Found this quote in the opening of Paper-Making by Jules Heller:

RAGS make paper,
PAPER makes money,
MONEY makes banks,
BANKS make loans,
LOANS make beggars,
BEGGARS make
RAGS.

What a great start to a book on making paper by hand.

#papermaking #rags #quotes

🇺🇦 #NowPlaying on #BBC6Music's #BLOCKPARTYWithHueyMorgan

Rags:
🎵 Ice (feat. Bambi)

#Rags

https://mikegalbraith.bandcamp.com/track/ice-box

"Without clear signposting, an agent might miss your API entirely. If it does discover it, large language models (LLMs) may stumble with undocumented behaviors, hallucinate methods or flood your servers with random calls.

So, it’s important to get it right. Thankfully, strategies are emerging to position APIs for AI agents, from new standards to underground tricks. And, it’s more than just “get an MCP server” (though that’s a crucial step).

The jury’s out on what strategy will be most effective. So I’ve structured this guide to start with broadly agreed-upon best practices, then explore ones still taking shape. Most tips apply equally to public, partner and private APIs."

https://thenewstack.io/how-to-prepare-your-api-for-ai-agents/

#AI #GenerativeAI #AIAgents #AgenticAI #LLMs #API #OpenAPI #RAGs #MCP

#AI #GenerativeAI #Journalism #News #LLMs #RAGs #Science #ScienceJournalism #Media: "Surprisingly, GPT-4 with the abstracts performs a little bit better than GPT-4 with RAG over the article text, both in terms of accuracy (96.6% vs. 93.5%) and win percentage (29.2% vs. 27.8% — the rest were ties). This suggests that more context from the scientific article did not necessarily lead to higher accuracy or better understandability (i.e. our first assumption about using RAG here as described above was not met). A deeper investigation of the retrieved snippets and their actual relevance to the jargon term may help understand if this is an issue with the quality of the context, or if there may be other causes such as the similarity threshold we used. To apply RAG successfully, it’s essential to explore and test different parameters. Without this kind of careful experimentation, RAG on its own might not provide the desired results.

It may also be the case that the large size of GPT-4’s pre-training dataset enables it to draw from other sources to generate definitions. This can be as much a concern as a benefit though — it can make it harder to override irrelevant information from the pre-training data, or eschew its limitations based on cutoff dates for model training.

We also found that the effectiveness of these approaches varied based on the reader’s expertise. For instance, the less experienced annotator found similar value in both methods (higher tie percentage), while the more expert reader noticed more differences. Further evaluation with a larger set of annotators may help to replicate and understand these differences."

https://generative-ai-newsroom.com/making-sense-of-science-llms-for-helping-reporters-understand-complex-research-f1292720a09c