Danish Pruthi
@danishpruthi.bsky.social
Faculty at the Indian Institute of Science in Bangalore. Interested in LLMs, NLP, culture and people.
Reposted by Danish Pruthi
I've become a superpredictor just through this rule of thumb: future AI will improve faster than you think, but current capabilities are always exaggerated. Another example from @danishpruthi.bsky.social
All That Glitters is Not Novel: Plagiarism in AI Generated Research
Automating scientific research is considered the final frontier of science. Recently, several papers claim autonomous research agents can generate novel research ideas. Amidst the prevailing optimism,...
arxiv.org
February 25, 2025 at 10:50 PM
I've become a superpredictor just through this rule of thumb: future AI will improve faster than you think, but current capabilities are always exaggerated. Another example from @danishpruthi.bsky.social
Reposted by Danish Pruthi
Yep! To add to the insult, we know that a lot of "AI-generated original research ideas" are plagiarised. I ended up writing a blog post about it jilltxt.net/can-an-ai-ag... - see also @danishpruthi.bsky.social thread on X about their study finding the plagiarism x.com/danish037/st...
Can an AI agent really do “PhD level research”?
OpenAI plans to charge $20,000 (USD) a month for an AI agent that can do “PhD level research”. Maybe all the PhDs and postdocs recently fired by DOGE should band together and sell their…
jilltxt.net
March 6, 2025 at 11:03 AM
Yep! To add to the insult, we know that a lot of "AI-generated original research ideas" are plagiarised. I ended up writing a blog post about it jilltxt.net/can-an-ai-ag... - see also @danishpruthi.bsky.social thread on X about their study finding the plagiarism x.com/danish037/st...
My #ICLR2025 story: Reviewed 3 papers. Blocked a couple of hours to engage with authors. Of the three papers, 1 is withdrawn, and 2 have no author responses (for any of the reviews).
And yes, no reviewer engagement on the paper we submitted.
And yes, no reviewer engagement on the paper we submitted.
November 25, 2024 at 7:50 AM
My #ICLR2025 story: Reviewed 3 papers. Blocked a couple of hours to engage with authors. Of the three papers, 1 is withdrawn, and 2 have no author responses (for any of the reviews).
And yes, no reviewer engagement on the paper we submitted.
And yes, no reviewer engagement on the paper we submitted.