David Van Dijcke

I'm excited to share my job market paper (for the 2025-26 market)! It introduces a new extension of RDD where outcomes are entire distributions: Regression Discontinuity Design with Distributions (R3D). Thread below 👇 (1/)

Following @packlesshepherd we are adding BYOK option to openaireview. Now you can easily review longer papers without setting up the package locally.

@lopez_prol @CoarseDotInk Great! Any idea what went wrong the first two?

@packlesshepherd @CoarseDotInk Update: it worked at the third try, thank you!

Big update on @CoarseDotInk: you can now run reviews at ~zero marginal cost through your AI subscriptions! Just drop your paper into coarse.ink, click "Review with my subscription" and follow the instructions See coarse.ink/setup --> "Use my subscription"

@lopez_prol @CoarseDotInk Yes, copy paste the command in codex terminal, after launching codex with "codex --yolo"

@packlesshepherd @CoarseDotInk I'm trying this with Codex but keep getting command approval requests. Is there any way to avoid the requests?

@_joaodias @CoarseDotInk Glad you like it and find it useful! Resuming the review is currently not implemented but seems worthwhile to do, I've added a GitHub issue for future versions. For now I'd recommend setting the reasoning effort lower or using the online version with an API key (pricier)

@packlesshepherd @CoarseDotInk Hey David, wonderful tool you built. I was using it on an early draft locally with my Codex Plus subscription and just ran out of limits. When the limits reset, is it possible to continue the @CoarseDotInk revision from where it stopped, or will it always start over?

1,000+ reviews served, 1.6K PyPi downloads, in a little over a week. Thanks for all the love and feedback :)

@PeiranXiao @guilhermejd1 @CoarseDotInk Just chunk them together into one file! I can add support for multiple files if you think that'd be useful

@packlesshepherd @guilhermejd1 @CoarseDotInk Does it work if I have multiple tex files for the same paper?

@fatihkansoy @CoarseDotInk is not for profit, costs typically less than 2$, and can be run at zero marginal cost on your local subscription though :) Good prompting can get you a good chunk of the way but doesn't make up for a good harness on top

Bunlara yüzlerce dolar para vermenize gerek yok. Hepsinin çalışma prensibi aynı ve bunlar sizde de mevcut. Yapılan şey: iyi bir prompt ile seçili bir LLM modeline aynı işi tekrarlatmak. Dolayısı ile güzel bir prompt ile benzer hakem raporlarını siz de oluşturabilirsiniz.

Akademisyen arkadaşlara: kaşıkla kazandığınızı kepçeyle vermeyin :) Son dönemde yapay zeka ile bir sürü araç çıkıyor. literatur taraması yapan/yazan vs ve son dönemde hakem değerlendirmesi (referee/review) yapan araçlar çoğalmaya başladı. örneğin refine ink, course ink vs +

@sanjayrajsingh @ChatGPTapp @CoarseDotInk @RefineDotInk @reviewer3com To be clear, for the local version the only API requests are for the PDF OCR the rest runs on your code subscription

@sanjayrajsingh @ChatGPTapp @CoarseDotInk @RefineDotInk @reviewer3com Thank you! Yes I'd be happy to do that. Could you request this through a GitHub issue?

@guilhermejd1 @CoarseDotInk Yeah tex files skip the ocr step. The agent might still ask for a key in the way it's currently prompted. Let me fix that!

@packlesshepherd @CoarseDotInk It seems to require OpenRouter for OCR. But what if I introduce the tex file? Or then a txt from a pdftotext?

Post-script: Since this allowed me to run more reviews without breaking the bank, I've updated the comparison table on coarse.ink/compare with @ChatGPTapp-5.4 results. @CoarseDotInk outperforms @RefineDotInk, @reviewer3com, and Stanford Agentic Reviewer across the board!

I believe this is a big step towards making not-for-profit, open-source AI reviews cheap and available to as many academics as possible. @ChicagoHAI @ChenhaoTan This is a first release, so please submit any issues here: github.com/Davidvandijcke…

400+ reviews served in just a couple of days! All running on free-ish infrastructure with no institutional backers. 'coarse <3 that you <3 open-source. Keep 'em coming!

If you have working papers that are not too polished and you're willing to let Coarse use them as demonstrations on its examples page, let us know - we'll give you endless not-for-profit full reviews as a token of appreciation. Please DM or reply.