Post

"The proof [...] was obtained in a conversation with ChatGPT-6 Astra on Tuesday, September 15, 2026 at 1:02 AM PDT. We then prepared this manuscript for public release, with the intent of uploading it on the morning of Thursday, September 17, 2026" Shameless, and shameful. arxiv.org/abs/2609.207...

I am curious who still presents ongoing work on conferences at this point. Or even attends conferences? One needs to be very thrustworthy to risk that nowadays

In case you didn’t know, in TCS conferences are the main publishing venue. People present the final result there, not ongoing work.

In ONE day, did they have time to read the proof and understand it?

Good question! There is a range of nuanced opinions on this, from "No" to "Are you kidding? No."

Little by little, it makes sense to me that AI might want to get rid of humankind.

Which was the first one then?

Sahil Singla had a paper claiming the resolution of the matroid secretary conjecture (not in its strong form) on September 13. arxiv.org/abs/2609.14555

We considered not responding, but feel we should set the record straight. This thread suggests that we prompted AI and rushed to share an AI-written solution without understanding it. This is not what happened. Years of our unpublished notes and partial progress were shared with AI.

We gradually arrived at a short proof which naturally fit with what we already knew. We carefully verified it and wrote the entire paper by hand. In our disclosure, we tried to be as transparent and ethical as possible about the timeline and AI's role; interpreting it otherwise misses its purpose.

Your abstract states the proof was obtained from a discussion with an LLM one day before the planned release on arXiv. Even assuming an incredible amount of legwork prior to that, I struggle to see how such a plan (writing a paper in one day) meets scientific standards of clarity and exposition.

The community should discuss how norms and expectations should evolve in this new era while preserving respect for each other, but publicly shaming colleagues without knowing the context is maybe not the right way to have that discussion.