Brad LeVeck (@bradleveck) header
Brad LeVeck (@bradleveck) profile picture

political scientist at UC Merced Likes: 🌄📸, 🧠🔄🧠, 📊🧪

Mountain View
Joined April 2008
Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)Photo by Brad LeVeck (@bradleveck)
Brad LeVeck retweeted
🚨Job alert 🚨 UVA is hiring an assistant professor in CP! Application link below 👉🏻
3
23
84
18,650
Brad LeVeck retweeted
Great thread
I follow the AI & math discourse a bit - without understanding the actual math. I thought I might offer some thoughts as a scientist in structural biology - a field transformed by various tech over the years, and recently by AlphaFold. It might perhaps interest @nasqret et al.🧵
2
15
2,636
Brad LeVeck retweeted
The challenge & plea fall on the readers: it is their job to know that parallel trends means "I am assuming away specific non-causal explanations for a correlation." Proper calibration of sales pitch must be incentive compatible for authors. That comes from readers.
-> "My plea to economists: stop overselling your difference-in-difference results. They don't say anything about aggregate effects, so don't pretend they do." The challenge is that academia tends to reward heavily overselling results.
1
1
7
1,206
This guy did a true public service: if you didn't know that interchange, it was impossible to know which way to go to get on I-5 North. It was totally confusing, so after Ankrom did this, the city left the sign up for 8 years, and eventually replaced it with a sign of its own.
Before dawn, Richard Ankrom gets in his truck, drives to a Los Angeles overpass, and installs an amendment to an existing highway sign: “NORTH,” with the Highway 5 logo. He calls it “guerrilla public service” - helping traffic when nobody else would.
3
7
164
25,618
Brad LeVeck retweeted
Stable trust in AAUP is not a structural constant of the universe. It’s a result of actions over many years, specifically not doing stuff like this. AAUP has no special expertise on this race. Its self-politicization will damage the organization with no effect on elections.
6
20
3
147
27,766
Brad LeVeck retweeted
When scientific journals started endorsing candidates, this led to a drop in trust in science and had no meaningful electoral consequences. All downside for science. No upside for the endorsed candidates.
4
91
3
641
20,003
Brad LeVeck retweeted
I am not an AAUP member, but I think it is a really bad idea for the American Association of University Professors to be endorsing candidates, especially in a factional Democratic primary.
32
70
86
698
488,416
Brad LeVeck retweeted
🚨 New in APSR! 🚨How do the origins of authoritarian regimes shape their subsequent trajectories? @kbclarke @jackpaine_prof and I are excited to share "Rebel Victory and Authoritarian (Dis)Order" /1
3
41
1
155
10,436
It took longer than expected but it’s finally here. And yes, it’s brutal: The core change Duration of status (D/S) ends for F-1 students. - You now get a fixed admission end date instead of an open-ended stay. - New admissions are for your program length, capped at 4 years. - If your program runs longer than 4 years, you must file an Extension of Stay (EOS) with USCIS. - To stay past your fixed date, you either file an EOS or leave and re-enter. Current students (transition protection) - If you are already here on D/S, you may generally stay until your I-20 program end date. - But no longer than 4 years after the rule's effective date, whichever comes first. - Years you have already spent in F-1 do not count against that 4-year window. - DHS can delay parts of the new restrictions for up to 2 years after the effective date. Extensions of Stay (EOS) - The DSO issuing a new I-20 no longer extends your status by itself. - You must file an EOS with USCIS to stay past your fixed admission date. - EOS filings may require biometrics (fingerprints, etc.). - If you timely file an EOS, you may keep studying full-time while it is pending. - Poor academic progress is generally not an acceptable reason for an extension. - Academic probation, suspension, or repeated inability or unwillingness to finish your program does not justify an extension. Extension denials - If USCIS denies your EOS after your fixed date has passed, you may have to leave immediately. - You may start accruing unlawful presence at that point. - Form I-539 denials have no ordinary appeal (a motion can be filed but does not automatically protect you). For Graduate students - You generally cannot change your educational objective at any point during a graduate program. - You generally cannot transfer to another school during your program. - The only transfer exception is narrow: SEVP-approved extenuating circumstances (school closure, loss of SEVP certification, program elimination, major disasters). - Advisor conflict, loss of funding, or changing research interests are not clearly protected. Undergraduate students - Cannot transfer or change educational objective during the first academic year. - After completing one academic year, lateral changes and transfers become possible again. Degree-level rules - After finishing a program (post-effective date), you may generally only start a program at a higher educational level. - You cannot move to the same or a lower educational level as an F-1 student. OPT and work authorization - After the transition window, post-completion OPT and STEM OPT generally require two forms: I-765 (work permit) and I-539 (F-1 extension). - They can be filed together but are adjudicated separately. - A pending extension lets you keep studying, but new OPT employment generally cannot start until both the extension and work authorization are approved. - If your EOS is still pending after your admission period expires, you may get an auto-extension of existing employment authorization, up to 240 days, but only under severe economic hardship / emergent-circumstance rules. Transition OPT exception (helpful for you) - Current D/S students filing post-completion OPT or STEM OPT within the first 6 months after the rule takes effect only need the I-765, not a separate I-539. Grace period - The post-completion grace period drops from 60 days to 30 days. - Some current D/S students keep 60 days during the transition. - If you end study or training early, you and your dependents must depart or take action to maintain status within 30 days. Travel - If you leave and re-enter after the effective date, CBP will generally issue a fixed-date I-94, moving you into the new system early. - Travel can also affect a pending EOS application. Dependents (F-2) - Your spouse and minor children still each need their own I-20. - Their authorized stay cannot exceed yours as the principal F-1. Language training student's
Updates coming on CPT/OPT, I-20, and maintenance of stays soon. Likely March. It looks brutal.
12
42
8
237
119,005
Brad LeVeck retweeted
The Wittman-Calvert model has broken containment
The only reason why Democrats don’t *have* to move to the center is because the Republican Party is so bad.
1
1
2
1,235
Brad LeVeck retweeted
I don't have expensive tastes or any particular desire for luxury. But I've done a lot of travel this year, and I've concluded there are three things I want a hotel room to have: - Blackout curtains - Power outlets next to the bed - One button to turn off all the lights.
80
27
5
1,266
278,059
These results make a lot of sense, but I wonder if, on a longer time-scale, students will learn to use AI more like a tutor.
In einer grossen Studie wurde untersucht, wie sich die Leistung von über 26'000 Schülern in China während 30 Monaten veränderte, wenn sie anfingen, KI-Chatbots zu nutzen. Ihre Hausaufgaben wurden rund 20% besser. Sie benötigten für die Hausaufgaben rund 20% weniger Zeit. Das ist super. Aber: Bei Prüfungen (wo KI verboten ist) wurden sie rund 20% *schlechter*. Das ist eine massive Verschlechterung. KI kann Denkkompetenz aufbauen, wenn sie als eine Art Tutor eingesetzt wird. Dann spricht man von kognitivem Scaffolding. Die Realität ist aber, dass die Strategie des kognitiven Offloading der Weg des geringsten Widerstands ist: Denkarbeit an Chatbots auszulagern, ist instrumentell gesehen rational. Ein Fehlanreiz. Diese Entwicklung ruiniert Bildung. Und sie ist ein systemisches Risiko: Was passiert, wenn eine ganze Generation noch weniger als frühere Generationen lernt, eigenständig zu denken?
1
107
Brad LeVeck retweeted
Really interesting post...
The New York Knicks and the martingale property of calibrated probability forecasts (with some simulation and R code) statmodeling.stat.columbia.e…
1
1
543
The Center for Shared AI prosperity is hiring an Executive Director. Help shape a progressive economic vision for a world with AI-driven job loss! We are advancing ideas grounded in both public opinion research and policy analysis. Job description link in next post.
1
10
32
9,510
Brad LeVeck retweeted
Seems like time for a ballot measure to require the UC system to use test-based admissions
California is turning UC Berkeley (whose professors won 4 Nobel Prizes last year) into remedial school. No SAT for admissions and a new mandate to make everyone's "outcomes similar." No mention of excellence. This is how bureaucrats kill the nation’s best public university.
14
23
3
532
53,721
Congratulations to Professor Anil Menon on receiving the Emerging Health Politics Scholar Award from the Health Politics and Policy Section of APSA!
1
1
3
115
MORE: An “open” strait under Iranian management is not a return to the pre-war status quo and would mean that Iran has accomplished a key war aim. ⬇️ Iranian Foreign Affairs Minister Abbas Araghchi discussed the strait at length and noted that it is under Iranian and Omani sovereignty and that both countries will collect fees after the MoU. Iran will also manage the strait. Araghchi rejected the term “tolls” but defended Iran charging “service“ fees for passage through the strait, which is effectively an attempt to rebrand Iran’s protection racket as legal administration. Araghchi added that Iran would provide safe passage through the strait for civilian vessels but create separate arrangements for military vessels. This system of management is at odds with both US policy and long-established maritime legal precedent. US President Donald Trump has repeatedly said, for example, that the strait must be open with no fees or Iranian management. A US official told Reuters that the agreement would “open” the strait, but an “open” strait under Iranian management would be very detrimental to US interests, as ISW-CTP has previously argued in @brian_cartr's essay in the post below. Iran also continues to use force in an attempt to impose the reality that it controls the Strait of Hormuz. US Central Command (CENTCOM) reported on June 12 that US forces intercepted multiple Iranian drones targeting commercial vessels in the strait. Iranian media separately reported explosions near Qeshm and Sirik islands and said the sounds came from warning shots that Iranian forces fired to enforce control over the strait.
NEW: Iranian statements regarding the contents of the US-Iran memorandum of understanding (MoU) indicate that some terms of the agreement are at odds with stated US positions on the Strait of Hormuz in recent months. The agreement has two parts: a first stage that deals with an “end to the war” on all fronts, a resumption of traffic in the Strait of Hormuz, and economic benefits for Iran, including reconstruction funds, sanctions relief, release of frozen funds, and an end to the US blockade. The second stage would deal with the nuclear issue and “one or two other [unspecified] issues.” Other Key Takeaways: Iran is likely attempting to structure the MoU and the phasing of a final agreement to reduce US leverage before later negotiations over the nuclear program. Iran is attempting to access at least some of its frozen assets early in the MoU process, for example, which would give it some economic relief before nuclear talks begin. Iranian media appears to be portraying a possible MoU as a tactical pause in the war rather than a final settlement. The Ghalibaf-affiliated Khorasan outlet argued on June 13 that the emerging agreement only aims to end the current war and does not resolve the underlying issues between Iran and the United States. Iranian Foreign Affairs Minister Abbas Araghchi’s readout of the MoU is largely consistent with IRGC-affiliated media’s readout of the latest possible agreement, which may indicate a growing consensus among Iranian regime leaders on Iran’s red lines for negotiations. That Araghchi’s June 12 readout and IRGC media’s account of the latest US-Iran MoU are very similar suggests that Vahidi or elements close to him may have successfully driven consensus on their preferred policy outcomes.
32
189
31
490
248,222
I hadn't thought about the fact that this result turned 50 this year. Also see @Afinetheorem's excellent, short summary of the theorem, and why it's a good bang-for-buck in terms of mathematization -> non-obvious and important insight. afinetheorem.wordpress.com/2…
Replying to @sapinker
More precisely, rational agents with the same understanding of the world, after examining evidence and sharing their assessments, must arrive at the same conclusion. Or as stated in the paper itself:
169
Brad LeVeck retweeted
because they affect the outcome you care about. We can't take the market response to UC degrees as a constant. Everyone should agree with this, but sadly some seem not to get it or don't want to. ->
2
1
10
709
Brad LeVeck retweeted
There are two points here, (1) the technical point that we can't get good results from partial equilibrium thinking, (2) the substantive point that we have to understand our objective. Point 1 means that you have to consider market adjustments to your policies, ->
I spoke with someone who contributed to another report, separate from the academic senate report, that advised Napolitano in her SAT decision-making. What he said, paraphrasing: “Jelani, what if the admissions objective is not to admit the students who will perform the best academically, but to admit the students who would benefit the most from a Berkeley degree, say as measured by increase in lifetime earnings.” Of course, if you start producing graduates that are substantially weaker, high-paying employers will notice and cut back on hiring your graduates. Their counter is some idealized/wishful thinking that the university experience is so transformational that it can turn any student into a diamond coming out, even if they didn’t know how to add fractions coming in. Link to the second report: ucop.edu/institutional-resea…
1
3
16
2,997