Mike Burnham
@ML_Burn
Followers
808
Following
656
Media
65
Statuses
683
Incoming AP @tamupols, Postdoc @PUPolitics, dual Ph.D. @psupolisci & @CSoDA_PSU. Text analysis & deep learning, methods, American politics and accountability.
Princeton, NJ
Joined May 2019
New Pre-print out today! We're releasing Political DEBATE -- a new set of language models for zero/few-shot classification of political text. The models are open source, small enough to run on your laptop, and as good as proprietary LLMs within domain.
2
48
206
You want mandatory voting to improve the health of our democracy. But I want it so that I don't have to listen to debates about voter enthusiasm vs. appealing to the median.
1
0
1
Some will interpret this as outsourcing your work to an LLM. It’s not any more than peer review is outsourcing your work. LLMs are a cheap way to get feedback on empirics and the models are good enough now that if it finds criticisms you should at least consider them.
There’s no reason to be making these mistakes anymore. I’m reviewing my third paper in a row with obvious specification problems. Wrong fixed effects, wrong clustering, etc. In all three, I checked and GPT’s o3 also caught the main problems. If you’re a little unsure, ask it.
0
0
6
Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.
398
661
3K
Imagine you draft an article, use and LLM to help edit, and accept some of the proposed revisions after reviewing them. Then you submit to a journal which accuses you of plagiarism because a particular keyword is in there one too many times.
1
0
14
I just don't think this is reasonable. LLMs can be productively used in editing and drafting while still maintaining authorship of the text. Calling these uses plagiarism seems like a rather draconian position to take.
2
0
17
@MattGrossmann It draws a pretty clear line that using an LLM in writing is plagiarism: .". the line between plagiarism and acceptable use of LLMs is clear. Like Wiki sites, LLMs can provide good background information and brainstorming, but should not be used for direct text development".
1
0
3
To clarify, it's the article cited and not @MattGrossmann making this assumption, and it makes some. bold declarations. "Using this keyword detection method, this article shows that the spike in AI-generated manuscripts has almost certainly already compromised the profession".
1
0
10
I don’t think we should equate any use of AI assistance in writing with plagiarism. There are valid and productive ways of using AI to improve your communication. Important data, but assuming plagiarism from keyword use is just too much IMO.
Popular ChatGPT keywords suggest AI use for writing political science academic articles.
1
5
75
Breaking: Trump eliminates all tariffs to China on grounds that “everything is computer”.
Tim Apple wins again: "Trump Exempts Phones, Computers, Chips From ‘Reciprocal’ Tariffs"
0
1
7
Genuinely impressed by the craftsmanship that went into this bait.
1
1
8
Occasionally, I get emails from high school students looking to do part time RA work. I find this depressing. This is not how I want my kids spending their time when they are 16.
3
1
73
Sure it's petty. But nowadays if you're not using AI to protect yourself that's on you.
0
0
4
Look I'm obviously pretty pro AI, but this whole Ghibli thing should make clear how profoundly unfair this technology can be. The fact that someone's distinct contribution to the world can be commoditized against their wishes and with no compensation should give us pause.
0
1
8
The most outrageous part of these accusations is that we have any clout.
he is getting a lot of hate for it but I think Matt Yglesias is (unfortunately) absolutely, 100% correct. Unfortunate, because academic social science has a lot of clout still and people tend to blindly trust it, but a lot of it (though not all) is pretty lacking in rigor.
0
0
20
Going to start turning all of my papers into board books. No child of mine is going to enter kindergarten without a research agenda.
It's release day for How Politicians Polarize!. To celebrate, my husband made a board book version so we can teach our 14-month-old about political representation in an age of negative partisanship. Some key findings from the book (in baby-friendly language)🧵
1
5
52
Can someone help me estimate the expected value of this bet? I’m not a Harvard trained economist.
"Would you bet your paycheck on a weather forecast for tomorrow? If not, then why should this country bet billions on ‘global warming’ predictions that have even less foundation?". — Thomas Sowell.
0
1
10
If literacy rates are so poor then explain how everyone read and formed an opinion about that abundance book in less than 24 hours.
1
0
4
Thousands of people speculating about what is entirely an artifact of the algorithm used to plot the network.
0
0
4