David Alami·Jun 3Agentic AI Failure Series #1: The Infinite Loop as a Security RiskThis article is part of an ongoing series on agentic AI failure modes and their security implications.
David Alami·Apr 6, 2025Why Every AI Agent Needs a Tool Registry — And How to Build OneAI Agents are attracting an incredible amount of attention because of their ability to perform dynamic tasks on behalf of users. The idea…
David Alami·Mar 28, 2025Model Context Protocol — What, Why and HowIn today’s rapidly evolving technological landscape, the way we interact with software is being redefined by modern Large Language Models…
David Alami·Mar 26, 2025How to Teach an LLM Without Buying It — Part 2: Evolutionary AlgorithmsA black-box LLM is one where the user sees only the input/output interface but has no direct knowledge of (or access to) the underlying…
David Alami·Mar 25, 2025How to Teach an LLM Without Buying It — Part 1: Meta-PromptingIn recent years, the remarkable progress of large language models (LLMs) has led to a wave of new techniques for getting these models to…
David Alami·Mar 24, 2025Mitigating LLM Hallucinations: Reliability and SecurityHallucinations in large language models (LLMs) have become a central concern for researchers and practitioners striving to build reliable…
David Alami·Apr 4, 2023Uncovering OpenAI’s Data Sources: From Crowdsourcing to CollaborationOpenAI, one of the leading research institutes in artificial intelligence (AI), relies on a diverse range of sources to obtain datasets…
David Alami·Nov 1, 2022Violent NLP — Part 3.Today I will show you how to big a targeted wordlist basing on a huge password leak dump — rockyou2021. There is a lot of ML and Big Data…
David Alami·Oct 8, 2022Violent NLP — Part 2. Patterns in passwords of Russian-speaking Internet usersThis article continues a series on applications of natural language processing in the area of information security and penetration testing…
David Alami·Oct 6, 2022Violent NLP — part 1.This article begins a series on applications of natural language processing in the area of information security and penetration testing…