←
Sinan P. = NP2,542 posts
Sinan P. = NP avatar

Sinan P. = NP

@sinanspd

Son of Perdition. Quantum Computing And Concurrency dude @CtrQuantumTech & @VMware | Ex @Twitter Ex @BrownUniversity | CEO Vizion | Singer/Guitars @ Alakazam

◷ Joined August 2013▧ 259 Photos and videos
Steve McCormick avatar
⟳ Sinan P. = NP retweeted
Steve McCormick@Quasilocal·Nov 17

This drives me mad -- I've had AI check things a bit now, and very often it will want to add a heap of details to stretch out something that the reader already understands, and then cling to that part as the main thing to focus on

Human reviewers pull 8.4% of their criticisms from a paper's appendix. Frontier LLM reviewers: 62%. A new study had GPT-6 and Claude 5.5 models review ICLR 2027 submissions, then mapped each weakness to the section it came from. The AI critiques keep going past page 30 of the appendix. They skew toward detail audits: more seeds, significance tests, confidence intervals. Rarely the core idea. Same pattern via API, web chat or coding agent, which points at the models, not the interface. arxiv.org/abs/2611.17811 #PeerReview #LLMs #AcademicTwitter

Sinan P. = NP avatar
Sinan P. = NP@sinanspd·Nov 10

What does everyone use as a blogging tool? I have some things to say about this AI situation

Suresh Govindarajan DLM avatar
⟳ Sinan P. = NP retweeted
Suresh Govindarajan DLM@modularform·Oct 31

It appears that some guy has created a website iopscienceiop.org and posted a paper that is formatted to look like it us from the Astrophysical Journal with several big names as co-authors. @Kaju_Nut Note the missing dot after iopscience in the URL.

Sinan P. = NP avatar
Sinan P. = NP@sinanspd·Oct 24

If this transition is to happen, the right way to do this is probably to have the venue run a predetermined prompt that focuses on errors instead of gaps and provide that report to the reviewer, instead of letting referees run loose & wild with it

Sinan P. = NP avatar
Sinan P. = NP@sinanspd·Oct 24

In order to remotely accept this idea I need to see evidence & explanation on how AI 1) Avoids the endless clarifications & criticism loop today's models suffer 2) Avoids the NeuralIPS statistical significance test fiasco 3) Adapts to field specific norms and expectations

By every measure, Opus 4.8 beats human experts as peer reviewers in a study of 124 Harvard social science papers covering a range from statistical analyses to formal models to qualitative work. As academics, we need to figure out how to adapt the practice of science to AI, fast.

Post attachmentPost attachment
Gautam Kamath avatar
⟳ Sinan P. = NP retweeted
Gautam Kamath@thegautamkamath·Oct 23

How do you tell if authors fabricated their research results? Sometimes: they just tell you. Read, e.g., this caption carefully. Another banger of a post by my friend Nihar B. Shah on his blog, where he found at least 20+ such #ICLR2027 submissions: researchonresearch.blog/2026…

Post attachment
Remy Levin avatar
⟳ Sinan P. = NP retweeted
Remy Levin@RemyLevin·Oct 6

AI is creating new frontiers in professional bullying. Extraordinary.

I am pleased to announce a new professional milestone: Professor Polson was sufficiently unhappy with my post, and the ensuing Washington Post article, that he decided to email *my Dean* an AI-generated review of my body of work. 1/

Post attachment
Sinan P. = NP avatar
Sinan P. = NP@sinanspd·Oct 3

New paper drop: Continuation of my talk at PLDI, my paper on showing how destructive interference can be avoided in classical simulation by creating race conditions, with minimal loss has been published in the IOP Quantum Science Technology journal iopscience.iop.org/article/1…

Sinan P. = NP avatar
Sinan P. = NP@sinanspd·Oct 2

Stanford re-discovered a light version of Matthias Felleisen's code walks

Talked to a Stanford prof last night who told me that AI has led to students all acing all the homeworks, never showing up to office hours, and then failing the final exams in-person. So in response the department is changing the TA's purpose, by eliminating office hours and requiring students meet with TAs 1:1 and walking them through each coding assignment to demonstrate they actually understand what's going on in the code.

arXiv.org avatar
⟳ Sinan P. = NP retweeted
arXiv.org@arxiv·Oct 1

arXiv has updated our policy on rate limiting for all submitters. This update was made to fairly distribute moderator time & support the arXiv community of staff, volunteers, readers & authors. Please read our announcement to learn more: blog.arxiv.org/2026/10/01/up…

Post attachment
Sinan P. = NP avatar
Sinan P. = NP@sinanspd·Aug 7

3 cancelled flights... thank you for leaving me stranded in the most expensive city in the country @united without providing any accommodation

Post attachmentPost attachmentPost attachment