iiinova — where your digital self keeps evolving

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 17h

New "Do you even lift?"/"Homie what set you from" just dropped.

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 02/10/2026

Interesting associations in this research, including: younger men were more likely to endorse white nationalism, but women showed a U-shaped relationship between WN endorsement and age; "Not sure" and "Did not vote" correlated with white nationalism about as much as conservatism/GOP/Trump support.

000

Reposted by Splainer (Material Conditions Acknowledger)

Quinn Slobodian @quinnslobodian.com · 29/09/2026

Interesting that the ICC is the test case for genuine digital sovereignty.

37620

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 01/10/2026

This post got me to try recalling when Netanyahu (first) did his pointing-at-a-stylized-Iranian-bomb-diagram thing at the UN. My guess was 7 years too late.

Screenshot of the start of September 28, 2012 Reuters article "Netanyahu's Iran cartoon bomb timed to make big impact" (https://www.reuters.com/article/world/netanyahus-iran-cartoon-bomb-timed-to-make-big-impact-idUSBRE88R12L/). Included are the lead image's caption, byline, and first two paragraphs.

> Israeli Prime Minister Benjamin Netanyahu draws a red line on a graphic of a bomb as he addresses the 67th United Nations General Assembly at the U.N. headquarters in New York September 27, 2012. REUTERS/Lucas Jackson Purchase Licensing Rights
>
> By Jeffrey Heller
>
> NEW YORK (Reuters) - The "Bibi bomb" was born of days of discussions between Israeli Prime Minister Benjamin Netanyahu and a brains trust of close advisers on how to make a powerful impact in yet another speech on Iran's nuclear program.
>
> "The diagram made his address special," a senior official in Netanyahu's entourage said on Friday about the cartoonish drawing of a bomb the Israeli leader, who is nicknamed "Bibi," used at the U.N. General Assembly as a prop to illustrate what he sees as Iran's drive for an atomic weapon.

010

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

OK, I'm calling you out for repeatedly amplifying Gaza-denying transphobes on this platform, and for spontaneously defending Neera Tanden's boosting of Holocaust-denying neofascist Nick Fuentes over on X. Feel free to remove yourself from Bluesky anytime.

Screenshot of Will Stancil's March 3 blueskeet

> Given that Tanden was recently going after Platner for antisemitism, I’m pretty sure this is just her pointing to and mocking intra-right infighting and not any kind of endorsement of Fuentes. Its like if I retweeted Trump saying something absurd - I wouldn’t expect anyone to take that as pro-Trump

quoting Hannah (@hannahgais.bsky.social) Gais's blueskeet "lol ok" over a screenshot of Neera Tanden on X/Twitter reposting Nick (@NickJFuentes) Fuentes's then-recent "Still better than Kamala?" xeet.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

How many days ago was it that you had to delete a post trying to dogpile someone who asked how you slept, "son"?

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

What "lot of stuff" is it "pretty good" at explaining? It's a partial explanation of the rise of the radical right over the last 30 years, which I am guessing is the pet example you tacitly have in mind, but even then only a partial explanation.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

Do you have evidence for that latest sweeping claim beyond your single follow-up example of the COVID-19 pandemic deranging unemployment rates in 2020?

020

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

I saw this ABD, he was counting his citations. It's like, dude, you have to _wait_.

110

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

I retract that new graph — looking at my code with fresh eyes, I see I forgot to update a variable name, so used the wrong model object to make that new graph. Fixing the bug and re-generating, here's what I get without the egg regressors. (Dashed line again 2022-02-01. In-sample _R_² is again 70%.)

Line chart of the actual monthly price-blame time series (light blue) and a _corrected_ re-fit of my final model, with the two egg variables removed (pink), to the pre-2020 price-blame series. A vertical dashed line represents January 1, 2022.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 30/09/2026

I think that's broadly fair; I don't think he literally _asked_ in the sense of posing an explicit question, but he has in the past, and I wasn't sure how else to take his "LOOK AT THE GRAPH" spluttering. So I took the spluttering seriously as a request for a solution, and proposed one.

011

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

And Stancil himself was literally demanding that people "LOOK AT THE GRAPH" and saying the "dynamic" represented by the vibecession gap between lines on the graph "will kill progressive policy". Do people need to address the puzzle to save progressive policy? Or are we being baited in bad faith?

130

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

I didn't wake up today expecting to think of the political slogan "We are the 1%", and approve of it, but here I am.

010

Reposted by Splainer (Material Conditions Acknowledger)

Gregk Foley @gregk.co.uk · 27/09/2026

The Internet, Social Media and AI are facilitating the churning out of obscene atrocity denial propaganda to deterritorialised global audiences in a way that feels unlike any prior epoch, and I feels like nobody in power is even willing to confront that because they expect to exploit it themselves.

224464

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

Not obvious at all, actually, not least because even that answer's ambiguous. Do you mean "predictive" in the sense of demonstrating skill on held-out data, in the sense of demonstrating skill on data yet to be observed at the time of model construction, or something else?

120

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

Not an answer to my question. I asked which explanations would clear the bar of achieving actual interest, not for a (very) partial ordering of goodness.

110

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

One could reject pretty much any Your Regression Omits Relevant Regressors-type explanation as "P-hacking", and if one's ruling out that entire class of explanations, in what possible explanations are you/Stancil actually interested?

130

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

I personally don't believe Stancil's obliged to pursue an answer. I do, though, believe that if there's no answer he'd actually accept, he should stop demanding people pay attention to the puzzle while using catastrophic rhetoric to hype up the importance of answering it.

010

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

And "We no longer have access to the mind-state" is tantamount to conceding that Stancil's trolling. If all solutions to a puzzle are delegitimized because contaminated by the puzzle having been formulated, waving the puzzle in people's faces to bait them into offering solutions is a troll action.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

False premise. The standard models would only have "seemed to work perfectly" in 2022 to people not looking closely at the charts. Example here: I doodled green brackets on a chart to mark where a classic inflation-and-unemployment model systematically mis-predicted for years at a time pre-2022.

Screenshot of figure 4A of https://www.nber.org/system/files/working_papers/w32163/w32163.pdf, NBER working paper 32163 "The Cost of Money Is Part of the Cost of Living: New Evidence on the Consumer Sentiment Anomaly" by Bolhuis, Cramer, Schulz, & Summers. It's a line chart of consumer sentiment (in blue) from 1977 onward and "expected consumer sentiment given unemployment and inflation" in red. I've drawn on straight-lined green brackets marking, approximately, 1982 to 1987, 1987 to 1990, 1996 to 2001, and 2007 to 2014, periods within which the blue line was either consistently above or consistently below the red line.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

(I also notice that, one, the Standard Model analogy seems to put me on Team Standard Model, and two, the Standard Model is routinely described as one of the most empirically successful theories in physics, if not in science generally. Just sayin'.)

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

A month ago you described the "trend" "in the anomaly" as "a straight-line monotonic increase" over the course of about 6 years. Taken seriously, that description rules out the "deviation is abrupt" hypothesis.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

You might get a kick out of this. Cutting the egg-price variables out of the model and re-fitting actually shifts the model's expectation upward, _more_ in line with the survey numbers, for 2022 onward.

110

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

Because I _did not know_ whether it's "going to have predictive validity" across the multiverse. To repeat myself without the cute pun, I was and am _agnostic_ about that. Being open to the possibility that it is, I fitted a model accordingly.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

You're not the only one, I saw someone else catch a suspension for a two-week-old post. Why are Bluesky mods still doing this time-delayed kirking shit?!

010

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

Bluntly, I think you're selectively demanding rigor from me here, holding me to double standards. Am I right to guess that you don't have a quartet of posts leveling these critiques against the (often worse) models of people arguing for the reality of a "vibecession"?

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

... And "serial correlation due to common trends" can itself be a source of association.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

...YARDCG, to whom you originally responded, did the same thing (also without causal language) when linking my postletter. My chart and Stancil's further imply that what I'm doing is a curve-fitting exercise, rather than something purporting to establish and quantify causation all by itself. ...

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

My final paragraph describes how what I'm presenting is a regression model with reduced residuals for recent years, which isn't a causal claim. When I linked my postletter, I quoteblueskeeted Stancil's "I made a chart" post about two lines on a chart diverging, including the chart itself. ...

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

And I allude to that statistical, not-in-itself-causal sense by using the phrase "explain most variation" in my title (and repeating the word "variation" in the subhed). My intro explains that what I'm doing is trying to "reconcile" two lines on a chart.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

I don't agree that a "reasonable person would interpret" my use of "explains" "to imply a causal claim". It's quite normal to use "explains" in a statistical rather than causal sense, as with characterization of the _R_² (or η²) as the proportion of "variance explained".

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

What you're glossing as "the model says sentiment should recover" is the model saying that sentiment should be about as bad, or worse, than it was in the Great Recession, and about as bad as it was during the shoulders of the recession(s) circa 1981.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 28/09/2026

I personally expect people to take egg prices into account to some extent. Why? Because they cite them curiously often. And here's what pre-2022 (and post-) look like if I take out both egg regressors. Vertical dashed line is 2022-02-01. (In-sample (pre-2020) _R_² dips a bit to 70%.)

Line chart of the actual monthly price-blame time series (light blue) and a re-fit of my final model with the two egg variables removed (pink) to the pre-2020 price-blame series. A vertical dashed line represents January 1, 2022.

110

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

I don't think that (I am egg-nostic, if you like) and I certainly don't think "basically all the egg price variance is post-2022"!

110

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

How? I fitted the model to pre-2020 data, so how would the model know to rely on egg prices if their variance were negligible pre-2020?

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

I think your speculation has some plausibility, working as it does from the premise that sentiment's affected by exogenous material supply shocks via prices. But that concedes the wrongness of the "vibecession" notion that sentiment's become disconnected from objective measures like prices!

110

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

Eggs, specifically, are central to your wilful misunderstanding. You've twice now asked what my claim is, apparently without even skimming the postletter I wrote, and after explicitly told you what my claim is. You're just trolling.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

I explicitly wrote above that my choice of prices was informed by prior knowledge, so if replacing eggs with wheat or milk made the explanation vanish (and I wouldn't presume that it necessarily would!), I can simply respond, "That's no surprise. You're flying in the face of prior knowledge".

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

Standard deviation of residuals pre-2021, 4.73 points. Standard deviation from 2021-01-01 on, 7.29 points. And all of them at once:

White-on-black line chart of monthly residuals from the model under discussion, January 1978 through August 2026.

000

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

In which case "the core vibecession claim" would seem to be substantially a disguised claim about how regression works. If a variable changes its level in 2022, a regression model accounting for that change is going to have variables that, individually or collectively, changed level in 2022.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

Consumer prices "in general" can in fact "do that", I just documented that they did. "[P]rices writ large" in the sense of the overall CPI can't, and I documented that too. And as for replacing eggs with wheat, I wouldn't know, I haven't run those numbers.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

Yes, and indeed I quantify the cumulative size of the post-2021 residuals, in the process documenting how much of the "vibecession" the change in model addresses.

Screenshot of the "How successful is this explanation?" section of the postletter under discussion.

> How successful is this explanation?
>
> Now I check my eye’s hunch that the eggs-gas-rent-interest model accounts for most of the price-blame vibecession. That vibecession is represented by the gap between Stancil’s orange line, “annualized change in CPI”, and his blue line, price blame. I fit a simple regression model of the latter to the former and add up its residuals (the gap between model and reality, measured in percentage points) over the relevant time period: 1145 point-months if I count from September 2021 (reflecting Stancil’s “consistent 5-year trend” remark), 1159 point-months if I count from the start of 2022 (following Stancil’s reference to 2022).
>
> Summing the residuals of my final model over the same time periods gives 501 point-months (from September 2021), and 474 point-months (from January 2022). Thus most (56%–59%) of the error is gone, and so most of this vibecession is explained by improving the model. That fact is especially meaningful because I deliberately fitted my final model only to the data through 2019, whereas I fitted the simple CPI-only model to all of the data. The model not trained on the 2020s does much better at matching the 2020s.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

Well, yeah, obviously any regression accounting for a variable going up in 2022 and staying up is going to have regressors that also persistently changed (individually or collectively) in 2022. That follows from how regression works.

100

Splainer (Material Conditions Acknowledger) @splainer.bsky.social · 26/09/2026

No, the claim is that prices can account for most temporal variation in how often people in the University of Michigan's consumer-sentiment survey blame higher prices for making them worse off.

100