iiinova — where your digital self keeps evolving

Clint Harris @clintent-to-treat.bsky.social · 07/10/2026

I want to set up a betting platform, a la Polymarket, where you ask questions you want an LLM to answer, and the hype guys take your bet and ask your question as you phrase it or however they can phrase it. If they can get a correct answer they win (money well spent), if they don't you win.

000

Clint Harris @clintent-to-treat.bsky.social · 07/10/2026

This just makes me see Jesus walking toward Ken Paxton with a millstone on a chain, giving off Witch King of Angmar vibes. But scarier, because he is usually nice.

100

Reposted by Clint Harris

Daniel A Collier, PhD @dcollier74.bsky.social · 02/10/2026

My manuscripts are a means to an end - it isn't a labor of love or romanticized shit - the means to the end is to help the people I've done the research on. That's it. Citations are nice and I want them but being as right as I can be and getting shit into decision-makers hands is the goal.

static.klipy.com

Lanterns Hal Jordan

ALT: Lanterns Hal Jordan

032

Clint Harris @clintent-to-treat.bsky.social · 02/10/2026

"Have you tried adding and subtracting the thing causing problems from your equation?" is truly the "Have you tried turning it off and back on again?" of applied econometrics.

010

Reposted by Clint Harris

Sam @samd.bsky.social · 30/09/2026

should be instructive for democrats that Magyar is living up to his promise to prosecute Orban officials for corruption and is continuing to poll extremely highly

231635450

Clint Harris @clintent-to-treat.bsky.social · 30/09/2026

The extremely niche content I want to see is rubix cube races where the contestants are given a set of potential outcomes for each color as a function of adjacent colors, and they are scored on how quickly and how well they solve the assignment problem to maximize welfare.

010

Clint Harris @clintent-to-treat.bsky.social · 26/09/2026

Just saw 2 blad eagles circling above my house. Toyally get it! Those things are majestic!

010

Clint Harris @clintent-to-treat.bsky.social · 25/09/2026

Bayesianism is just Frequentism when our universe is a sample draw from the population of universes; this is not even hard. Send tweet. (Cunningham's Law says I'm about to learn about Bayesianism.)

010

Reposted by Clint Harris

Manasvini Singh @vinisingh.bsky.social · 08/06/2026

New WP, just in time for #ASHEcon26! What started off as a project to learn LLM use for text analysis turned accidentally into a paper about stigma. [1/n]

26524

Reposted by Clint Harris

☀️ Starshine @starshine.bsky.social · 22/09/2026

What’s your fake starwars academic paper

33360674

Reposted by Clint Harris

ellie lockhart (she/they) @eleanor.lockhart.contact · 15/09/2026

I absolutely think that having a sense of agency or control over one's life comes off as an active threat or condemnation to people who've rejected that. I can't tell you how many times just talking about my life has led to people getting *furious* and implying that I'm saying they're not victims

3695

Clint Harris @clintent-to-treat.bsky.social · 16/09/2026

If so, prosecute them. "Your boss told you to break the law, you complied, then you went to jail", should be the normal version of this sort of thing, and it has to be if rule of law is going to work. Going after the big boys is stupid and pointless. There is a measure 1 mass of them waiting.

000

Reposted by Clint Harris

Mark Copelovitch @mcopelov.bsky.social · 14/09/2026

"Which song on that album do you love best?" Look, I love all my children equally, & "Mr. Jones" is still an absolute banger despite it being the overhyped pop hit. But I love "Anna Begins" the best.

youtube.com

Anna Begins

YouTube video by Counting Crows - Topic

14361

Clint Harris @clintent-to-treat.bsky.social · 16/09/2026

This is just weirdo posting over here on Bluesky, but... the US government announcing that they are against "effective" (the only kind of) altruism, is essentially just saying "Matthew 22:39 for losers." The 2nd most central tenet of the religion that ~62% of Americans follow. MAGA=Death cult.

020

Clint Harris @clintent-to-treat.bsky.social · 15/09/2026

Ok friends, how do you have Claude not burn through your session limit within like 2 questions? I'm on Pro and recently this has been happening. Usually use Opus on extra effort, which I realize is going ti burn fast, but 2 exchanges or so is what I am getting as of this week.

020

Clint Harris @clintent-to-treat.bsky.social · 09/09/2026

The same AI that can prove Fermat's last theorem will also tell you to do a one-sided t-test on a weighted average of a dozen effects based on your favorite guess regarding effect sizes. Thanks buddy!

000

Clint Harris @clintent-to-treat.bsky.social · 09/09/2026

Rolling with the food example, You do not have to use multiple pans or deseed/deskin produce. Whatever recipe you have: 1 pot Heat aromatics Add starches Add proteins Add spices Add stuff that will melt if added sooner Don't drain 1. The nutrients out of your food 2. The time out of your life

000

Clint Harris @clintent-to-treat.bsky.social · 09/09/2026

A wild thing that you realize if you ever interact with the court system in the US is that the "law and order" meme about how justice is applied in the US should literally just be, "People complain to a judge, who has many degrees of freedom due to vague legislation, who rules by fiat."

000

Clint Harris @clintent-to-treat.bsky.social · 09/09/2026

Put tomatillos in Madras curry. Columbian exchange remains unbeatable.

000

Clint Harris @clintent-to-treat.bsky.social · 08/09/2026

Not to get all religious, but the thing that LLMs really lack is inspiration. They just don't have that (doG)' in them.

001

Reposted by Clint Harris

Micah Warren @micah541.bsky.social · 08/09/2026

cloak and dagger navier stokes cims.nyu.edu/~tristanb/st...

cims.nyu.edu

1342278

Reposted by Clint Harris

lastpositivist.bsky.social @lastpositivist.bsky.social · 04/09/2026

If you want to read the paper that will soon ruin @danielmalinsky.bsky.social's life you can check it out here. Forthcoming in Journal of Causal Inference. www.liamkofibright.com/uploads/4/8/...

liamkofibright.com

76719

Clint Harris @clintent-to-treat.bsky.social · 04/09/2026

It seems like an optimal house purchase contract would require the seller (buyer) to pay for inspections if they (don't) find an undisclosed defect. Buyer's optimal inspections unchanged. Seller with known undisclosed defects adds inspection costs to price or refuses deal. Both reveal defects.

100

Clint Harris @clintent-to-treat.bsky.social · 04/09/2026

I watched Interstellar. I'm such a busy dad that it took me 12 years to watch a 3 hour movie that I'm pretty sure came out a few months ago. Best movie about fatherhood ever?

010

Clint Harris @clintent-to-treat.bsky.social · 03/09/2026

As a person with a B- understanding of both empirical welfare maximization and bayesian statistics, it is very surprising to me that the former does not seem to systematically rely on the latter.

000

Clint Harris @clintent-to-treat.bsky.social · 02/09/2026

It's wild that people will insist on bounds and such if you estimate treatment effects using survey data with nonresponse, but are fine with parametric predictions for differences in differences. I will concede that these might be different people.

static.klipy.com

Pam: They're The Same Picture - The Office

ALT: Pam: They're The Same Picture - The Office

000

Clint Harris @clintent-to-treat.bsky.social · 02/09/2026

May all swing voters get lost on their road trips.

000

Clint Harris @clintent-to-treat.bsky.social · 15/08/2026

Loved this. We need more Yi(d,x1,x2,....)-Yi(d',x1,x2,...) in causal inference and less Yi(d)-Yi(d'). And maybe make sure x1,x2,... aren't set to externally invalid values.

140

Clint Harris @clintent-to-treat.bsky.social · 14/08/2026

The funniest case of this sort of thing is of course private educational records, because a huge amount of the relevant human capital theory assumes not only that all educational activity is public knowledge, but that its observability is a first-order reason students engage in it.

100

Reposted by Clint Harris

Ian Hussey @ianhussey.mmmdata.io · 14/08/2026

I’ve been told for years that it’s legally and ethically impossible to share medical RCT data. In 2024, the BMJ made sharing anonymised data mandatory, and now lots* of folk do it. Interesting how the barriers evaporated like that, isn’t it? *not everyone, compliance problems exist.

medrxiv.org

Availability of clinical trial individual patient data in the BMJ before and after adoption of a stringent data-sharing policy, compared with recent rates at other major medical journals

Background In 2024, the BMJ updated its data-sharing policy for clinical trials, requiring open deposit of deidentified individual participant data (IPD) before publication. We considered whether data...

510040

Clint Harris @clintent-to-treat.bsky.social · 13/08/2026

Started thinking about a synthetic control analog for survey nonresponse using always observed covariates and realized I was reinventing MAR adjustments.

000

Reposted by Clint Harris

Anna Hasselqvist @annahasselqvist.bsky.social · 10/08/2026

New WP! ✨How does joint vs. sole custody shape life after divorce? Novel data from court records spanning 25 yrs merged w. SWE admin data. We find that joint custody boosts dads' earnings and mental health, no effects for moms. Kids enjoy better school outcomes and no change in mental health.

131

Clint Harris @clintent-to-treat.bsky.social · 11/08/2026

The way I think about this is an intent-to-treat IV where the emails/letters/etc inviting subjects to treatment aren't actually sent. 50% chance of a positive in-sample first stage entirely driven by chance correlation with endogenous treatment choice. Ergo, weak IV is biased toward OLS.

000

Clint Harris @clintent-to-treat.bsky.social · 06/08/2026

Inference enthusiasts will literally assume empirical estimates are selected draws from the distributions of actual answers to research questions rather than going to therapy.

020

Clint Harris @clintent-to-treat.bsky.social · 05/08/2026

www.nber.org/papers/w35544 I enjoy this sassy abstract from @johnlist.bsky.social. I wonder how well the number of enrollment invites each subject receives prior to enrolling would work as IVs in lieu of incentives, a la Behaghel et al (2015) or my working paper arxiv.org/abs/2404.17693.

nber.org

Two Selection Problems, One Bias Term: Experimental Sign-Up Is Treatment Choice in Disguise

Founded in 1920, the NBER is a private, non-profit, non-partisan organization dedicated to conducting economic research and to disseminating research findings among academics, public policy makers, an...

131

Clint Harris @clintent-to-treat.bsky.social · 30/07/2026

Random comment about housing policy: I recently moved into a brand new mass-produced spec home. It was the same price as similar sqft houses nearby, everything is new/green/efficient, nicest place I have ever lived by far. I guess some people have just decided to not let other people have this?

000

Reposted by Clint Harris

Matt Darling @besttrousers.bsky.social · 29/07/2026

Politics aside, do people believe federal government gas stations with deeply discounted prices are a good POLICY idea? What are they intended to accomplish, in your view? What problem do they solve? www.cnn.com/2026/07/09/b...

cnn.com

Trump promotes ‘Freedom Fuel’ stations that sell cheap gas. How they’re able to do so is unclear | CNN Business

President Donald Trump is heralding a new chain of gas stations that are selling gas for $3.479 a gallon, well below both market prices and wholesale costs.

9791389

Clint Harris @clintent-to-treat.bsky.social · 28/07/2026

Litmus test for "do something about corruption" Democrats. Ball is your court, Newsom. Do we have to follow laws or not? Just the ones we like? Repubs have said Dems don't care about rule of law for decades. Are they right, or do we just enforce laws that make sense to us (just like they do)?

010

Clint Harris @clintent-to-treat.bsky.social · 25/07/2026

This reminds me of my question "are peanut butter and jelly substitutes or complements?" Introductory Micro question. The right answer is it depends on the person.

000

Clint Harris @clintent-to-treat.bsky.social · 23/07/2026

Hoho, Claude Opus just said an RCT that randomly assigned all men or all women to treatment and used the other group as a control would be "an identification catastrophe". They must be hawking Fable. Clearly this is a consistency catastrophe. Identification is fine.

000

Clint Harris @clintent-to-treat.bsky.social · 16/07/2026

Borusyak and Hull (2023) is so simple that I was able to apply their method after only reading the abstract (complimentary).

100

Clint Harris @clintent-to-treat.bsky.social · 15/07/2026

Planning on writing a paper called "Double Debiased AI Bullshit" where I ask ChaptGPT for a Model-Based identification strategy And Claude for a Design-based identification strategy Then cram them together.

121

Reposted by Clint Harris

Mayor Zohran Kwame Mamdani @mayor.nyc.gov · 14/07/2026

This morning in Biddeford, Maine, a 26-year-old man said goodbye to his wife and daughter and left for work. Moments later he was dead, shot in the head by ICE agents, the second man ICE has killed in six days. ICE is killing our neighbors. ICE cannot be reformed. Abolish ICE.

10746505120335

Reposted by Clint Harris

Mark Copelovitch @mcopelov.bsky.social · 14/07/2026

This was yet another wildly illegal, immediately presidency-ending scandal if any Democrat since Truman had done it. And we just yawned & moved on as a society, along with all the other wildly illegal things. A total failure of all 535 members of Congress to uphold their oaths & do their damn jobs.

424496

Clint Harris @clintent-to-treat.bsky.social · 10/07/2026

As an aside, perhaps companies should be liable for damages from false reports. I had a snag with my mortgage - my credit report showed $xxxxxx student loans that had been forgiven. I was fine - but if a company can't keep their records straight, people shouldn't get shafted.

110

Reposted by Clint Harris

The Review of Economic Studies (REStud) @reveconstudies.bsky.social · 01/07/2026

🚨🚨 NEW ECONOMICS JOURNAL 🚨🚨 We are excited to announce the launch of The Review of Economic Studies: Insights, a new open-access journal in the REStud family. Mission: publish concise, high-quality research on important questions across economics and economic policy.🧵1/13

17941

Clint Harris @clintent-to-treat.bsky.social · 26/06/2026

Preregister your Diff in Diffs! Watch out for those Ashenfelter dips!

000

Reposted by Clint Harris

derek guy @dieworkwear.bsky.social · 16/06/2026

RIP Mark Rothko. You would have loved how the hydrogen peroxide bleach is only turning the Reflecting Pool blue around the perimeter

1441382347414

Clint Harris @clintent-to-treat.bsky.social · 15/06/2026

Is it inside trading if I buy low and sell high 15 times in a row?

000

Clint Harris @clintent-to-treat.bsky.social · 24/05/2026

Negroni revision 2 parts gin 1 part Aperol 1/2 part sweet vermouth 1/2 part St Germain 1 splash fernet branca menta

010