You’ve no doubt noticed that AI – artificial intelligence – has seeped into every aspect of business and personal life. CRT mentor Dan Gatti has served as CEO of three Silicon Valley technology companies, and is well-versed in this space. Therefore, to make your life easier we pressed him into service to provide a list of definitions of terms and concepts relevant to the study of AI. We hope it helps.
—————————-
GPT stands for Generative Pre-trained Transformer. It refers to a type of large language model developed by OpenAI that generates human-like text. The model is built on a deep learning architecture known as a Transformer, which is highly effective for processing sequences of data, such as language. GPT models are trained on vast amounts of text data and can generate, summarize, and understand natural language.
LLMs (Large Language Models) LLMs are advanced AI models trained on massive amounts of text data to understand and generate human-like language. They use deep learning techniques, particularly neural networks. Can handle diverse tasks like summarization, translation, and answering questions. The “large” in LLM refers to the size of the model (number of parameters) and the vast training data.
LLMs
| Name | Developer | Type | Relation to GPT |
| ChatGPT | OpenAI | GPT engine | Based on OpenAI’s GPT architecture. |
| Claude | Anthropic | LLM | Separate LLM, not based on GPT. |
| Gemini | Google DeepMind | LLM | Independent of GPT; Google’s alternative. |
| Perplexity | Perplexity.ai | Search assistant | Uses GPT models but is not a model itself. |
| LLaMA | Meta | LLM | Not GPT; Meta’s own LLM design. |
DeepSeek Hangzhou LLM Open AI using software logic
Natural Language Processing (NLP) is a field of artificial intelligence (AI) that focuses on the interaction between computers and human (natural) languages. It enables machines to understand, interpret, generate, and manipulate human language in a way that is both meaningful and useful. NLP combines computational linguistics with machine learning and deep learning techniques to process and analyze large amounts of natural language data.
AI Bots- AI bots are software applications designed to autonomously interact with users or other systems, typically through conversation or task execution.
AI Tools-AI tools are broader applications or platforms that use artificial intelligence to assist users in performing specific tasks or solving problems.
Inference-in AI refers to the process of using a trained model to make predictions or generate outputs based on new input data.
Chatbot A chatbot works by using artificial intelligence (AI) and natural language processing (NLP) to simulate conversations with users. The chatbot can understand, interpret, and respond to text or voice inputs, creating an interactive experience.
How AI agents work Autonomous agents are AI systems that can pursue a goal without needing humans to tell them what to do and when. If given an objective, they can create tasks, complete those tasks, and re-prioritize them until reaching the objective. Integral components of these systems include foundation models (for reasoning and analysis), tool use (to interact with the internet and other software or apps), and memory access
Technologies Behind Chatbots
- Machine Learning (ML)(see below) AI chatbots use ML algorithms to learn from user inputs, analyze large datasets, and improve their accuracy over time.
- Neural Networks: Some advanced chatbots use neural networks, particularly deep learning models, to understand more complex sentences, recognize patterns, and respond with more human-like accuracy.
- APIs: Chatbots often connect to external APIs (Application Programming Interfaces) to pull or send data (e.g., weather services, payment systems) to enrich the user experience.
Artificial General Intelligence (AGI)- refers to a form of AI capable of performing any intellectual task that a human can do, with flexibility across multiple domains. It mimics human cognitive abilities, such as reasoning, problem-solving, and learning in diverse contexts.
- Example: A “hypothetical” AI system capable of autonomously designing a rocket, diagnosing diseases, or writing poetry without being specifically trained in these areas.
Agentic AI-Agentic AI refers to AI systems that operate as autonomous agents with a degree of “agency.” This means they are designed to act toward specific goals or objectives in dynamic environments, often without direct human intervention.
- Example: A self-driving car navigating traffic to reach a destination or an AI assistant proactively managing a user’s schedule.
- Operator: ChatGPT feature that performs a system function such as selecting ingredients form a menu like linguine and clams ordering from store and delivering to your home.
Virtual Assistant (VA) is a software application or AI-powered system designed to perform tasks or provide services for individuals or businesses. Virtual assistants use natural language processing (NLP), machine learning, and other AI techniques to interact with users, understand their needs, and execute tasks through voice commands, text input, or pre-defined triggers.
General-Purpose Virtual Assistants
- Siri (Apple): Provides voice-activated assistance for tasks like sending messages, setting reminders, and controlling smart devices.
- Google Assistant (Google): Offers multi-platform assistance with tasks like navigation, managing calendars, or even translating languages.
- Alexa (Amazon): Operates Amazon’s smart ecosystem, answering questions, playing music, and controlling smart home devices.
Hallucinations in AI inference are a known challenge, particularly for generative models. By improving data quality, implementing verification processes, and enhancing model architectures, it is possible to reduce the frequency and severity of hallucinations. However, complete elimination remains an ongoing research focus, as AI systems continue to balance creativity with factual accuracy.
Virtual Reality
- Digital environment that replaces the real world
Augmented Reality
- Overlays Digital Elements onto the Real-World Environment
Computer Power
- Traditional-based on Binary-bits on or off
- Quantum-based on qubits bits being on and off simultaneously called Entanglement
Machine learning
This is the key subset of AI, and frankly, why we’ve gathered you here today. ML enables systems to find patterns, make predictions, and draw conclusions without explicit programming. An ML system can do its thing without its minders needing to hand-code every rule. With any luck, ML researchers train their algorithms, deploy them, and let the system “learn” from datasets. ML has many subsets:
- Supervised learning: You feed the algorithm loads of labeled training data, i.e. an image of a dog with the annotation “dog.” Supervised learning systems learn to associate inputs with the correct output.
- Unsupervised learning: You give the algorithm unstructured, unlabeled data and it makes inferences on its own. A common use case is clustering, where input data is divided into groups or patterns and rated on similarity. In more plain terms, this could help you find overlap between friends on a social graph.
- Semi-supervised learning: Semi-supervised learning systems can handle datasets that are noisy or missing many labels.
- Reinforcement learning: A program that learns by trial-and-error, and feedback, in dynamic environments. RL agents are commonly found acing multiplayer games or guiding robots.
- Transfer learning: A model trained for one task that is reused as a starting point for a separate task.
Dan Gatti is Managing Partner of Innovative Capital Ventures and serves as Executive Director of the Big Data IoT Forum and Executive Director Health Cloud Solutions, a forum for cloud- based solutions. Dan also serves as an Adjunct Professor at San Diego State University Graduate Program in Homeland Security. You can reach him at dan.gatti@icventures.net.