The article discusses the fundamental concept of AI, which revolves around finding the best-fitting line for a set of data points. It explores how various AI models, including LLMs and prediction models, rely on this principle. Th...
About:
Director of AI at ddx, with a passion for programming and entrepreneurship, and a background in software architecture and management.
Website:
Specializations:
Interests:
Outgoing Links:
Subscribe to RSS:
The article discusses the challenges of chunking in Retrieval-Augmented Generation (RAG) systems, particularly in isolating relevant information from source documents while maintaining context. It highlights the issues of creating...
The author offers professional coding services to help businesses improve their MVPs (Minimum Viable Products) that were quickly developed using LLM-based coding tools. They address common issues such as security vulnerabilities, ...
The article explains the concept of embeddings in AI, using dogs and books as examples. It discusses the problems with classifying books using word counts and the solution of using embeddings. It also explains how Word2Vec and Rec...
The article discusses the challenges of developing a chatbot for an online crime game that must handle real-time, interruptible conversations. It describes the techniques used to ensure the chatbot can handle quick follow-ups, cla...
The article discusses the benefits of using in-memory free-text search (IMFTS) to improve the performance of LLM-driven NPCs in a crime-based online game. It explains how IMFTS helps in reducing inference calls, improving answer q...
The text advises against using large language models (LLMs) for core application logic, such as decision-making and executing business logic, due to their limitations in performance, debugging, and other areas. It suggests using L...
The text discusses the different definitions of 'Agentic AI' and the importance of having a shared understanding of the term. It presents four definitions, including marketing, overly simplistic, broad agency, and narrow agency, a...
The article discusses the challenges of using LLM-based NPCs in a constrained-reality online game and the role of chain-of-thought prompting in keeping LLM-generated responses accurate. The author explains the process of using CoT...