Explore definitions and explanations of essential AI and technology concepts.

Solving tasks with no or a handful of examples embedded in the prompt (in-context learning).

On-body devices (glasses, rings, earbuds) that can host/mediate AI and help make computing ambient.

A storage/indexing system for embeddings to enable fast semantic search and retrieval-augmented generation.

• Guide for Vectorizing Content for Internal Search & Analytics

Ubiquitous computing is a new paradigm where intelligent systems are seamlessly embedded into everyday environments, making technology effectively invisible while continuously supporting human activity. Unlike traditional computing that requires explicit prompts and fragmented devices, ubiquitous computing leverages context-awareness, memory, and integration to deliver adaptive, anticipatory, and user-centric experiences.

A decoding setting controlling randomness/creativity; higher values yield more diverse outputs.

A neural architecture built on attention mechanisms; foundation of modern LLMs.

The unit of text (or audio/video chunk) a model processes; counts affect cost/latency.

AI-generated data used to supplement scarce or sensitive real data (with care for drift and artifacts).

Supervised learning stage using instruction–response pairs to teach task formats and tone.

Empirical relationships between data/parameters/compute and model performance that inform roadmap decisions.

Policies, controls, and lifecycle practices that mitigate misuse, harm, and failure modes (safety filters, audits, governance).

Fine-tuning with human preference data to align outputs with desired style/behavior.

The instruction/context you give a model for it to complete a task; includes system prompts, examples, constraints, and tool schemas.

Fetching relevant context (via search/vector DB) and injecting it into the prompt to ground responses in current or proprietary data.

Initial large-scale training on broad data before specialized tuning (SFT, RLHF).

Learned numeric values inside a model (billions+ for frontier LLMs) that encode patterns from training data.

Mechanisms that let models call tools (search, databases, emails, code) and stitch steps together into workflows—key for agents. Agents will need to conduct widespread orchestration across different websites, apps, and platforms.

For example, for an agent to purchase a plane ticket, it needs to navigate across many channels to search for the ticket, select the seat, add to cart, etc.

A layered function approximator trained to minimize loss via gradient descent. Breakthroughs in AI stem from this idea of a neural network where machines can make connections and think in the same way a human brain functions, rather than having to follow basic commands that tell it step-by-step, right from wrong.

A transformer-based model trained on massive text corpora to generate/analyze language; increasingly multimodal. Ex: ChatGPT, Claude, Gemini, etc.

• Guide for Structuring Content for Optimal LLM / Agentic Consumption

A model that consumes and/or produces multiple data types (text, images, audio, video). ChatGPT is multimodal because it can provide text and images to you.

Numeric parameters inside a neural network that encode what the model has learned; 'open-weight' models expose weights for local use/modification; 'closed-weight' do not.

Dynamically choosing the best model for a given prompt based on predicted quality/cost/latency—often a meta-model trained on evaluation data.

ChatGPT now has a built-in model router that takes your question and determines which of its LLMs are needed to answer the prompt based on its complexity, desired outcome, etc. This helps optimize for the most efficient outcome vs latency, cost, etc.

Round-trip time from request to response; an experience bottleneck for agents and wearables.

Running a trained model to generate outputs (production use). The costs and latency of inference shape the 'unmetered intelligence' thesis. In simple terms, the cost to generate outputs using AI is rapidly declining, leading to a world of abundant knowledge.

Confident but false output from a model; mitigated by retrieval, calibration, routing, and post-processing. The root cause of much of the fear of AI.

When you ask ChatGPT about yourself and it provides incorrect details about where you grew up or went to school.

The latest point in time included in a model’s pretraining data. The model lacks builtin knowledge from after this cutoff, unless updated or retrieved.

A large, broad-training model adaptable to many downstream tasks (e.g., LLMs, multimodal models). Ex: GPT family, Claude, Mistral, Gemini.

Additional training on a narrower dataset to specialize a base model (via SFT, LoRA, etc.). Leveraging your own data to make a large model better (i.e. training a model with your company’s data so it can better respond to queries regarding certain processes or customers).

Methods to make model behavior understandable (feature attributions, summaries, rule extraction); emphasized as a solution to AI ethical concerns.

A claim that traditional schooling structures (pace, assessment, curricula) must be reimagined for an AI-rich world that values creativity, curiosity, and agency over memorization.

A numeric vector representation of text, images, or other data that captures semantic meaning for search, clustering, and retrieval-augmented generation.

A constrained, user-supervised assistant embedded in apps (code, docs, CRM).

The aggregate processing power used to train or run models (GPUs/TPUs, memory bandwidth, networking). A central accelerator of AI capability growth.

While the cost of compute to train models is still quite high (and the reason new developments in AI are so expensive), the inference cost to use AI is dropping significantly. This means the cost to actually interact with the pre-existing models is dropping, leading to expanded model access.

A prompting/analysis style where a model reasons step-by-step. Powerful, but often hidden from end users in production for safety/privacy.

A model whose internal reasoning is opaque to users; creates a desire for explainable AI requirements in regulated or high-stakes contexts.

Systematic errors that produce unfair outcomes (e.g., by gender, race). In this framing, explainability and transparency are prioritized alongside bias mitigation.

Note that non-algorithimic bias is not bad and actually is required for AI to function.

Hardware and software that read/write brain signals to control devices or restore functions; used as a horizon example of “computing beyond screens.” NeuraLink is Musk’s company working on BCI.

• TNR Newsletter: AI-Powered Accessibility: The End of Human Limitations

Structured tests (tasks, datasets, rubrics) used to compare models or track progress; crucial for routing decisions, safety, and QA.

Stand alone criminals not empowered by a larger organization. AI risks the empowerment of low resource bad actors, giving “small scale” criminals increased resources and capability to create far more harm.

Nation states with lots of power and Bond villains)

Research and practices to ensure AI systems pursue goals consistent with human values and constraints (safety, reliability, controllability). Alignment means the AI working on behalf of goals for human betterment.

Teens wanting to hurt themselves used to be able to Google practical advice for causing physical harm. If you ask ChatGPT this, it will provide resources as to why you should not hurt yourself, representing alignment.

People committing crimes with malintention that are harmful to society.

Software that plans and executes multi-step tasks toward a goal with varying autonomy (tool use, memory, self-correction). Distinct from “copilots,” which remain user-in-the-loop. Agentic AI is presented in TNR and TNGC keynotes as the second part of Zack's “Phases of AI Integration”.

Sample example data for more than one example development
foofoofoofoo

Another sample example for the dataset for development

• TNR Newsletter: AI Agents and The End of Shopping as We Know It
• TNR Newsletter: Prepare for an Entirely New Agentic Internet

The programmatic interface apps and services used to talk to AI models and tools (e.g., for retrieval, function calls, or routing).

Think of API as a set of instructions for one piece of software to talk to another. Developers can invoke OpenAI’s API as a step within their pipeline.

The core operation in transformer models that lets them weigh relationships between tokens to predict the next token.

Likely brought up in reference to the research paper “Attention is All You Need” which introduced the transformer model, an integral part of the development of LLMs.

A hypothetical AI system that can understand, learn, and apply knowledge across most tasks at or beyond human level. In the book’s framing, AGI is expected around 2030 and functions as the inflection behind the “Next Renaissance.”