AI Glossary

• Agents – a type of advanced AI that needs only be given a desired outcome, then it autonomously finds its own path and intermediate steps to accomplish the goal.
• AGI (Artificial General Intelligence) – a theoretical type of AI that has abilities indistinguishable from a human, and hypothetically can pass the Turing test.
• Chatbot – this term is sometimes used as a synonym for public LLMs, but can also be used to refer to a custom-trained LLM generated by individual users, not the large company like OpenAI.
• GenAI (Generative AI) – a type of AI that “generates” an output, such as text or images. Large language models like ChatGPT are generative AI
• Generative Inbreeding – this term refers to AI being trained on previously-generated AI output, which may contain biases and hallucinations. These incorrect facts may be considered correct if they are prevalent enough, and the model becomes less reliable over time.
• LLM (Large Language Model) – a type of software / generative AI that accesses large databases it’s been trained on to predict the next logical word in a sentence, given the task/question it’s been given. Advanced models have excellent “perplexity” (plausibility in the word choice) and “burstiness” (variation of the sentences).
• Reasoning Models – advanced versions of LLMs that intentionally take longer to answer in order to provide better-reasoned answers. Ideal for knotty/thorny problems.
• Superintelligence – a hypothetical type of AI that far exceeds human intelligence. Sci-Fi has postulated such instances as Skynet or Ultron.