Sparse Vector

How Dense, Sparse, and Hybrid Retrieval Work in RAG

Three ways a RAG system can generate and combine candidate passages.

Oct 16 • Rishabh Gupta

1

2

GitHub is Rebuilding Git for Agent Scale Development

The important change is not simply more generated code. Agent workloads reshape the read, write, and coordination patterns beneath software development.

Oct 9 • Rishabh Gupta

1

3

How Embeddings and Similarity Search Work in RAG

How text becomes a vector, how a query is compared, and what a similarity score can tell you.

Oct 7 • Rishabh Gupta

4

2

3

September 2026

Why RAG Systems Split Documents into Chunks

How text boundaries shape what retrieval can return.

Sep 30 • Rishabh Gupta

4

4

How RAG Works: From Document Indexing to Generated Answers

How documents become searchable context for a language model

Sep 23 • Rishabh Gupta

9

5

5

August 2026

What the Model Reveals, and What It Just Makes Up

Series finale — OWASP LLM07 & LLM09

Aug 13 • Rishabh Gupta

2

1

The Free Cody Proxy That Sourcegraph’s Own Traffic Spike Exposed

A case study in OWASP LLM10: Unbounded Consumption

Aug 12 • Rishabh Gupta

1

The LangChain Bug That Turned a Math Question Into Remote Code Execution

A case study in OWASP LLM05: Improper Output Handling

Aug 9 • Rishabh Gupta

1

AI Just Made 30 Years of DNA Evidence Vulnerable to Tampering

A fresh finding

Aug 5 • Rishabh Gupta

1

The First Documented Case of One AI Agent Attacking Another

A fresh finding — published today

Aug 4 • Rishabh Gupta

1

2

1

How a Model on Hugging Face Became a Reverse Shell

A case study in OWASP LLM03: Supply Chain

Aug 2 • Rishabh Gupta

1

1

July 2026

How Slack AI Could Leak Data With No Clever Attacker Required

OWASP LLM06 — Excessive Agency

Jul 29 • Rishabh Gupta

1

2

2