor of the Canadian founders' deep Christian faith, by referencing Psalm 72:

He Shall have Dominion From Sea to Sea

Blogger Bernard Korzeniewicz November 21, 2020 7:42 AM  

Because the President wants to save the USA, not a burned out husk of a former union of countries.
You hit to kill or the enemies will hit back.

Blogger Difranco November 21, 2020 7:44 AM  

For posterity

Blogger Difranco November 21, 2020 7:45 AM

For posterity

Blogger Mr.MantraMan November 21, 2020 7:52 AM  

The author of those posts says in her public twitter profile "Airforce Veteran" with American flag emoji, maybe a Flynn loyalist. Also DOE shows up again with this Kegelmeyer fella.

So if Dominion/Smartmatic are "CIA" then the Obama admin and it's flunky Brennan installed Maduro and who will walk in and buy up that broken country's assets at fire sale prices, Soros & CO., shades of the Soviet bust up. Repeat this process till all the world is "healed" and damned if us untermensch sexists, racists and homophobes don't need a "healin."

Blogger pyrrhus November 21, 2020 7:57 AM  

O/T Covid tests are far too sensitive, 90% of positives carry hardly any virus

Blogger dienw November 21, 2020 7:58 AM  

The Tweet mentions the Lame Cherry blog which I read in tandem with Vox's every morning religiously.
Lame Cherry continues the Dominion reveal in today's blog posting

Blogger Critias November 21, 2020 8:00 AM  

If there is tech for manipulating any election, what about referendums? A number of those have happened over the last decade that have advanced globalhomo and caused constitutions to be amended.

Blogger Gregory the Tall November 21, 2020 8:13 AM  

SMOTE (Synthetic Minority Over-sampling Technique) =
Affirmative Action for Votes

Blogger phil g November 21, 2020 8:36 AM  

Great list. I'd add: the recommended American diet will make you sick, fat and low T.

Blogger Joeplanet November 21, 2020 8:46 AM  

The language: dominion, smote.
These people cannot hide the symbolism.

And it will be their downfall.

Blogger Silent Draco November 21, 2020 9:15 AM  

The paper is a giveaway. Look at the graphs toward the end, and note how after 4-5 iterations the precision and F-value make large jumps upward. Now, what if this is not done on training data, but instead on the entire data set of record? What if the minority population isn't 5%, but maybe 30-40% (call it B), and the algorithm retrains the real data set to boost B from minority to a slight majority of 5-8%, as a desired end state. What sort of large steps and changes would you expect to see? I'm competent with signal processing and pattern recognition, not high expert, but I see a large rat emerging.

Blogger Seeingsights November 21, 2020 9:19 AM  

I read the Wikipedia article on Oversampling and undersampling in data analysis. It has a section about SMOTE.

Quote from the article: "Both oversampling and undersampling involve introducing a bias to select more samples from one class than from another, to compensate for an imbalance that is either already present in the data, or likely to develop if a purely random sample were taken."

Doesn't that sound fishy? And if that is an accepted technique
in data analysis, I find that disturbing.

Blogger AJ November 21, 2020 10:05 AM  

@17 - No, it's not fishy.
You have to adjust minority classes of data because that's the real world and ML algos can't be trained otherwise (Well, right now it can't be).
In the real world, cancer cells are small concentrated minority of the tissue sample or MRI scan. If you don't artificially boost this minority class in order to train the classifier, the ML algo will never learn. But that's not what you do in the actual dataset that you're trying to predict/classify - there you leave it as is.
It's actually just like us - we concentrate on niche cases i.e the exception to the rules to figure out the rules for the larger system.

Blogger AJ November 21, 2020 10:07 AM  

@16 - Actually, this line of thinking coheres with what is prevalent in the vote tally dataset. It pieces together a few weird things and I think this is the way the fraud was perpetrated.

Blogger Zastavnik Džemo November 21, 2020 10:09 AM  

It is all true.
I know some of developers personally.

Blogger Fishslinger_Bear November 21, 2020 10:20 AM  

It is here too:

Blogger Ominous Cowherd November 21, 2020 10:54 AM  

Seeingsights wrote:Doesn't that sound fishy? And if that is an accepted technique in data analysis, I find that disturbing.
If you are trying to estimate rockfish populations, you have to oversample the areas containing rockfish. That is radically different from actually counting 100% of votes.

Blogger Stg58/Animal Mother November 21, 2020 11:10 AM  

These people will all leave other nuggets around.

Speaking at conferences, contributing to white can bet they all have huge trails to follow.

Blogger Crunchy Cachalot November 21, 2020 11:25 AM  

Both oversampling and undersampling involve introducing a bias to select more samples from one class than from another

Why would voting, especially in our purported "one man, one vote" system, require any sampling whatsoever? We're not building a statistical model, we're supposed to be counting actual votes, which can be easily done with a pencil, a sheet of paper, and a pair of eyes (or even just one eye in a pinch), no magic "algo" required.

Blogger Ominous Cowherd November 21, 2020 11:34 AM  

Crunchy Cachalot wrote:Why would voting, especially in our purported "one man, one vote" system, require any sampling whatsoever?
It doesn't. Sampling is the antithesis of counting.

Blogger AJ November 21, 2020 11:45 AM  

That’s answered by what the target is. In this case the target could be predicting how many votes need to be produced for a candidate at the end
What could have happened is that the turnout for Trump was quite high and therefore the algorithm was not tracking. So they had to suspend counting in the 5 states and produce ballots pronto - ergo the single candidate marked ballota

Blogger Reprehensible Adam November 21, 2020 11:45 AM  

All this information is great but we need some action soon. DOJ should assemble an inquisition now. Start getting them one at a time and work till there is no one left to arrest. This dealing with the enemy needs to stop here and now.

Blogger Dos Voltz November 21, 2020 11:53 AM  

Haha, "Lazaravic."

Same name as the arch villain in video game Uncharted 2.

Blogger SerbBear November 21, 2020 12:08 PM  

Did some research on the Serbs involved with Dominion. Most of the patents have been filed by a Goran Obradovic in 2012, along with John Poulos, James Hoover, Nick Ikonomakis.

This line says it all and is in all the patents that have Dominion as the assignee:

"The ballot processing application may process the digital image to define the voting results based on criteria established by election officials, including ambiguous mark criteria. The audit trail enables election officials to verify that particular paper ballots have been processed correctly in accordance with these criteria."

Blogger Terran Ghost November 21, 2020 12:23 PM  

@10 "that's not a bug, that's a feature!"

With all the low-info corona cultists out there, that sounds like an excuse to pump up number of cases to justify more totalitarian gov't BS. Most of my co-workers in my department make smalltalk about how "the numbers are so bad!" and how people need to put public safety first. Dumb normie advocacy for giving up rights and civil liberties because of herd mentality.

After all, it's just a mask! Is it that much to ask people NOT travel to visit family? /s

Their inattention to what's going on is exactly why a high-sensitivity test is used.

Blogger Dan B November 21, 2020 12:35 PM  

Having gone to uni, I can definitely confirm it middle the mind.

Blogger Zastavnik Džemo November 21, 2020 1:40 PM  

This people will probably not be prosecuted. Lazarevic was born in Serbia but he is a USA citizen. Probable CIA (Temple university).

He joined Dominion 13 years after the primary software was written, so he was there to do "adjustment" work.

I have personally had 2 encounters with Serbian-Americans rediscovering their love for homeland and coming here to work for 1/5 USA salary in the serbian branch. They were so obviously spooks it was painful to listen to their "background" stories.

The best thing you can learn from all this is that if you need an extremely good but cheap coder, a Serbian with ETF diploma is a sure bet.

Blogger Jack Amok November 21, 2020 2:21 PM  

No, it's not fishy.
You have to adjust minority classes of data because that's the real world and ML algos can't be trained otherwise

It's still fishy. Oversampling isn't fishy for training models, but synthetic oversampling sure is. Why generate fake data points to artificially boost a population when you have the far more simple method of restricting the training dataset to boost the interesting data points? Your models can still be trained to detect the unusual events, but with the benefit of learning to detect real events and not fabricated ones.

That entire paper is the sort of thing that drives me nuts about modern "scholarship." There's a much simpler option available, but you can't publish a paper on doing that, so an over-complicated solution get's cooked up.

Blogger Gregory the Tall November 21, 2020 8:04 PM  

Actually not OT
Just another example of smote

Blogger Jamin990 November 21, 2020 11:58 PM  

The author of the twitter feed believes Lazarevic is Gucifer 2.0. Theres also talk by someone rlse that the photo is artificially genersted because, amongst other reasons, the eyes are at different levels. Everyone should read the full twitter conversation. Its fascinating.

Blogger LoArmistead November 22, 2020 12:37 PM  

I want to believe this, but one thing that doesn't jive in my mind is how a software vote can be changed (plausible) without the paper ballot audit trail not exposing it. For instance if Dominion manipulated Georgia votes via software, would a hand count of all paper ballots that voters verified before submitting not reveal the discrepancy?

Blogger AJ November 23, 2020 9:17 AM  

@36 Depends where you're looking for the evidence. I'm looking at the vote tallies from GA. The following are the votes subtracted from each of the candidates from the NYT feed

Trump : Removed 40,075 votes out of 2,456,727 (1.6%)
Biden : Removed 28,990 votes out of 2,471,707 (1.1%)
Third parties : Removed 99,600 votes out of 64,914 (153% - wut?)

The third party numbers make no sense. So, is anybody checking the third party votes - why are so many votes removed from third parties?

In any case, if votes were switched to Biden using the tabulation software - there needs to be paper ballots behind them. Because, a recount would show the difference. You can be sure that fraudsters would plan for that. So, if I wanted actual hard evidence of fraud, I would try something like this.

As the election night counting unfolds, the need to switch more votes to Biden above and beyond becomes apparent in the 5 cities highlighted. If the appropriate warrants have been executed, those comms are being watched. So the software starts switching votes to Biden with the obvious deficit in paper ballots which need to be printed and shuttled over to the counting centers. Plenty of anecdotal evidence that this happened in the wee hours. So, if I were DOJ/whatever other agency - I would intercept this shipment (after all in the wee hours, what's the traffic going to be like heading to the counting centers?), and take those operators into custody, replace the ballots printed with something quite silly - like Biden only votes on paper that is of different quality (affidavits point this out as well). So, you have the evidence, the bad actors and the software in question. If I were prepared for stopping the steal, then the action would be something like this that connects all the dots and the reported evidence.

Blogger testingthisout November 30, 2020 3:07 PM  

@20 Zastavnik Džemo - Did you work with Aleksandar?

Post a Comment

Rules of the blog

<< Home

Newer Posts Older Posts