America’s mega-tech firms and their destruction of the USA’s internal competitiveness and talent development has consequences.

archive.ph/2025.08.21-2...

🧵 AGENTIC AI

"Agentic AI" is a class of artificial intelligence that focuses on autonomous systems that can make decisions and perform tasks WITHOUT human intervention.

Quite literally:

The AI is an "agent" that acts for you.

#LLMZoomcamp 🧵

"Our best guess is that SQLite is the second mostly widely deployed software library, after libz."

#LLMZoomcamp

www.sqlite.org/mostdeployed...

DuckDB has 0 external dependencies, which is fairly cool.

#LLMZoomcamp

duckdb.org/why_duckdb

Here's the dlt "quickstart" if you're trying to find it

#LLMZoomcamp

dlthub.com/docs/tutoria...

Embeddings =

Turning non-numerical data into numerical data, while preserving meaning and context. Similar non-numerical data, when entered into an embedding algorithm, should produce similar numerical data.

#LLMZoomcamp

Embedding -

A representation learning technique that maps complex, high-dimensional data into a lower-dimensional vector space of numerical vectors.

Meaningful patterns or relationships are preserved.

See, also: dimensionality reduction.

en.wikipedia.org/wiki/Embeddi...

#LLMZoomcamp

There is current an explosion of vector search tools, as people seek to capitalize on the current absolutely insatiable appetite for LLM software.

#LLMZoomcamp

Follow-up to previous thread with some background links about vector search and semantic similarity.

You use vector search tools in order to find documents with similar semantics. i.e. Use vector search to find information you want.

#LLMZoomcamp

Semantic Similarity (DEF) -

Ametric defined over a set of documents or terms, where the idea of distance between items is based on the likeness of their meaning or semantic content as opposed to lexicographical similarity.

en.wikipedia.org/wiki/Semanti...

#LLMZoomcamp

LLMs are rapidly advancing in capabilities, and foundational LLMs can be connected to knowledge bases that can provide more specialized context to user queries.

In the coming weeks, we will be talking a lot about:
1. Retrieval-augmented Generation
2. Model Context Protocol Servers

#LLMZoomcamp

Language models aren't new, but since the release of ChatGPT in 2022, LARGE language models have dominated the conversation.

What sets these models apart is the LARGE number of parameters they use and the LARGE amounts of data they've been trained on.

#LLMZoomcamp

en.wikipedia.org/wiki/Large_l...

Converting words to numerical representations allows us to take advantage of vector spaces.

#LLMZoomcamp

en.wikipedia.org/wiki/Vector_...

Tokenization = first & last step of text processing & modeling. Text is broken into "tokens" and each token is assigned a numerical representation, or "index", which can be used to feed into a model.

#LLMZoomcamp

docs.mistral.ai/guides/token...

#LLMZoomcamp

github.com/DataTalksClu...