juri retweeted
TMLR has faced a deluge of submissions, necessitating stricter desk rejection policies due to limited reviewer capacity
Co-EiC Nihar Shah reached out to authors of 10 papers slated for desk reject. Could they answer questions about their *own* submission?
medium.com/@TmlrOrg/asking-a…
juri retweeted
I don't get it. When it comes to code, AI should be way smarter than me. Yet there are dozens of examples where I ask the LLM to optimize something, it completely fails, I suggest "What if we do this?" and it says "Wow, great idea. That will work." And of course, I'm skeptical--maybe it's being sycophantic--but then we implement my idea and it's legitimately much more performant!
I am not some hotshot software engineer; I don't understand how I'm able to easily find improvements that frontier LLMs overlook. And this happens all the time!
What am I bringing to the table that it doesn't already have?
juri retweeted
Does your company need more attention or investment? Just tell the world that there's a 10% chance your product will kill us all.
juri retweeted
Years ago you only needed papers when applying for job as post-doc or faculty. Now secondary-school students are using LLMs to write papers, to help them get into a good UG university... Inevitable result is wave of AI slop, which (as below) "authors" may not even understand.
We are seeing a new trend in submissions to @arxiv (and presumably to conferences and journals): Authors submitting papers whose contents they likely do not understand. 1/
juri retweeted
Anthropic and OpenAI delaying their IPO, because their S-1 will reveal that they are bleeding money, and have no path to profitability.
Solution? "AI slowdown", so they cut costs for training new models.
It has everything to do with IPO, and nothing to do with safety.
juri retweeted
I'm fascinating by how miserly AI progress is.
A lot of us assumed that once AI was as capable as it is now, we'd see insane progress in the things we actually cared about like number of useful projects completed. But that hasn't really happened.
It amazes me that the machines have literally proved a millennium prize problem, one of the hardest challenges in math, and yet somehow the result is still almost useless to us.
It's just a mountain of crap we humans have to dig through in hopes of salvaging crumbs of insights.
AI gives answers but in the least useful way possible.
juri retweeted
LLM review weirdness...
Just renaming an uploaded pdf from "paper.pdf" to "paper_final_draft_pdf_ready_for_review.pdf" boosts average scores
(gpt-5.6-terra)
juri retweeted
Brief update on the progress of the EMNLP 2026 AI Reviewing Experiment and a timeline for releasing the AI reviews to opted-in authors 2026.emnlp.org/ai-reviewing-…
#EMNLP2026
juri retweeted
When Claude writes code, it signs the commit: Co-Authored-By: Claude. I'm paying for the model, and it still takes a credit line. Okay, fine.
But co-authorship is a claim about contribution. And that logic should work both ways.
So what about all those people who involuntarily "contributed" the training data? You know, half the internet. Don't they deserve co-authorship on the models?
juri retweeted
Testing Text Embeddings in 275+ languages!
Happy to share our recent work.
#nlproc #machinelearing
arxiv.org/abs/2607.00171
juri retweeted
the situation with writing right now is really bad! you can almost feel the mode collapse. everyone is writing with AI all the time. and when they don't write with AI, they produce text that sounds so much like AI, you can barely tell. literally getting worse every month, scary
juri retweeted
Kolumne: Was wir mit #KI gerade erleben, ist etwas anderes als der Einzug von Taschenrechnern, Navigationssystemen, Google oder der Wikipedia in den Alltag - wenn ChatGPT oder Gemini die Arbeit machen, denkt niemand mehr. Der kognitive Geizkragen freut sich,aber das hat Folgen.
ALT Kolumne: meinung •Auf Knopfdruck generierte Texte Warum ich seit KI anders über digitale Technik denke Eine Kolumne von Christian Stöcker Politiker prompten Gastbeiträge, Medienunternehmer lassen ihre Meinungsartikel von KI schreiben, Schüler ihre Hausaufgaben. So neu die Technik sein mag: Die kognitive Psychologie weiß längst, was das bewirken wird. 21.06.2026, 11.35 Uhr
juri retweeted
Apparently ≥3 authors have ≥40 submissions to this ARR cycle.
I think we can reasonably cap submissions per author at something significantly lower than this tbh.
aclrollingreview.org/may26-c…
juri retweeted
Very annoyed with 2-year+ old papers whose "code and data" are "released" in an EMPTY REPO. Should we check if everything is out there before acceptance? If the main contribution is the CORPUS and the corpus is NOT THERE, shouldn't that be an ISSUE? 😩😩😩
juri retweeted
There's a lot of research papers on AI for science focused on ML engineering, but as of yet I haven't seen any examples of an AI-discovered ML algorithm that has actually seen mainstream acceptance.
Are there any examples of this?
