iiinova — where your digital self keeps evolving

Alex Baker @aabaker.bsky.social · 08/10/2026

From ASU's charter. Vance can kick rocks. Or eat them, if he so prefers.

Screenshot of ASU's charter statement, which reads "ASU is a comprehensive public research university, measured not by whom it excludes, but by whom it includes and how they succeed; advancing research and discovery of public value; and assuming fundamental responsibility for the economic, social, cultural and overall health of the communities it serves."

010

Alex Baker @aabaker.bsky.social · 13/08/2026

We'll get an indirect proxy for how thatd have gone in a few weeks.

Screenshot of a news article talking about how Nick Sirianni "discovered ChatGPT" in the 2026 off-season.

110

Alex Baker @aabaker.bsky.social · 17/07/2026

He apparently had quite the reputation.

Rabbi Yirmeya raised a dilemma before Rabbi Zeira: If the bird is big and contains such a large amount of blood that it effaces the water, rendering it indistinguishable, or if the bird is small and contains so little blood that its blood is effaced due to the water and indistinguishable, what is the halakha?

Rabbi Zeira said to him: Haven’t I told you not to take yourself out of the bounds of the practical halakha? Do not ask questions about impossible eventualities. The Sages measured the ratio of blood to water specifically with regard to a sparrow. There is no sparrow big enough to efface the water, nor is there one small enough to be effaced due to the water.

2443

Alex Baker @aabaker.bsky.social · 06/07/2026

Reminds me of how back in 2018 the Arizona State House voted 56-3 to expel habitual sex pest Don Shooter. The two Republicans who opposed the measure - alongside Shooter himself - were David Stringer and Noel Campbell. Their wiki articles got additions in the following years:

A screenshot of Campbell's wiki article that reads "n December 2020, Campbell’s wife accused him of domestic violence according to a police report taken by the City of Prescott Police Department. According to Campbell’s wife, the former state representative pushed her multiple times until she fell over and then “struck her multiple times in the neck and multiple times in the face.” Campbell’s wife did not press charges as she believed Campbell’s violence was a result of early stages of dementia and that they would be seeking medical treatment."

A screenshot of Stringer's wiki article that reads "On January 25, 2019, the Phoenix New Times reported that Stringer had been charged in 1983 with five sex offenses, including two child pornography charges, while he was living in Baltimore. He had accepted a plea bargain on charges of paying two boys for sex, one of whom was developmentally disabled, and was sentenced to five years' probation.[16][17] In 1990, he had the charges expunged from his record. In response to calls for his resignation, Stringer confirmed that he had been arrested and charged with the offenses but stated that he would not be resigning from the Arizona House.[18] On March 27, 2019, facing a deadline to hand over documents regarding his 2018 comments and the 1983 sex offenses, he abruptly resigned from the Arizona House."

010

Alex Baker @aabaker.bsky.social · 23/06/2026

Munsil's best known as the sacrificial lamb to the all-conquering tide that was Gov Napolitano's reelection campaign in '06. He ran on kitchen-table issues like griping that the 9/11 Memorial included Balbir Singh Sodhi, and that Piestewa Peak got renamed from its previous racial+gendered slur name.

The wikibox from the 2006 Arizona gubernatorial election showing Republican Len Munsil somehow managing to lose every single county to incumbent Democratic Governor Janet Napolitano.

010

Alex Baker @aabaker.bsky.social · 12/06/2026

Heck, since I'm bored: Do you believe this excerpt from their PA lawsuit to be true? If yes, what kinds of examples of that occuring in other (non-Russian/Ugandan) elections would you be interested in, and how many? I've got old races, foreign ones, Dem. primaries, nonbinding referenda, the works.

Excerpt from ETA's PA lawsuit that reads "90. In 2012, Dr. Klimek, Yegorov, Hanel, and Thurner published research
demonstrating that fraudulent elections exhibit distinctive statistical patterns. In fair or normal
elections, vote share for winning candidates remains stable across different turnout levels.
Fraudulent elections show unusual clustering at high turnout/high vote percentages.
91. Instead of tight clustering around typical values, fraudulent elections show data
spreading or smearing toward the upper-right corner of vote-turnout plots."

120

Alex Baker @aabaker.bsky.social · 12/06/2026

3. Claimed to have "controlled for demographics" despite the 6 variable combo they used generating far lower (5-10x) partial r^2s for SigC in my run, and the 100+ variable set I threw at it generating even smaller partial r^2 (100-200x lower than claimed).

Partial r^2 test on GOPReg totals (FraudSigC) using just the 7 demographic controls ETA claimed to use, already only 1/5th of their magnitude.

Partial r^2 results for SigC after applying a full demographic battery at the county level. Further drops can be induced by substituting precinct-level stats.

110

Alex Baker @aabaker.bsky.social · 12/06/2026

1. Wrote everything they did about FraudSigA without disclosing the only reason 1/GOPReg has such a big coeff. is an outlier precinct (Edgewater 1.5 in Kenton) where there are 7 registered GOP voters and 1 voter turned out (for Masse). Had that voter voted Gallrein the signature would have flipped.

Basic graph of 1/GOPReg in the KY-CD4 race with Massie winning Edgewater 1.5 (upper-right corner) 1-0.

Same graph as before, just now Massie has lost Edgewater 1.5 (bottom right) 1-0 instead of winning it 1-0.

120

Alex Baker @aabaker.bsky.social · 07/06/2026

Really loved this cutscene from of Age of Mythology:

A photoshopped image of Nithya Raman's head over the physical embodiment of the Egyptian god Osiris from the game Age of Mythology, seen here saying "I have come upon Sunset Boulevard and with my own two feet taken possession", a slight modification of the in-game "I have come upon the earth and with my own two feet taken possession" (derived purportedly from the Book of the Dead), spoken by Osiris right before he starts flinging meteors from the game's bad guys from the sky.

1456

Alex Baker @aabaker.bsky.social · 07/06/2026

The State Treasurer race is also functionally indistinguishable from the Pres. one re. distribution shape. The same scary graph shapes ETA loves flagging can be derived from either one. If such distributions were impossible under normal US voting behavior we'd have heard of PA RLA discrepancies.

Side-by-side KDE heatmaps of the correlation between turnout and GOP vote share in the 2024 PA Presidential and State Treasurer races, both showing nigh-identical shapes, and if anything the "anomalous" correlation is stronger in the Treasurer race.

Side-by-side scatterplots of the correlation between total precinct votes and GOP vote share in the 2024 PA Presidential and State Treasurer races, both showing nigh-identical shapes, and if anything the "anomalous" correlation is stronger in the Treasurer race.

Scatterplot of GOP Presidential vote share vs. GOP Treasurer vote share in PA in 2024, showing a >0.985 r^2 correlation.

140

Alex Baker @aabaker.bsky.social · 27/05/2026

Were this method actually worthwhile to use in detecting potentially fraudulent anomalies, it'd flag results out of places like Nebraska, Kansas, Oregon, and Minnesota as all visibly benefitting Harris.

Graph showing Harris vote share in Nebraska in 2024 increasing as precinct size does

Graph showing Harris vote share in Minnesota in 2024 increasing as precinct size does

Graph showing Harris vote share in Oregon in 2024 increasing as precinct size does

Graph showing Harris vote share in Kansas in 2024 increasing as precinct size does

110

Alex Baker @aabaker.bsky.social · 27/05/2026

If you ignore municipal boundaries (which are another limitation on just which precincts get drawn where) and only limit yourself on contiguity you can force the r^2 to grow obscenely high. Non-contiguous configurations make the r^2's even larger. These all result in the same countywide vote totals.

Simulation of a theoretically-possible combination of Allegheny County precincts ignoring municipal boundaries that leaves a strongly positive correlation between EDay vote totals and Harris.

Simulation of a theoretically-possible combination of Allegheny County precincts ignoring municipal boundaries that leaves a strongly negative correlation between EDay vote totals and Harris.

110

Alex Baker @aabaker.bsky.social · 27/05/2026

If you toy around with which precincts you combine together, you can instead force Pittsburgh to have mega precincts on the right side of the graph, despite changing 0 votes in the process. And the existing negative Harris/positive Trump correlation is only ~1/3rd as strong as it could be.

Simulation of a theoretically-possible combination of Allegheny County precincts that leaves a strongly positive correlation between EDay vote totals and Harris.

Simulation of a theoretically-possible combination of Allegheny County precincts that leaves a strongly negative correlation between EDay vote totals and Harris.

110

Alex Baker @aabaker.bsky.social · 27/05/2026

Time and again in this current political distribution of coalitions you're going to see smaller precincts in Allegheny County go blue and bigger ones go red. It's even visible when you search for splits in primary participation by precinct. This points to something structural from year-to-year.

Split of partisan primary preference by precinct in Allegheny County, 2022

210

Alex Baker @aabaker.bsky.social · 27/05/2026

Yeah, I've got some stuff about Allegheny County and can go a bit more indepth about why the GOP precincts are consistently bigger. First off, the plot of registration total vs. GOPPct, and some races from the 90s showing correlation between votes cast and GOPPct even then.

Graph of precinct count of registrants in Allegheny County vs. what % of them are registered GOP

Comparison between Ballots Cast at an Allegheny County polling place in 1992 and GOP% for the Auditor General race

Comparison between Ballots Cast at an Allegheny County polling place in 1996 and GOP% for the Treasurer race

Comparison between Ballots Cast at an Allegheny County polling place in 1996 and GOP/Dem% for the Presidential race

110

Alex Baker @aabaker.bsky.social · 25/05/2026

Bit further eastwards but Roosevelt and Canal is an interesting corner, to say the least.

Google Maps walking directions from the closest Dollar Tree store to me to the closest Nordstrom store.

010

Alex Baker @aabaker.bsky.social · 22/05/2026

Sorry, since this is the 3rd reply I've gotten in this vein I fear I might be being misread: I want to be 1000% clear that I concur, that there is no irony in my first comment, and that I hope Galindo goes the way of Duke. Just felt like drawing a throughline to the last example in a Dem. primary.

A screenshot of my 3 most recent donations through ActBlue, including 2x$25 to Johnny Garcia in Texas.

240

Alex Baker @aabaker.bsky.social · 16/05/2026

The funny/bizarre/depressing/unsettling/your adjective of choice thing is that apparently he very much was:

Op-ed coauthored by Platner in the wake of Nord Ost telling Americans to "always ask why" terrorists do terrorism because "one man's terrorist is another man's freedom fighter".

Coverage of Platner getting kicked out of a Bush-Collins rally for protesting Iraq in Oct. 2002

Old Platner yearbook photo where he's holding a poster calling for the liberation of Kurdistan, Palestine, Chechnya, Tibet, Kashmir, and Kosovo

Excerpt from a Platner 2011 Reddit post talking about how he was politically active in HS.

110

Alex Baker @aabaker.bsky.social · 16/05/2026

He said as much in a 2011 Reddit post, coincidentally. Explicitly asked for infantry and for Iraq despite believing it to be the less-justifiable of the two wars. www.reddit.com/r/AskReddit/...

Platner comment from Reddit in 2011 about how he explicitly asked for the infantry and Iraq upon enlisting in the USMC.

1201

Alex Baker @aabaker.bsky.social · 07/05/2026

This image of 2-18 participating in it during the 2010 parade now probably looks about as bizarre as a unicorn grazing in Central Park.

Members of the US 2nd Battalion, 18th Infantry Regiment during dress rehearsals for the 2010 Moscow Victory Day Parade

021

Alex Baker @aabaker.bsky.social · 06/05/2026

Eh, you can find those strings upon searching: www.justice.gov/epstein/file...

Epstein email from late 2016 where he uses the phrases "watcha", "bust out", and "cryin"

0100

Alex Baker @aabaker.bsky.social · 03/05/2026

Their two tests can also send out conflating signals. In Hendricks County IN for example Harris has positive TotalVote correlation, but Trump has a positive turnout correlation. The same has been taking place across elections in MN since 2016 (though Dem correlations always stronger than GOP).

Timeline showing strength of TotalVotes->Dem and Turnout->GOP correlations in Minnesota since 2008. A noticeable uptick for both metrics across races (President/Senator/Governor) can be seen beginning in 2016, but the strength of Dem correlations (which would be deemed anomalous by ETA) consistently outstrips GOP turnout correlations. In 2018 all 3 statewide races show near-0 correlation between GOP share and turnout.

Graph showing Harris vote share in Hendricks County, IN increasing as precinct total votes cast do in 2024

Graph showing Trump vote share in Hendricks County, IN increasing as precinct turnout does in 2024

120

Alex Baker @aabaker.bsky.social · 03/05/2026

And now turnout: * Favoring the Hungarian opposition in 2022 vs. Orban * Favoring Le Pen in France in 2022 * Favoring the LibDems and Tories (and hurting Labour) in the UK in 2024 * Hurting NZ Labour in 2023

Graph showing Hungarian opposition vote share in 2022 increasing as precinct turnout does

Graph showing Marine LePen vote share in 2022 increasing as precinct turnout does

Graph showing UK Conservative/LibDem vote share in 2022 increasing as precinct turnout does while Labour vote share decreases

Graph showing NZ Labour vote share in 2023 decreasing as precinct turnout increases

220

Alex Baker @aabaker.bsky.social · 03/05/2026

The trends are visible in international elections as well. First a small sample of TotalVotesCast correlations: * Against the AFD in Saxony 2025 * Favoring the SPD in Saxony-Anhalt 2025 * Favoring the Canadian Tories in Labrador 2025 * Favoring the Canadian Greens in Outremont 2025

Graph showing AFD vote share in Saxony 2025 decreasing as total precinct votes cast increases

Graph showing SPD vote share in Saxony-Anhalt 2025 increasing as total precinct votes cast increases

Graph showing Tory vote share in Labrador riding 2025 increasing as total precinct votes cast increases

Graph showing Green vote share in Outremont riding 2025 increasing as total precinct votes cast increases

110

Alex Baker @aabaker.bsky.social · 03/05/2026

Their last one swaps total votes cast for turnout. By ETA's logic the following races show pro-Dem anomalies: * 2024 Hawaii for Harris * 2024 San Francisco for Harris * 2024 King County WA (Seattle) for Harris * 2024 Dearborn MI for Harris

Graph showing Harris vote share in Hawaii in 2024 increasing as precinct turnout does

Graph showing Harris vote share in San Francisco in 2024 increasing as precinct turnout does

Graph showing Harris vote share in Dearborn MI in 2024 increasing as precinct turnout does

Graph showing Harris vote share in King County WA in 2024 increasing as precinct turnout does

110

Alex Baker @aabaker.bsky.social · 03/05/2026

Approach #2 is measuring whether or not there's a positive correlation between total votes cast per precinct and candidate vote share. By ETA's logic, the following races show pro-Dem anomalies: * 2024 Nebraska for Harris * 2024 Minnesota for Harris * 2024 Oregon for Harris * 2024 Kansas for Harris

Graph showing Harris vote totals increase as Nebraska precinct size increases

110

Alex Baker @aabaker.bsky.social · 03/05/2026

But that's just not true. 4 of 7 don't have auto recounts at all. In the others the margins easly cleared thresholds. AZ: R+5.53 vs 0.5% threshold GA: R+2.19, no auto recounts MI: 80.1k vote margin vs. 2k vote threshold NV: R+3.1, no auto NC: R+3.21, no auto PA: R+1.71 vs. 0.5% WI: R+0.86, no auto

A table of recount thresholds by state:

AZ: Auto under 0.5%, no manual requests
GA: No auto recounts, manual under 0.5%
MI: Auto under 2K vote margin, no restrictions on manual
NV: No auto recounts, no restrictions on manual
NC: No auto recounts, manual subject to varying limits - statewide race threshold is 0.5%
PA: Auto recounts under 0.5%, no threshold restrictions on manual
WI: No auto recount, manual available under 1%

350

Alex Baker @aabaker.bsky.social · 03/05/2026

Non-zero, though he switches it up with "from the standpoint of". I think both get used semi-often in the JAJ impersonation on SNL.

Various web links of Trump comments where he says the phrase "as it pertains to"

A screenshot of Trump at a Jan 3rd 2026 presser about how the US will have a presence in Venezuela as it pertains to oil.

030

Alex Baker @aabaker.bsky.social · 02/05/2026

43 years as an airline, 34 of which (rounding up) were explicitly as Spirit.

A screenshot of Spirit Airlines Wikipedia page, showing when it first became an airline under a different name 43 years ago, and became Spirit ~34 years ago.

040

Alex Baker @aabaker.bsky.social · 02/05/2026

Theyre also undermined by the fact that half of what they have is diametrically opposite from the other half. They have tried and will (thankfully) stupidly continue to try to brand him both as Mr. Totenkopf and the leftie peacenik who wrote we should seek to understand why terrorists do terrorism.

Laura Loomer simultaneously calling Platner a Nazi and a Communist, because words mean nothing in the modern era.

An NRSC press release from Sep. 2025 highlighting a HS-era op-ed Platner wrote in the aftermath of 9/11 about how we should "always ask why" terrorists do terrorism.

An NRSC press release from several months later about how Platner is, like, totally a Nazi.

020

Alex Baker @aabaker.bsky.social · 27/04/2026

If this is how she responded to the word "vehicle" I presume something like "motor pool" will leave her fascinated and fixated for days.

Far too much pablum from Bari Weiss's sister for me to bother transcribing on my phone, but it ends with her writing "I lost count of the number of times I heard about 'vehicles' coming and going, the security-minded man's term of endearment for cars".

020

Alex Baker @aabaker.bsky.social · 19/04/2026

Graphs like those or these (in Twin Cities Dem. strongholds the turnout correl. is 3x as strong for Harris as statewide Trump turnout correl. is, incl. precincts in the 90%+ vote share and turnout corner) would absolutely have been flagged by ETA were the colors flipped. They go unmentioned in MN.

020

Alex Baker @aabaker.bsky.social · 19/04/2026

This is what that correlation looks like for MN'24. A correlation 2.5x stronger than Turnout->GOP, that spiked far more in '16, that has persisted even as turnout has fluctuated, and which also coincidentally is by far the weakest (1/10th as big as statewide Harris r^2) in hand-counted precincts.

Bucketed average of precinct size vs. vote share correlation in MN2024-Pres

Rolling average of precinct size vs. vote share correlation in MN2024-Pres

Year-over-year dashboard comparing r^2 strength of turnout->GOP and TotalPrecinctVotes->Dem correlations

Correlation between Total Precinct Votes and Candidate Vote Share in 2024 MN handcounted precincts.

110

Alex Baker @aabaker.bsky.social · 19/04/2026

And there's a test of theirs missing from the MN article. Do you recall this graph from their PA release? They wrote "...do we see a relationship between the number of votes cast at a precinct and how well a candidate performs in that precinct? If so, does that trend benefit only one candidate?"

Screenshot of ETA-generated graphs of San Mateo County CA's 2024 presidential results showing a flat line when plotting turnout vs. candidate vote share correlations, (incorrectly) claiming that this is what constitutes a "normal" election/deviations from it are anomalous.

110

Alex Baker @aabaker.bsky.social · 19/04/2026

When running on precinct-level demog. data you can get a host of combos where both a) partial r^2 for turnout is dropped to functionally zero and b) the slope of the turnout correlation turns flat/negative. I personally only think A is important but ETA uses both. Breakdowns in image alt-texts.

Partial r^2 for correlation between turnout vs. TrumpPct in Minnesota in 2024 after controlling for:

* Precinct White%
* Precinct Black%
* Percent of precinct under age 18

(Black% and White% have some collinearity - though a manageable amount - but introducing a second race field helps differentiate between various nonWhite areas. The indigenous community at Little Earth and on the Native reservations, the Somali community in Cedar-Riverside, non-immigrant Black Americans in Brooklyn Center, Hmong/other SE Asian immigrants to Frogtown, Hispanic voters in Powderhorn all have their own political behavioral variations)

(All values are from 2024 ACS)

Partial r^2 for correlation between turnout vs. TrumpPct in Minnesota in 2024 after controlling for:

* Precinct White%
* Precinct Male%
* % of voters in precinct who registered the day of the 2024 general election

Partial r^2 for correlation between turnout vs. TrumpPct in Minnesota in 2024 after controlling for:

* Precinct White%
* Precinct Black%
* Precinct Native American%

(I promise this is a different graph than the last one)

Partial r^2 for correlation between turnout vs. TrumpPct in Minnesota in 2024 after controlling for:

* Precinct White%
* Precinct Native American%
* Percent of precinct adult men who work in the mining industry
* Percent of voters who registered on same day as election

110

Alex Baker @aabaker.bsky.social · 19/04/2026

Now, I don't actually think Obama rigged his 2000 House primary run. If he did he did an atrocious job of it, because he lost by over 30 points. But the outlier areas in the race (precincts around UChicago/other White neighborhoods) will throw a wrench into models trained on unimodal European data.

130

Alex Baker @aabaker.bsky.social · 19/04/2026

The script I used is github.com/aabaker1997/..., based on an R package I found at github.com/kkalininMI/E.... It finds the rough numbers Klimek reported for Russia'12 and Uganda'11, finds functionally nothing for MN 2024, and finds a lot for Obama's 2000 US House primary against Bobby Rush.

120

Alex Baker @aabaker.bsky.social · 19/04/2026

Between '84 and '24 the main mass of the heatmap moved upwards and the appendage below it became more pronounced. This is the graph result of a) the farming and mining areas of MN getting redder and b) big cities getting bluer rel. to them, both of which have happened across the country.

120

Alex Baker @aabaker.bsky.social · 19/04/2026

I also don't have any clue what the tail they're citing for MN in their heatmaps is. Philly - for all my thoughts on their approach there - at least has the "smaller segment streaking up and to the right" element for Trump's vote share that they've referenced previously. You don't see that in MN.

ETA 2024 MN heatmap of turnout vs. vote share

My heatmap of turnout vs. TrumpPct in Philadelphia in 2024

100

Alex Baker @aabaker.bsky.social · 19/04/2026

Actually hell why not 4 more - Chicago's 1983 Mayoral, the 2020 IL Presidential, the 2017 Alabama special, and the 2012 New York presidential. In all four the peaks at 90%+ would satisfy the RFE definition of the tail. All four correspond to heavily Black, authentically Dem. precincts.

1983 Chicago mayoral results for Harold Washington

130

Alex Baker @aabaker.bsky.social · 19/04/2026

One such curve from a race I think we'd all agree it'd be silly to call rigged (though I guess if one really wanted to be pedantic Poroshenko was the more pro-Europe/anti-Russia candidate then so there was something of a cui bono factor present): Ukraine's presidential runoff elections in 2019.

231

Alex Baker @aabaker.bsky.social · 09/04/2026

They built the Zero-Day Finder from the hit novel "Don't You Dare Ever Build The Zero-Day Finder".

Graph showing exponential increase in successful Firefox JS shell exploitation by Mythos Preview vs. past Anthropic models.

0160

Alex Baker @aabaker.bsky.social · 04/04/2026

2 pics from the Vietnam War

Chicago's flag flies over a US military outpost in Quang Tri

Chicago's flag flies over an American artillery position in Vietnam

000

Alex Baker @aabaker.bsky.social · 04/04/2026

That's "2024 3rd place finisher in the Fountain Hills mayoral election Joe Arpaio" to you!

010

Alex Baker @aabaker.bsky.social · 04/04/2026

Jessica Lynch. One of 3 women captured in the ambush alongside her best friend from training Lori Piestewa (died of wounds in captivity, becoming the first indigenous woman to die fighting for the US) and Shoshana Johnson. You can likely guess why Lynch got the lion's share of the media attention:

Shoshana Johnson at an NAACP dinner, July 18, 2003

Lori Piestewa a few weeks before her death in 2003

Jessica Lynch is awarded the Bronze Star, Prisoner of War, and Purple Heart medals on July 22, 2003

120

Alex Baker @aabaker.bsky.social · 02/04/2026

Better call the French and British then:

Plot of French presidential 2nd round election in 2022, showing LePen's vote share per turnout bucket increasing as turnout increases while Macron's decreases.

Plot of the British parliamentary election in 2024, showing Labour vote share decreasing as turnout increases, while Tory and LibDem vote share increases. Tory share flatlines at around 65% turnout while LibDem just keeps spiking.

110