Assistant Prof @sbucompsc @stonybrooku. Researcher → @SFResearch Ph.D. → @ColumbiaCompSci Human Centered AI / Future of Work / AI & Creativityhttps://bsky.app/profile/tuhinchakr.bsky.socialhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3muzcf3fk4226Brilliant read in todays @newyorker.com essay on utility of books in LLM pre-training with quotes and inputs from @mmvty.bsky.social and @mellymeldubs.bsky.social08 Sep 2026 15:06 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3muzcf3fk4226https://bsky.app/profile/tuhinchakr.bsky.social/post/3mux5m5irck2fBMC is a pooled metric should not be surprising ? It was explicitly mentioned. k=5 is deliberately sensitive and does contribute to BMC but central extraction doesn’t depend on 5 word matches. Some books have 24-69% of entire text covered by exact spans > 50 words and single generations > 100 words [contains quote post or other embedded content]07 Sep 2026 18:36 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mux5m5irck2fhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3munz42mqis2r👇👇👇 [contains quote post or other embedded content]04 Sep 2026 03:21 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3munz42mqis2rhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3mumrlhl5pk2pAnd guess what @nature.com published it :) Clearly no institutional bias [contains quote post or other embedded content]03 Sep 2026 15:34 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mumrlhl5pk2phttps://bsky.app/profile/tuhinchakr.bsky.social/post/3mtwv7jgea22hHappy Pub Day to Rachel Cusk. If I don’t respond to your emails on time this week forgive me.25 Aug 2026 22:40 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mtwv7jgea22hhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3mtwrny2fz222Secured my copy!! [contains quote post or other embedded content]25 Aug 2026 21:37 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mtwrny2fz222https://bsky.app/profile/tuhinchakr.bsky.social/post/3mtwjghptxc2eWill James Taranto write an oped now stating why SlopEds are okay ? [contains quote post or other embedded content]25 Aug 2026 19:09 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mtwjghptxc2ehttps://bsky.app/profile/tuhinchakr.bsky.social/post/3mtmgpzka622zBlake Lefray has a very convincing video that proves why Daggermouth might be AI generated. For folks that don’t believe AI detectors these kind of patterns occur disproportionately in AI writing. It’s no magic that AI detectors that are neural networks catch them.21 Aug 2026 18:54 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mtmgpzka622zhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3mtjzcmo5pc2rAfter the Daggermouth story broke out a group of angry mob on BookTok and Threads said my hypothesis on rare expression correlation with AI writing was weak. Still stand by it. Here is an evidence. You will only find this v awkward cliche in post 2024 books. Is AI changing writing?????20 Aug 2026 19:49 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mtjzcmo5pc2rhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3msxuyirwe22x@jameszou.bsky.social might be able to comment better. Also wouldn’t it be nice to write a paper with new results in the light of mounting evidence that not all AI detectors are biased ? Almost everyone piggybacks on this I know it came out in 2022 which is ancient given how fast things move in AI [contains quote post or other embedded content]13 Aug 2026 14:44 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3msxuyirwe22xhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3msh2kyeayk27The most popular comment on the @theatlantic.com article talked about the futility of a bestseller list. So its only ironic that Daggermouth is officially a @nytimes.com best selling book.06 Aug 2026 22:08 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3msh2kyeayk27https://bsky.app/profile/tuhinchakr.bsky.social/post/3mry2y25e2c2hVery grateful to @ai2.bsky.social for tools like OLmO Trace and Infingram for helping us to draw the boundary between whats human written and whats ripped off from the training data 😈 [contains quote post or other embedded content]31 Jul 2026 23:06 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mry2y25e2c2hhttps://bsky.app/profile/tuhinchakr.bsky.social/post/3mrxsdaqmhk2kEchoes @vtanger.bsky.social premonition !! [contains quote post or other embedded content]31 Jul 2026 20:31 +0000at://did:plc:itpy7gpwnd36ac6pm47tuwuu/app.bsky.feed.post/3mrxsdaqmhk2k