An agent is a computer program or system that is designed to perceive its environment, make decisions and take actions to achieve a specific goal or set of goals. The agent operates autonomously, meaning it is not directly controlled by a human operator.

The broad field of computer science focused on creating machines that can perform tasks that typically require human intelligence, such as learning, reasoning, and problem-solving.

Bias in AI systems refers to systematic, unintended favoritism or discrimination in an algorithm's outputs, often caused by imbalanced training data, flawed assumptions, or incomplete modeling. It results in outcomes that unfairly advantage or disadvantage certain groups or perspectives.

The rules and guardrails ensuring that the data your company inputs into an AI system isn't used to train public models or exposed to competitors.

Generative AI is an advanced technological approach that enables the creation of content including text, images, and videos. By analyzing and discerning patterns within extensive training datasets, generative AI can autonomously construct material that shares comparable characteristics to its training input. This capability stems from the AI’s understanding of data patterns and its ability to replicate or innovate based on these patterns.

GPT (Generative Pre-trained Transformer) Generative Pre-trained Transformers (GPT) are a type of advanced artificial intelligence model primarily used for natural language processing tasks. GPT models are based on the transformer architecture, which allows them to efficiently process and generate human-like text by learning from vast amounts of data. The “pre-trained” aspect refers to the initial extensive training these models undergo on large text corpora, allowing them to understand and predict language patterns. This pre-training equips the GPT models with a broad understanding of language, context, and aspects of world knowledge.

We call the occurrences where large language models generate factually inaccurate or illogical answers due to data and architecture constraints hallucinations.

A type of artificial intelligence that excels at processing, understanding, and generating human language.

These models are trained on vast amounts of text data and learn patterns and rules of language, allowing them to perform tasks like answering questions, summarizing text, and generating content.

A lightweight text formatting language using simple symbols (like # for headers, for bold, and - for bullets). LLMs love Markdown. It uses fewer tokens than HTML and helps the AI understand text hierarchy cleanly. You should always structure your system prompts using Markdown headings.

The phenomenon where a previously effective prompt becomes less reliable or accurate over time due to changes in the underlying AI model's behavior, data, or the complexity of the application it's used in.

Defined, code-based capabilities given to an AI so it can interact with the outside world. For example, If you want an agent to look up a flight, "Search_Web" or "Access_Amadeus_API" is a skill you explicitly attach to it.

The basic unit of data processed by an LLM. It’s usually a fragment of a word (roughly 4 characters or 0.75 words in English). Models have a Token Limit (or context window), which is the maximum amount of text they can read or write at one time.

The dataset used to teach an AI model how to recognize patterns, make predictions, or generate outputs.