Hacker News: Giwwihttps://news.ycombinator.com/user?id=GiwwiHacker News RSShttps://hnrss.org/hnrss v2.1.1Thu, 23 Apr 2026 18:30:27 +0000<![CDATA[New comment by Giwwi in "The Graveyard of Unabsorbed Knowledge"]]>Hi HN, I'm the author. I recently finished a draft of a book on the political economy of AI, arguing that our main economic bottleneck is shifting from generating information to verifying it.

This essay is a spin-off from that research. It’s about 'stranded knowledge' — research that is mathematically or scientifically sound but gets buried because it lacks institutional legitimacy at the time (like Hochreiter’s 1991 thesis before LSTM took off).

My argument is that as AI drives the cost of generating new papers to near zero, the archive decays 'attentionally'. I think the highest-return use of AI right now isn't writing more papers, but building a triage system to recover the neglected ones.

I’d love to hear your thoughts, especially from anyone working on metascience or AI discovery tools.

]]>Mon, 13 Apr 2026 14:57:27 +0000https://news.ycombinator.com/item?id=47752955Giwwihttps://news.ycombinator.com/item?id=47752955https://news.ycombinator.com/item?id=47752955<![CDATA[The Graveyard of Unabsorbed Knowledge]]>Article URL: https://verificationbottleneck.substack.com/p/the-graveyard-of-unabsorbed-knowledge

Comments URL: https://news.ycombinator.com/item?id=47752933

Points: 2

# Comments: 1

]]>Mon, 13 Apr 2026 14:56:24 +0000https://verificationbottleneck.substack.com/p/the-graveyard-of-unabsorbed-knowledgeGiwwihttps://news.ycombinator.com/item?id=47752933https://news.ycombinator.com/item?id=47752933