AI is remaking the world—and, in the process, inventing an entirely new lexicon to describe it. Sit in on any product meeting, pitch, or panel today, and you'll hear people toss around terms like LLMs, RAG, and RLHF. And as of this week, there's "opaque recurrence," the reasoning technique in OpenAI's new Astra model that has AI safety researchers rattled. The vocabulary is moving fast enough to make even seasoned tech insiders feel a bit out of their depth.
This glossary is our attempt to help. We offer plain-English definitions of the AI terms you're most likely to encounter—whether you're building with these systems, investing in them, or simply keeping up with TechCrunch and industry podcasts. We update this list regularly as the field evolves, so consider it a living document, much like the AI it describes.
AGI
Artificial general intelligence (AGI) is a slippery concept, but it generally refers to AI that outperforms an average human across most—if not all—cognitive tasks. OpenAI CEO Sam Altman once described AGI as the “equivalent of a median human that you could hire as a co-worker.” Meanwhile, OpenAI's charter defines it as “highly autonomous systems that outperform humans at most economically valuable work.” Google DeepMind offers a slightly different take, viewing AGI as “AI that's at least as capable as humans at most cognitive tasks.” Confused? Don't worry—even AI researchers aren't sure.
AI Agent
An AI agent is a tool that uses AI to perform a series of tasks on your behalf—going beyond what a basic chatbot can do—such as filing expenses, booking tickets or a restaurant table, or even writing and maintaining code. However, as we've noted before, this emerging space is full of moving parts, so “AI agent” can mean different things to different people. The infrastructure to deliver on its full promise is still developing. At its core, though, an agent implies an autonomous system that may draw on multiple AI models to complete multi-step tasks.
API Endpoints
Think of API endpoints as “buttons” on the back of software that other programs can press to make it act. Developers use these interfaces to build integrations—like letting one app pull data from another, or enabling an AI agent to control third-party services directly without a human clicking through each step. Most smart home devices and connected platforms have these hidden buttons available, even if end users never see them. As AI agents become more capable in 2026, they're increasingly adept at finding and using these endpoints on their own, unlocking powerful—and sometimes surprising—automation possibilities.
Chain of Thought
Given a simple question, a human brain can answer without much deliberation—like “Which animal is taller, a giraffe or a cat?” But for more complex problems, you often need pen and paper to work through intermediate steps. For instance, if a farmer has chickens and cows with a total of 40 heads and 120 legs, you'd jot down a few equations to solve it (2 * chickens + 4 * cows = 120, chickens + cows = 40). AI’s chain of thought mimics this process: it breaks problems into logical steps before arriving at an answer.
via TechCrunch AI