Tootfinder

Purely Semantic Indexing for LLM-based Generative Recommendation and Retrieval
Ruohan Zhang, Jiacheng Li, Julian McAuley, Yupeng Hou
https://arxiv.org/abs/2509.16446 https://

Purely Semantic Indexing for LLM-based Generative Recommendation and Retrieval
Semantic identifiers (IDs) have proven effective in adapting large language models for generative recommendation and retrieval. However, existing methods often suffer from semantic ID conflicts, where semantically similar documents (or items) are assigned identical IDs. A common strategy to avoid conflicts is to append a non-semantic token to distinguish them, which introduces randomness and expands the search space, therefore hurting performance. In this paper, we propose purely semantic index…

@borkdude, in your 'scittle with additional libraries' files, you have, for each file in the library, a line
SHADOW_ENV.evalLoad( <filename>, <boolean>, <???>);
where ??? is a string of JS source code with some punctuation replaced by hex codes.
The second last line appears to invoke the file itself (recursively!?!), and the last is always
SHADOW_ENV.evalLoad("shadow.module.<filename>.append.js", false, <???>);
Wha…

I think it's fine to document the many AI fuckups, but there's so often a tone of, essentially, "But what if this search results page was read by someone with shallow literacy?"
(I sometimes mentally append the line from the heroin-using Chris Morris character: "But what about other people less stable, less educated, less middle-class than me?")

Do you want to use append-only WORM (Write Once Read Many) files for logging?
For ~ $100, you can use a Raspberry Pi as an external USB hard disk with an ext4/exFAT filesystem providing some preconfigured append-only WORM files. Try Socarrat:
https://gitlab.eif.urjc.es/publications/soca<…

Note that the headline treats this product’s output as established fact:
❝Major AI conference flooded with peer reviews written fully by AI❞
They didn’t even have the journalistic decency to append “claims AI detection company.” Note the passive voice in the sub-headline:
❝Controversy has erupted after 21% of manuscript reviews for an international AI conference were found to be generated by artificial intelligence.❞
“were found”…by whom? Nobody. It’s just a magic fact.
3/

Context Lineage Assurance for Non-Human Identities in Critical Multi-Agent Systems
Sumana Malkapuram, Sameera Gangavarapu, Kailashnath Reddy Kavalakuntla, Ananya Gangavarapu
https://arxiv.org/abs/2509.18415

Context Lineage Assurance for Non-Human Identities in Critical Multi-Agent Systems
The proliferation of autonomous software agents necessitates rigorous frameworks for establishing secure and verifiable agent-to-agent (A2A) interactions, particularly when such agents are instantiated as non-human identities(NHIs). We extend the A2A paradigm [1 , 2] by introducing a cryptographically grounded mechanism for lineage verification, wherein the provenance and evolution of NHIs are anchored in append-only Merkle tree structures modeled after Certificate Transparency (CT) logs. Unlik…