Verified AI announcements and plain-English explainers, without the noise.https://whatisai.fyihttps://whatisai.fyi/a2a-protocolhttps://whatisai.fyi/a2a-protocolA2A is an open protocol for AI agents: one agent can find another, hand it a task and track the result, even when different companies built them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/accuracyhttps://whatisai.fyi/accuracyAccuracy is the share of a model's predictions that were right, the correct answers divided by all answers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/activation-functionhttps://whatisai.fyi/activation-functionAn activation function is the non-linear step a neuron applies to its weighted sum, so a network of layers can learn curved patterns, not only straight lines.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-evaluationhttps://whatisai.fyi/agent-evaluationAgent evaluation tests an AI agent on set tasks, checking both its final result and the steps it took, over several tries.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-memoryhttps://whatisai.fyi/agent-memoryAgent memory is how an AI agent saves useful information outside the model and loads the right pieces back into its prompt later.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-orchestrationhttps://whatisai.fyi/agent-orchestrationAgent orchestration is how an app with several AI agents decides which agent works, in what order, and who picks what happens next.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-planninghttps://whatisai.fyi/agent-planningAgent planning is how an AI agent splits a big task into smaller steps first, then carries out those steps one by one.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-statehttps://whatisai.fyi/agent-stateAgent state is the running record an AI agent keeps while it works, so a task can pause, resume, or recover from a crash.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agentic-modelshttps://whatisai.fyi/agentic-modelsAn agent plans, acts, observes the result, adjusts its approach, and repeats until the task is complete.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agentic-workflowshttps://whatisai.fyi/agentic-workflowsAn agentic workflow runs a task as several language model and tool steps, joined by a path that your code fixes in advance.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agihttps://whatisai.fyi/agiAGI is a proposed category for AI at least as capable as a person across most thinking tasks. Definitions differ; a 2023 Google DeepMind framework marked the level matching many definitions as not yet achieved.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agnohttps://whatisai.fyi/agnoAgno is open-source software for building AI agents and running them as a live service. You can build single agents, teams of agents and step-by-step workflows.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ai-acceleratorshttps://whatisai.fyi/ai-acceleratorsAI accelerators are chips built to speed up the maths inside AI models, which is largely multiply-add sums.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ai-agentshttps://whatisai.fyi/ai-agentsAn AI agent is a language model that chooses its own next steps and tools while it works through a task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ai-modelhttps://whatisai.fyi/ai-modelAn AI model is the part of an AI system that has learned from data. It is a structure plus a set of learned numbers that turns an input into a prediction or new content.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/algorithmhttps://whatisai.fyi/algorithmAn algorithm is a precise, step-by-step procedure for getting a result. In AI, learning algorithms are the procedures that turn data into a trained model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/amazon-bedrockhttps://whatisai.fyi/amazon-bedrockAmazon Bedrock is an AWS service that lets developers call AI models from several companies through one set of APIs, with extra tools for search and safety.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/amazon-bedrock-agentcorehttps://whatisai.fyi/amazon-bedrock-agentcoreA set of AWS services, called Amazon Bedrock AgentCore, for running AI agents in the cloud, with hosting, memory, sign-in, tool access and monitoring.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/amazon-sagemaker-aihttps://whatisai.fyi/amazon-sagemaker-aiSageMaker AI, from AWS, lets you build, train and run machine learning models without managing your own servers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/anomaly-detectionhttps://whatisai.fyi/anomaly-detectionAnomaly detection finds the few items or events that do not fit the usual pattern in data, such as a fraudulent card payment or an odd burst of network traffic.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/approximate-nearest-neighbor-searchhttps://whatisai.fyi/approximate-nearest-neighbor-searchApproximate nearest neighbor search finds stored vectors that are very close to a query quickly, by checking only a promising part of the collection and accepting a few misses.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/arc-agihttps://whatisai.fyi/arc-agiARC-AGI presents demonstration grid pairs to a system, then asks it to construct outputs for new test grids.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/artificial-intelligencehttps://whatisai.fyi/artificial-intelligenceAn AI system infers from received inputs how to generate predictions, content, recommendations or decisions for explicit or implicit objectives.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/attentionhttps://whatisai.fyi/attentionAttention lets a model build an output by assigning weights to available pieces of information and mixing their values.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/autoencodershttps://whatisai.fyi/autoencodersAn autoencoder learns an encoder that maps an input to a code and a decoder that reconstructs the input from that code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/autogenhttps://whatisai.fyi/autogenAutoGen is an open-source Microsoft framework for building apps where several AI agents talk to each other, and sometimes to people, to finish a task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/autoregressive-modelshttps://whatisai.fyi/autoregressive-modelsAn autoregressive model factorizes a sequence probability into conditional probabilities.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/aws-inferentiahttps://whatisai.fyi/aws-inferentiaAWS Inferentia is a family of computer chips that Amazon designed to run trained AI models cheaply and quickly in its cloud.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/aws-trainiumhttps://whatisai.fyi/aws-trainiumAWS Trainium is a computer chip Amazon designed for training and running AI models, rented through special Amazon EC2 cloud servers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/azure-openaihttps://whatisai.fyi/azure-openaiAzure OpenAI lets companies use OpenAI's models through Microsoft's Azure cloud, with Azure billing, safety filters and data controls.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/backpropagationhttps://whatisai.fyi/backpropagationBackpropagation computes each parameter's gradient by applying the chain rule from the last layer back to the first.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/base-modelshttps://whatisai.fyi/base-modelsA base language model is a pretrained model before further instruction or conversational fine-tuning.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/batch-inferencehttps://whatisai.fyi/batch-inferenceBatch inference sends many AI requests as one job that runs in the background, trading an instant reply for a lower price.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/batch-sizehttps://whatisai.fyi/batch-sizeBatch size is how many training examples are grouped before the weights change once.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/bayesian-inferencehttps://whatisai.fyi/bayesian-inferenceBayesian inference treats an unknown number as uncertain, starts from a belief about it, and uses Bayes' theorem to update that belief as data arrives.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/beam-searchhttps://whatisai.fyi/beam-searchBeam search keeps a fixed number of high-scoring partial sequences, expands them, and retains the best candidates at each step.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/benchmark-contaminationhttps://whatisai.fyi/benchmark-contaminationBenchmark contamination happens when test questions, or close copies of them, end up in a model's training data, so its score can look better than its real skill.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/benchmarkshttps://whatisai.fyi/benchmarksA benchmark is a fixed set of test tasks with a scoring rule, so different AI models can be compared on the same exam.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/bias-variance-tradeoffhttps://whatisai.fyi/bias-variance-tradeoffBias is average error across training sets, while variance indicates sensitivity to varying training sets.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/byte-pair-encodinghttps://whatisai.fyi/byte-pair-encodingByte-pair encoding builds a subword vocabulary by repeatedly merging the most frequent adjacent symbol pair.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/chain-of-thoughthttps://whatisai.fyi/chain-of-thoughtChain-of-thought prompting asks for or demonstrates intermediate reasoning text before a final answer, but that text is not a guaranteed view of hidden model internals.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/chat-templateshttps://whatisai.fyi/chat-templatesA chat template converts a structured list of messages into a formatted sequence.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/chatgpthttps://whatisai.fyi/chatgptChatGPT is OpenAI’s conversational product for asking questions, creating things and completing work with models and tools.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/chinchilla-scalinghttps://whatisai.fyi/chinchilla-scalingChinchilla scaling allocates a fixed training-compute budget by growing model parameters and training tokens in roughly equal proportions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/chunkinghttps://whatisai.fyi/chunkingChunking splits a long document into smaller pieces that can be embedded, searched, or fitted into a model prompt.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/claude-agent-sdkhttps://whatisai.fyi/claude-agent-sdkAnthropic's Claude Agent SDK is a Python and TypeScript library for building agents that run on the same loop, tools and context handling as Claude Code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/claude-apphttps://whatisai.fyi/claude-appClaude is Anthropic’s AI product for conversation, analysis and completing tasks with files, connected tools and editable outputs.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/claude-codehttps://whatisai.fyi/claude-codeClaude Code is Anthropic's coding assistant that reads a project, edits files and runs commands for you, asking before risky steps.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/closed-modelshttps://whatisai.fyi/closed-modelsA fully closed model keeps its weights and code proprietary for internal use.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/clusteringhttps://whatisai.fyi/clusteringClustering sorts data that has no labels into groups of items that resemble each other.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/code-interpretershttps://whatisai.fyi/code-interpretersA code interpreter lets an AI model write a small program, run it in a sealed-off sandbox, and use the result in its answer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/code-modelshttps://whatisai.fyi/code-modelsCode Llama is a family of language models for code with infilling and instruction-following capabilities.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/collaborative-filteringhttps://whatisai.fyi/collaborative-filteringCollaborative filtering recommends items by learning from what many people rated, clicked or watched, then predicting the gaps in your own history.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/comfyuihttps://whatisai.fyi/comfyuiComfyUI is a free, open-source app where you build AI image, video and audio generators by wiring boxes called nodes together.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/computer-use-modelshttps://whatisai.fyi/computer-use-modelsA computer-use model reads a screen and emits mouse or keyboard actions that an application can execute.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/confusion-matrixhttps://whatisai.fyi/confusion-matrixA confusion matrix is a table that counts a classifier's predictions by what it predicted and what was really true, so you see which mistakes it makes.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/constitutional-aihttps://whatisai.fyi/constitutional-aiConstitutional AI trains and evaluates model behavior against an explicit written set of principles.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/context-compactionhttps://whatisai.fyi/context-compactionContext compaction swaps the older part of a long AI conversation for a short summary, so the chat can keep going inside the model's memory limit.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/context-engineeringhttps://whatisai.fyi/context-engineeringContext engineering is choosing, trimming and updating the tokens a model sees at each step, so an AI agent works from a small, useful context.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/context-windowhttps://whatisai.fyi/context-windowA context window is all the text a language model can reference while generating a response, including the response itself.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/contextual-retrievalhttps://whatisai.fyi/contextual-retrievalContextual retrieval has a model write a short note that places each chunk in its document, then indexes the note with the chunk so search can find it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/continual-learninghttps://whatisai.fyi/continual-learningContinual learning trains a model on a stream of changing data while measuring whether later learning damages or helps earlier tasks.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/convolutional-neural-networkshttps://whatisai.fyi/convolutional-neural-networksA convolutional neural network uses sparse convolutions that reuse the same weights at multiple locations in an ordered grid.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/cosine-similarityhttps://whatisai.fyi/cosine-similarityCosine similarity scores how closely two vectors point the same way, ignoring how long they are. The score runs from −1 to 1.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/cpushttps://whatisai.fyi/cpusA CPU is the chip that runs a computer's software, carrying out instructions with just a few cores backed by lots of cache memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/crewaihttps://whatisai.fyi/crewaiCrewAI is an open-source toolkit, written in Python, for building teams of AI agents, each with a role, that work through a list of tasks together.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/cursorhttps://whatisai.fyi/cursorCursor is a coding tool with an AI agent that can read your project, edit files and run commands for you.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/data-augmentationhttps://whatisai.fyi/data-augmentationData augmentation makes extra training examples by changing existing ones in small ways that keep the answer the same, such as flipping or cropping a photo.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/data-labelinghttps://whatisai.fyi/data-labelingData labeling attaches the answer a supervised model should learn to raw examples such as images, text, audio or table rows.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/data-leakagehttps://whatisai.fyi/data-leakageData leakage is when a model learns from information it will not have when making real predictions, so its test score looks better than it really is.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/data-parallelismhttps://whatisai.fyi/data-parallelismData parallelism runs replicas of one model on different slices of a batch, then combines their gradients before a synchronized update.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/data-privacyhttps://whatisai.fyi/data-privacyData privacy is about what an AI company may do with your chats, such as training models, and the settings that let you limit it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/decoder-only-modelshttps://whatisai.fyi/decoder-only-modelsA decoder-only model predicts each next token from the tokens that came before it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/deep-learninghttps://whatisai.fyi/deep-learningDeep learning is machine learning with neural networks that stack many layers, so each layer can build a more abstract picture of the data than the one before.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/diffusion-modelshttps://whatisai.fyi/diffusion-modelsA diffusion model learns to reverse a process that gradually adds noise to data.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/dot-producthttps://whatisai.fyi/dot-productThe dot product multiplies two equal-length lists of numbers position by position and adds the results into one number set by the angle between them and their lengths.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/dpohttps://whatisai.fyi/dpoDPO teaches a language model to prefer the better of two answers by learning from the comparison itself, with no separate reward model trained first.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/dropouthttps://whatisai.fyi/dropoutDuring training, dropout randomly replaces some input tensor elements with zero.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/dspyhttps://whatisai.fyi/dspyDSPy is a Python framework where you describe an AI job by naming its inputs and outputs, and it tunes the prompts for you.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/embedding-layerhttps://whatisai.fyi/embedding-layerAn embedding layer turns integer indexes into dense vectors of fixed size.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/embedding-modelshttps://whatisai.fyi/embedding-modelsAn embedding model maps an input to a numeric vector whose distance from another vector can measure relatedness.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/embeddingshttps://whatisai.fyi/embeddingsAn embedding maps a discrete item such as a token ID to a dense vector of numbers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/emergent-abilitieshttps://whatisai.fyi/emergent-abilitiesEmergent abilities are task capabilities that appear absent in smaller models but present at a larger scale.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/encoder-decoderhttps://whatisai.fyi/encoder-decoderAn encoder-decoder model combines an encoder with an autoregressive decoder for sequence generation.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/encoder-only-modelshttps://whatisai.fyi/encoder-only-modelsAn encoder-only model uses bidirectional self-attention to represent a supplied sequence.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ensemble-learninghttps://whatisai.fyi/ensemble-learningEnsemble learning combines the predictions of several models, usually by a vote or an average, so that their separate mistakes partly cancel out.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/epochhttps://whatisai.fyi/epochAn epoch is one complete pass through every example used to train a model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/evalshttps://whatisai.fyi/evalsAn eval checks an AI system's work. You feed it an input and use a set of rules to score what comes back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/expert-systemshttps://whatisai.fyi/expert-systemsAn expert system stores a specialist's know-how as if-then rules and uses a separate inference engine to apply those rules to a new case and explain its advice.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/f1-scorehttps://whatisai.fyi/f1-scoreThe F1 score rolls a classifier's precision and recall into one number from 0 to 1, and it stays low if either of them is low.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/feature-engineeringhttps://whatisai.fyi/feature-engineeringFeature engineering turns raw data, like prices and colour names, into the lists of numbers a model can learn from.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/feature-importancehttps://whatisai.fyi/feature-importanceFeature importance gives each input column a score for how much a trained model relies on it, so you can see what drives its predictions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/featureshttps://whatisai.fyi/featuresFeatures are the input facts a machine learning model reads about each example, such as a car's mileage or colour, turned into a list of numbers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/few-shot-learninghttps://whatisai.fyi/few-shot-learningFew-shot learning is the ability to perform a task from only a small number of labelled examples.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/few-shot-promptinghttps://whatisai.fyi/few-shot-promptingFew-shot prompting places a small set of example inputs and outputs in the prompt so the model can imitate the task on a new input.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/fine-tuninghttps://whatisai.fyi/fine-tuningFine-tuning takes a model that is already trained and keeps training it on a smaller set of examples for one task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/flashattentionhttps://whatisai.fyi/flashattentionFlashAttention computes exact attention with tiling that reduces reads and writes between GPU memory levels.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/flopshttps://whatisai.fyi/flopsFLOPs count the small arithmetic steps, like one multiply or one add on decimal numbers, that a computer does to train an AI model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/foundation-modelshttps://whatisai.fyi/foundation-modelsA foundation model is trained on broad data and can be adapted for many downstream tasks.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/frontier-modelshttps://whatisai.fyi/frontier-modelsFrontier models are highly capable general-purpose AI models at or beyond the capabilities of the most advanced current models.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/function-callinghttps://whatisai.fyi/function-callingFunction calling lets a model request a named function by returning structured arguments that an application, or the provider, then runs.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ganshttps://whatisai.fyi/gansA GAN trains a generator to make samples and a discriminator to distinguish generated samples from training data.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gemini-apphttps://whatisai.fyi/gemini-appGemini is Google’s AI app: you ask questions, upload files for answers and summaries, and can have it research a topic across many sources.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/generative-aihttps://whatisai.fyi/generative-aiGenerative AI learns patterns from existing data, then uses those patterns to create new text, images, audio, video, code or other content.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/genetic-algorithmshttps://whatisai.fyi/genetic-algorithmsA genetic algorithm searches for good answers the way breeding does. It keeps a population of candidates, mixes the fitter ones and repeats over many generations.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/github-copilothttps://whatisai.fyi/github-copilotGitHub Copilot is an AI helper for programmers that offers the next lines of code while you write, explains code when you ask, and can take on jobs you give it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/google-adkhttps://whatisai.fyi/google-adkGoogle ADK is an open-source toolkit from Google for writing AI agents in code, giving them tools, a record of each chat and helper agents.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gpqahttps://whatisai.fyi/gpqaGPQA is a 448-question multiple-choice benchmark written by experts in biology, physics and chemistry.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gpushttps://whatisai.fyi/gpusA GPU is a chip built to run many similar calculations at the same time, which suits the matrix maths inside neural networks.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gradient-boostinghttps://whatisai.fyi/gradient-boostingGradient boosting builds one strong model from many small decision trees, added one at a time, each trained to fix the errors the trees before it still make.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gradient-checkpointinghttps://whatisai.fyi/gradient-checkpointingGradient checkpointing stores fewer activations by recomputing them when needed, exchanging extra computation for lower memory use.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gradient-descenthttps://whatisai.fyi/gradient-descentGradient descent repeatedly measures how loss changes with each parameter and moves the parameters a small step toward lower loss.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/graph-neural-networkshttps://whatisai.fyi/graph-neural-networksA graph neural network updates each node's hidden state based on messages.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/graphraghttps://whatisai.fyi/graphragGraphRAG turns your documents into a map of people, places and links, groups and summarises it, then answers questions from that map.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/greedy-decodinghttps://whatisai.fyi/greedy-decodingGreedy decoding generates a sequence by choosing the most likely next token at every step.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/grok-apphttps://whatisai.fyi/grok-appGrok is the assistant from SpaceXAI (formerly xAI) for conversation, web and X search, file analysis, voice and media creation.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ground-truthhttps://whatisai.fyi/ground-truthGround truth is the best available account of what really happened, the reality that a model's predictions are checked against.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/groundinghttps://whatisai.fyi/groundingGrounding gives a model relevant source material and asks it to base its response on that evidence.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/grouped-query-attentionhttps://whatisai.fyi/grouped-query-attentionGrouped-query attention lets each group of query heads share one key head and one value head.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gruhttps://whatisai.fyi/gruA GRU is a recurrent unit with reset and update gates.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/guardrailshttps://whatisai.fyi/guardrailsGuardrails are checks placed around an AI model that inspect what goes in and what comes out, and can block, change or check inputs and replies that are unsafe, off-topic or break the app's rules.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hallucinationhttps://whatisai.fyi/hallucinationA hallucination is generated content that sounds plausible but is false, unsupported by evidence, or inconsistent with the given input.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/haystackhttps://whatisai.fyi/haystackAn open-source Python toolkit from deepset that builds AI apps (search, RAG, agents) by linking small parts into a pipeline.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hidden-markov-modelshttps://whatisai.fyi/hidden-markov-modelsA hidden Markov model infers a sequence of states you cannot see from a sequence of observations you can, using the odds of each state following another and of each state producing each observation.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hnswhttps://whatisai.fyi/hnswHNSW finds the stored vectors closest to a query quickly by searching a stack of linked graphs, from a sparse top layer down to a dense bottom one.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hugging-facehttps://whatisai.fyi/hugging-faceHugging Face runs the Hub, a website where people share AI models, datasets and small demo apps so others can find, download and reuse them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hugging-face-transformershttps://whatisai.fyi/hugging-face-transformersTransformers is a free, open-source Python library that lets you download a ready-trained AI model by name and run or train it with a short piece of code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/human-in-the-loophttps://whatisai.fyi/human-in-the-loopWith this setup, a person checks an AI system's work at chosen points and can approve it, change it, reject it or step in.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/humanevalhttps://whatisai.fyi/humanevalHumanEval measures functional correctness for programs synthesized from docstrings.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hybrid-searchhttps://whatisai.fyi/hybrid-searchHybrid search runs a keyword ranker and a vector ranker on one query, then merges the two lists into a single order.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hydehttps://whatisai.fyi/hydeHyDE has a language model write a made-up answer to your question, then searches for real documents that look like that answer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hyperparametershttps://whatisai.fyi/hyperparametersHyperparameters are the settings people choose to control training, such as the learning rate, rather than values the model learns itself.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/image-editing-modelshttps://whatisai.fyi/image-editing-modelsAn image editing model changes a picture you already have.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/imbalanced-datahttps://whatisai.fyi/imbalanced-dataImbalanced data is a dataset where one class is far rarer than another, such as a few fraud cases among many ordinary card payments.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/indirect-prompt-injectionhttps://whatisai.fyi/indirect-prompt-injectionIndirect prompt injection is when an attacker hides instructions in content an AI app reads later, such as an email or web page, instead of typing them in.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/inferencehttps://whatisai.fyi/inferenceInference is using a trained AI model on new input to get an answer, such as a label, a number or a written reply, without changing what the model learned.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/inference-optimizationhttps://whatisai.fyi/inference-optimizationInference optimization is a set of methods, like caching, quantization and speculative decoding, that let a trained language model answer with less work or less memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/instruct-modelshttps://whatisai.fyi/instruct-modelsAn instruct model is a pretrained base version further fine-tuned on instructions and conversational data.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/instruction-followinghttps://whatisai.fyi/instruction-followingInstruction following is how well a model does what a written request asks, meeting each rule it sets, such as rules on content, style, format or numbers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/instruction-tuninghttps://whatisai.fyi/instruction-tuningInstruction tuning fine-tunes a pretrained model on many tasks written as instructions paired with desired responses.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/jailbreakshttps://whatisai.fyi/jailbreaksA jailbreak is a prompt written to trick an AI model into ignoring its safety training and producing something it would normally refuse.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/json-modehttps://whatisai.fyi/json-modeJSON mode is an API setting that asks a language model to reply in JSON, text a program can parse (read into data), and on some providers, such as OpenAI, enforces valid syntax, without fixing which keys or types appear.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/k-means-clusteringhttps://whatisai.fyi/k-means-clusteringK-means splits unlabelled data into k groups by putting each point with its nearest centre, then moving each centre to the average of its group, over and over.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/k-nearest-neighborshttps://whatisai.fyi/k-nearest-neighborsk-nearest neighbours labels a new example by finding the k most similar stored examples and taking their majority vote, or their average for numbers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/knowledge-cutoffhttps://whatisai.fyi/knowledge-cutoffA knowledge cutoff is the latest period a model can reliably know from its training alone.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/knowledge-graphshttps://whatisai.fyi/knowledge-graphsA knowledge graph stores facts as labelled links between real-world things, so software can follow the links to answer questions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/kv-cachehttps://whatisai.fyi/kv-cacheA KV cache lets a language model save work from earlier tokens, so each new token does not redo the same maths.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/labelshttps://whatisai.fyi/labelsA label is the answer attached to a training example, such as a category, a number or a ranking, that a supervised model learns to predict.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/langchainhttps://whatisai.fyi/langchainLangChain is an open-source framework for building apps and agents on top of large language models, with one standard way to talk to many AI providers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/langgraphhttps://whatisai.fyi/langgraphLangGraph is an open-source framework for building AI agents as graphs, where steps share one state that can be saved, paused and resumed.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/latencyhttps://whatisai.fyi/latencyRequest latency is the elapsed time between sending an inference request and receiving its final response.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/layer-normalizationhttps://whatisai.fyi/layer-normalizationLayer normalization normalizes selected activations independently for each example in a batch.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/learning-ratehttps://whatisai.fyi/learning-rateThe learning rate multiplies a gradient, and that product is how far the weight moves.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/linear-algebra-for-mlhttps://whatisai.fyi/linear-algebra-for-mlThe small part of linear algebra that models use every day, vectors, matrices, their products and their sizes, to hold data, run layers and learn.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/linear-attentionhttps://whatisai.fyi/linear-attentionLinear attention rewrites attention with feature maps so key-value summaries are formed before they are combined with queries.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/linear-regressionhttps://whatisai.fyi/linear-regressionLinear regression predicts a number by multiplying each feature by a learned weight, adding the results and adding a bias.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llama-cpphttps://whatisai.fyi/llama-cppllama.cpp is an open-source program that runs large language models on your own computer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llamaindexhttps://whatisai.fyi/llamaindexLlamaIndex is an open-source toolkit that brings your own files and data to a language model when you ask a question, so apps and agents can answer from them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llmhttps://whatisai.fyi/llmAn LLM predicts a token or sequence of tokens, sometimes many paragraphs long.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-as-a-judgehttps://whatisai.fyi/llm-as-a-judgeLLM-as-a-judge means asking a strong language model to grade answers to open-ended questions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-evaluationhttps://whatisai.fyi/llm-evaluationLLM evaluation means testing an AI feature with structured tests, so you can check how accurate and reliable it is even though its answers vary.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-gatewayshttps://whatisai.fyi/llm-gatewaysAn LLM gateway is one server that sits between your apps and AI model providers, adding limits, caching, fallbacks and logs to model calls.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-observabilityhttps://whatisai.fyi/llm-observabilityLLM observability means recording what an AI app does on each request, such as the prompt, each step, the time taken and the tokens used.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llmopshttps://whatisai.fyi/llmopsThe day-to-day work of running an app built on a language model, from managing prompts and testing answers to watching speed and cost.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/logistic-regressionhttps://whatisai.fyi/logistic-regressionLogistic regression weighs and adds up an example's features, squeezes the total into a probability between 0 and 1, and uses a threshold to pick a class.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/logitshttps://whatisai.fyi/logitsLogits are a model's raw, unnormalized prediction scores.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/long-contexthttps://whatisai.fyi/long-contextTransformer-XL reuses hidden states from previous segments instead of computing them from scratch for each new segment.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/lorahttps://whatisai.fyi/loraLoRA adapts a large model to a new task by freezing its weights and training two thin matrices whose product is added to them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/loss-functionhttps://whatisai.fyi/loss-functionA loss function turns the difference between a model's prediction and its target into a number that training tries to reduce.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/lost-in-the-middlehttps://whatisai.fyi/lost-in-the-middleLanguage models often use facts near the opening or closing of a long input better than facts buried in the middle.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/lstmhttps://whatisai.fyi/lstmAn LSTM is a recurrent unit whose multiplicative gates regulate access to its memory path.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/machine-learninghttps://whatisai.fyi/machine-learningMachine learning trains software on data so it can find patterns and make useful predictions or generate content for new inputs.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mambahttps://whatisai.fyi/mambaMamba is a sequence architecture whose state space parameters depend on the input, letting each token control what information is kept or forgotten.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mastrahttps://whatisai.fyi/mastraMastra is an open-source TypeScript toolkit for making AI agents, step-by-step workflows and the apps that use them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/matriceshttps://whatisai.fyi/matricesA matrix is a rectangular grid of numbers with a fixed number of rows and columns. Machine learning keeps data and layer weights in matrices and multiplies them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/max-tokenshttps://whatisai.fyi/max-tokensMax tokens sets an upper bound on how many tokens a model may generate for one response.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mcphttps://whatisai.fyi/mcpMCP standardizes how AI applications connect to external context and tools.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/meta-aihttps://whatisai.fyi/meta-aiMeta AI is Meta’s assistant across its apps, the web and a standalone app for conversation, search, voice and image creation.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/metadata-filteringhttps://whatisai.fyi/metadata-filteringMetadata filtering limits a vector search to records whose labels, such as category or year, match rules you set.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/microsoft-copilothttps://whatisai.fyi/microsoft-copilotMicrosoft Copilot is an AI helper from Microsoft that you can type to, talk to or show a picture, and it will reply, write something new or carry out a job for you.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/microsoft-foundryhttps://whatisai.fyi/microsoft-foundryMicrosoft Foundry is Microsoft's Azure platform for building AI apps and agents, with a large model catalogue, tools, testing and safety controls in one place.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/missing-datahttps://whatisai.fyi/missing-dataMissing data are empty cells in a dataset. You can drop them, fill them with reasoned guesses or use a model that accepts gaps, and why they are missing decides which is safe.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mixed-precisionhttps://whatisai.fyi/mixed-precisionMixed precision trains with both 16-bit and 32-bit numbers so the run is faster and uses less memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mixture-of-expertshttps://whatisai.fyi/mixture-of-expertsA mixture-of-experts layer uses a learned gate to select a sparse combination of expert subnetworks for each input.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mlopshttps://whatisai.fyi/mlopsMLOps is a way of working that makes building and releasing machine learning models simpler and more automatic.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mmluhttps://whatisai.fyi/mmluMMLU is a 57-task test of a text model's multitask accuracy.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/model-routinghttps://whatisai.fyi/model-routingModel routing sends each prompt to the model that suits it, so simple requests go to cheaper models and hard ones to stronger models.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/model-servinghttps://whatisai.fyi/model-servingModel serving means running a trained model on a server so apps can send it requests over the network and get answers back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/monte-carlo-methodshttps://whatisai.fyi/monte-carlo-methodsMonte Carlo methods estimate a number by running many random simulations and averaging what comes out, instead of working the answer out exactly.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/multi-agent-systemshttps://whatisai.fyi/multi-agent-systemsA multi-agent system is a group of AI agents, each a language model using tools in its own loop, that work together on one job.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/multi-armed-banditshttps://whatisai.fyi/multi-armed-banditsA multi-armed bandit applies an action, observes its reward, then continues the process with another action.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/multi-head-attentionhttps://whatisai.fyi/multi-head-attentionMulti-head attention consists of several attention layers running in parallel.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/multilayer-perceptronhttps://whatisai.fyi/multilayer-perceptronA multilayer perceptron can fit a nonlinear model to training data.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/multimodal-modelshttps://whatisai.fyi/multimodal-modelsA multimodal model can process more than one type of input, such as text, images, audio or video.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/music-generation-modelshttps://whatisai.fyi/music-generation-modelsA music generation model can generate music in the raw-audio domain.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/naive-bayeshttps://whatisai.fyi/naive-bayesNaive Bayes sorts things into classes by multiplying how common each class is by how likely each feature is in that class, then picking the top score.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/neural-networkshttps://whatisai.fyi/neural-networksA neural network is a model built from layers of simple units that each weigh their inputs, add them up and bend the result, so that together they can learn curved, complex patterns.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/neural-radiance-fieldshttps://whatisai.fyi/neural-radiance-fieldsA neural radiance field maps a 3D position and viewing direction to density and view-dependent color.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/next-token-predictionhttps://whatisai.fyi/next-token-predictionNext-token prediction estimates which token should follow the tokens already present in a sequence.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/npushttps://whatisai.fyi/npusAn NPU is a part of a chip built to speed up AI models at low power.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ollamahttps://whatisai.fyi/ollamaOllama is an open-source program that downloads open AI models and runs them on your own computer, with a simple command and a local API.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/omni-modelshttps://whatisai.fyi/omni-modelsAn omni model processes several input and output modalities within one multimodal model rather than routing everything through a text-only core.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/on-device-aihttps://whatisai.fyi/on-device-aiOn-device AI runs the model on your own phone, laptop or browser, so your data does not have to go to a server.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/one-hot-encodinghttps://whatisai.fyi/one-hot-encodingOne-hot encoding turns a category, such as a colour or a type of transport, into a list of 0s with a single 1 marking which category it is.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/open-webuihttps://whatisai.fyi/open-webuiOpen WebUI, a free chat app you host yourself, lets you talk to AI models running on your own computer or in the cloud, from one web page.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/open-weight-modelshttps://whatisai.fyi/open-weight-modelsAn open-weight model makes its trained weights publicly available for download.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/openai-agents-sdkhttps://whatisai.fyi/openai-agents-sdkOpenAI's Agents SDK is a small open-source toolkit from OpenAI for building AI agents that use tools, pass work to each other and can record each run as a trace.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/outliershttps://whatisai.fyi/outliersAn outlier is a data point that sits an unusually long way from the rest. It may be a mistake to fix or a rare real event worth keeping.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/overfittinghttps://whatisai.fyi/overfittingOverfitting happens when a model learns its training examples so specifically that it performs poorly on new examples.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/parameter-efficient-fine-tuninghttps://whatisai.fyi/parameter-efficient-fine-tuningPEFT adapts a pretrained model by training a small set of added or selected parameters while keeping most base weights frozen.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/parametershttps://whatisai.fyi/parametersParameters are the model values that training can change to improve how inputs map to outputs.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pefthttps://whatisai.fyi/peftPEFT is a free Hugging Face library that teaches a big AI model a new task by training a small set of extra weights instead of the whole model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/perceptronhttps://whatisai.fyi/perceptronA perceptron is a linear classifier.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/perplexityhttps://whatisai.fyi/perplexityPerplexity is an AI search product that searches sources, synthesizes an answer and links the evidence used.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/perplexity-metrichttps://whatisai.fyi/perplexity-metricPerplexity is the exponentiated average negative log-likelihood that a language model assigns to a token sequence.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/positional-encodinghttps://whatisai.fyi/positional-encodingPositional encoding adds token-order information to representations before attention reads them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/precision-and-recallhttps://whatisai.fyi/precision-and-recallPrecision asks how many of the things a model flagged were right. Recall asks how many of the real cases it managed to find.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pretraininghttps://whatisai.fyi/pretrainingIn one transfer-learning setup, a model is pretrained on a data-rich task before fine-tuning on a downstream task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/principal-component-analysishttps://whatisai.fyi/principal-component-analysisPCA finds the directions in which data spreads out most, then keeps only the first few, so many features become a few new ones.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/probabilityhttps://whatisai.fyi/probabilityA probability is a number from 0 to 1 that says how likely something is. Many machine learning problems need one as the answer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/prompt-cachinghttps://whatisai.fyi/prompt-cachingPrompt caching saves the work a model did on the unchanged start of a prompt, so the next request with the same start is faster and cheaper.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/prompt-chaininghttps://whatisai.fyi/prompt-chainingIn prompt chaining, you give a model one small job at a time in a set order, and each answer becomes the starting material for the next job.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/prompt-engineeringhttps://whatisai.fyi/prompt-engineeringPrompt engineering is writing and testing the instructions and examples you give a model so its answer meets a goal you can check.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/prompt-injectionhttps://whatisai.fyi/prompt-injectionPrompt injection is text that sneaks new instructions into what an AI model reads, so the model behaves in unintended ways.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pydantic-aihttps://whatisai.fyi/pydantic-aiPydantic AI is a Python framework for building AI agents whose tool inputs and final answers are checked against types you define.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pytorchhttps://whatisai.fyi/pytorchPyTorch is a free Python library for building and training neural networks, with fast maths on GPUs and automatic gradients.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/q-learninghttps://whatisai.fyi/q-learningQ-learning improves an estimate of each state-action value from sampled rewards and the best estimated value at the next state.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/qlorahttps://whatisai.fyi/qloraQLoRA fine-tunes a large language model by storing its frozen weights in 4 bits and training only a small add-on that sits beside them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/quantizationhttps://whatisai.fyi/quantizationQuantization stores a model's numbers with fewer bits, such as 8-bit integers instead of 32-bit decimals, so it needs less memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/raghttps://whatisai.fyi/ragRAG lets an AI model look up relevant documents first, then answer from what it found.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/rate-limitshttps://whatisai.fyi/rate-limitsA rate limit caps how much you can use an AI service in a set stretch of time, such as how many requests you send each minute.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/reacthttps://whatisai.fyi/reactReAct is a prompting method in which a language model alternates written thoughts with tool actions, reading each result before it picks the next step.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/reasoning-modelshttps://whatisai.fyi/reasoning-modelsReasoning models use intermediate reasoning tokens before producing a final response.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/recommender-systemshttps://whatisai.fyi/recommender-systemsA recommender system picks the few items, out of a catalogue too big to browse, that each person sees first, by predicting what that person is likely to want.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/recurrent-neural-networkshttps://whatisai.fyi/recurrent-neural-networksAt each time step, a recurrent neural network updates its hidden state from the previous hidden state and the current input.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/red-teaminghttps://whatisai.fyi/red-teamingRed teaming means deliberately testing an AI system the way an attacker would, to find its weak spots.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/regressionhttps://whatisai.fyi/regressionRegression predicts a number, such as a house price, a tree's life span or a rainfall total, from an example's features.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/reinforcement-learninghttps://whatisai.fyi/reinforcement-learningReinforcement learning trains a program, called an agent, to choose actions by trial and error, guided by a number called a reward that says how well it is doing.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/reluhttps://whatisai.fyi/reluReLU is an activation function that replaces every negative input with zero and leaves every positive input unchanged.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/reranker-modelshttps://whatisai.fyi/reranker-modelsA reranker sorts existing text candidates by semantic relevance to a specified query.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/rerankinghttps://whatisai.fyi/rerankingReranking takes the short list a first search returned and scores each item again with a slower, more careful model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/residual-connectionshttps://whatisai.fyi/residual-connectionsA residual connection combines a residual mapping F(x) with the block input x as F(x) + x.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/residual-networkshttps://whatisai.fyi/residual-networksA residual block adds a learned residual function F(x) to a shortcut carrying the block input x.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/rlaifhttps://whatisai.fyi/rlaifRLAIF runs the usual RLHF training loop, but another AI model supplies some of the preference labels that people would normally give.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/rlhfhttps://whatisai.fyi/rlhfRLHF uses human comparisons to learn a reward signal, then optimizes a model to produce responses that score better under that signal.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/robotics-foundation-modelshttps://whatisai.fyi/robotics-foundation-modelsA robotics foundation model is pretrained across many tasks, environments or robot bodies so it can serve as a starting point for new robot policies.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/roc-curvehttps://whatisai.fyi/roc-curveAn ROC curve shows, for every threshold a classifier could use, how many real positives it catches against how many false alarms it raises.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/rotary-position-embeddingshttps://whatisai.fyi/rotary-position-embeddingsRoPE encodes absolute position with a rotation matrix and adds explicit relative-position dependence to self-attention.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sandboxing-agentshttps://whatisai.fyi/sandboxing-agentsSandboxing an agent means running its commands inside a fence, set up by the operating system, that limits which files and websites they can reach.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/self-attentionhttps://whatisai.fyi/self-attentionSelf-attention updates each position by comparing it with positions in the same sequence and mixing their information.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/self-hosting-llmshttps://whatisai.fyi/self-hosting-llmsSelf-hosting means running a language model on a computer you control, so your prompts go to your own machine instead of a company's service.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/self-reflectionhttps://whatisai.fyi/self-reflectionSelf-reflection is when an AI model checks its own first answer, writes feedback on it, and then tries again using that feedback.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/self-supervised-learninghttps://whatisai.fyi/self-supervised-learningSelf-supervised learning trains a model on raw data by hiding part of each example and asking the model to predict it, so the data supplies its own labels.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/semantic-kernelhttps://whatisai.fyi/semantic-kernelSemantic Kernel is an open-source Microsoft toolkit that connects AI models to your own code, so a model can ask your functions to do real work.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/semantic-searchhttps://whatisai.fyi/semantic-searchSemantic search finds text by meaning, so a passage can match a question even when the two share no words.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/semi-supervised-learninghttps://whatisai.fyi/semi-supervised-learningSemi-supervised learning trains a model on a few labelled examples plus many unlabelled ones, so the unlabelled data can help when answers are scarce.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sglanghttps://whatisai.fyi/sglangSGLang is open-source server software that runs large language models on GPUs and answers many requests quickly, partly by reusing work shared between prompts.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/small-language-modelshttps://whatisai.fyi/small-language-modelsSmall-model research includes sub-billion-parameter language models for mobile deployment.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/smolagentshttps://whatisai.fyi/smolagentsSmolagents is a small open-source Python library from Hugging Face for building AI agents, including ones that act by writing short pieces of Python code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/softmaxhttps://whatisai.fyi/softmaxSoftmax determines a probability for each possible class, and those probabilities add up to exactly 1.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sparse-attentionhttps://whatisai.fyi/sparse-attentionSparse attention computes selected query-key connections instead of filling the entire attention matrix.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sparsityhttps://whatisai.fyi/sparsitySparsity means that many possible model values or computation paths are zero, absent, or inactive for a given input.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/speculative-decodinghttps://whatisai.fyi/speculative-decodingSpeculative decoding uses a target model to evaluate guesses in parallel.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/speech-to-text-modelshttps://whatisai.fyi/speech-to-text-modelsA speech-to-text model can transcribe a speech utterance into written characters.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/state-space-modelshttps://whatisai.fyi/state-space-modelsA continuous-time linear state space model uses x-dot = Ax + Bu for state evolution and y = Cx + Du for readout.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/statistics-for-mlhttps://whatisai.fyi/statistics-for-mlStatistics describes data and measures how far to trust a model's score, since every test set is only a sample of the cases the model will meet.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/stochastic-gradient-descenthttps://whatisai.fyi/stochastic-gradient-descentStochastic gradient descent steps the weights using a slope estimated from one randomly picked example.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/stop-sequenceshttps://whatisai.fyi/stop-sequencesA stop sequence is a configured character string that stops output generation.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/strands-agentshttps://whatisai.fyi/strands-agentsStrands Agents is a free AWS toolkit, released as open source under Apache 2.0, for making AI agents with Python or TypeScript, where the model itself plans the steps and picks the tools.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/structured-outputshttps://whatisai.fyi/structured-outputsStructured outputs constrain a model response to a declared schema so software receives expected fields and types instead of free-form prose.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sub-agentshttps://whatisai.fyi/sub-agentsA sub-agent is a helper AI that a main agent sends off to do one side task in its own workspace, then report back a short result.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/summarizationhttps://whatisai.fyi/summarizationSummarization is when a model turns a long text into a shorter version that keeps the important information.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/supervised-fine-tuninghttps://whatisai.fyi/supervised-fine-tuningSupervised fine-tuning keeps training an already-trained model on example questions paired with good answers, so it learns to answer that way.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/supervised-learninghttps://whatisai.fyi/supervised-learningSupervised learning trains a model on examples that already carry the right answer, called a label, so it can predict that answer for new examples that do not.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/support-vector-machineshttps://whatisai.fyi/support-vector-machinesA support vector machine separates two classes with the boundary that leaves the widest possible gap to the nearest training points of each class.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/swe-benchhttps://whatisai.fyi/swe-benchSWE-bench asks a system to generate a repository patch for a real GitHub issue.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sycophancyhttps://whatisai.fyi/sycophancySycophancy is when an assistant favors matching a user's beliefs over a truthful response.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/symbolic-aihttps://whatisai.fyi/symbolic-aiSymbolic AI writes knowledge down as readable symbols and if-then rules, then reaches answers by applying logic and search to them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/synthetic-datahttps://whatisai.fyi/synthetic-dataSynthetic data is artificial data built from seed data so some patterns of that seed can be reused.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/system-prompthttps://whatisai.fyi/system-promptA system prompt is high-priority context that sets an AI assistant's role, boundaries, style, and response rules before the user's request is handled.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/temperaturehttps://whatisai.fyi/temperatureTemperature enters softmax by dividing each logit by T.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/test-sethttps://whatisai.fyi/test-setA test set is a reserved group of labeled examples used for a final check of a model on data it did not train on.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/test-time-computehttps://whatisai.fyi/test-time-computeTest-time compute lets a language model improve its output by using more computation at test time.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/text-to-image-modelshttps://whatisai.fyi/text-to-image-modelsA text-to-image model can use text as a conditioning input to an image generator.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/text-to-speech-modelshttps://whatisai.fyi/text-to-speech-modelsA text-to-speech model synthesizes speech directly from text.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/text-to-video-modelshttps://whatisai.fyi/text-to-video-modelsGiven a text prompt, a text-to-video model generates a video.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/throughputhttps://whatisai.fyi/throughputThroughput is the amount of completed inference work divided by the time spent measuring it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/time-series-forecastinghttps://whatisai.fyi/time-series-forecastingTime series forecasting predicts future values of something measured over time, such as sales or electricity demand, from the patterns in its past.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/token-costshttps://whatisai.fyi/token-costsAI model APIs charge by the token, with separate prices for the text you send in and the text the model writes back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/tokenizationhttps://whatisai.fyi/tokenizationTokenization turns text into a sequence of vocabulary IDs that a model can process, then reverses generated IDs back into readable text.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/tokenshttps://whatisai.fyi/tokensTokens are chunks of text processed by text-generation and embedding models.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/tokens-per-secondhttps://whatisai.fyi/tokens-per-secondTokens per second reports how many tokens a server returns in a set amount of time.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/tool-usehttps://whatisai.fyi/tool-useTool use is a loop where a model proposes a call, a program outside the model runs it, and the result is written back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/top-k-samplinghttps://whatisai.fyi/top-k-samplingTop-k sampling keeps the k highest-probability next-token candidates, rescales their probabilities and samples one of them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/top-p-samplinghttps://whatisai.fyi/top-p-samplingTop-p keeps the smallest set of most probable tokens whose probabilities add up to at least p.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/tpushttps://whatisai.fyi/tpusTPUs are chips Google designed to speed up the maths of machine learning, rented out through Google Cloud.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/traininghttps://whatisai.fyi/trainingTraining is the repeated loop that nudges a model's internal numbers so its predictions get closer to the right answers in its examples.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/training-datahttps://whatisai.fyi/training-dataTraining data is the set of examples a machine learning model studies to learn its patterns, kept apart from the examples used to test it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/transfer-learninghttps://whatisai.fyi/transfer-learningTransfer learning reuses representations learned on a source task to improve learning on a different target task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/transformershttps://whatisai.fyi/transformersA transformer is a neural network that uses attention to decide which parts of a sequence matter to one another.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/turing-testhttps://whatisai.fyi/turing-testThe Turing test asks whether a judge, chatting by text with a hidden person and a hidden machine, can reliably tell which one is the machine.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/u-nethttps://whatisai.fyi/u-netU-Net combines upsampled decoder features with matching high-resolution features from a contracting path.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/unslothhttps://whatisai.fyi/unslothUnsloth is an open-source tool for training and running AI language models on your own computer, built to fine-tune them faster and with less graphics-card memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/unsupervised-learninghttps://whatisai.fyi/unsupervised-learningUnsupervised learning trains a model on examples that have no answers attached, so it finds structure on its own, such as groups of similar items or points that do not fit.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/validation-sethttps://whatisai.fyi/validation-setA validation set is held-out data used during development to compare model choices without training on the same examples.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/variational-autoencodershttps://whatisai.fyi/variational-autoencodersA variational autoencoder pairs a deep latent-variable model with a corresponding approximate inference model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vector-databaseshttps://whatisai.fyi/vector-databasesA vector database stores vectors and returns the saved records closest to a query vector, often with a score and any extra fields you ask for.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vectorshttps://whatisai.fyi/vectorsA vector is an ordered list of numbers that you can also picture as an arrow in space. Machine learning stores examples, words and images this way.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vercel-ai-sdkhttps://whatisai.fyi/vercel-ai-sdkThe Vercel AI SDK is an open-source TypeScript toolkit, free to use, that lets apps talk to AI models from many companies through the same code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vertex-aihttps://whatisai.fyi/vertex-aiVertex AI is Google Cloud’s platform for choosing AI models, tuning them on your own examples and running models and agents for real users.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vision-language-action-modelshttps://whatisai.fyi/vision-language-action-modelsA vision-language-action model maps visual observations and a language instruction to robot actions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vision-language-modelshttps://whatisai.fyi/vision-language-modelsA vision-language model can take visual data and text as input and produce text as output.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vision-transformershttps://whatisai.fyi/vision-transformersA vision transformer represents an image as a sequence of patches and processes that sequence with a transformer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vllmhttps://whatisai.fyi/vllmvLLM is open-source software that runs large language models on a server and answers many requests at once, using GPU memory carefully so fewer bytes go to waste.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vocabularyhttps://whatisai.fyi/vocabularyA model vocabulary supplies the mapping used to convert text into an ID sequence and back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/voice-agentshttps://whatisai.fyi/voice-agentsA voice agent is an AI app you talk to out loud; it listens, works out what you want, can use tools, and answers in speech.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/weightshttps://whatisai.fyi/weightsTrainable weights are model parameters that can be learned from data.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/world-modelshttps://whatisai.fyi/world-modelsA world model predicts how an environment may change after an action, so an agent can evaluate possible futures before acting.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/artificial-intelligencehttps://whatisai.fyi/artificial-intelligencePublished the full Artificial intelligence explainer. It is the starting point for the site's introductory learning path.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llmhttps://whatisai.fyi/llmPublished the full Large language model explainer. Readers need a grounded explanation of tokens, training and generation before following model news.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ai-agentshttps://whatisai.fyi/ai-agentsPublished the full AI agents explainer. It explains the model, tool and control loop behind agentic applications.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/raghttps://whatisai.fyi/ragPublished the full Retrieval-Augmented Generation explainer. It gives readers a concrete path from indexed documents to a grounded answer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/embeddingshttps://whatisai.fyi/embeddingsPublished the full Embeddings explainer. Embeddings connect token representations, semantic search and retrieval systems.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mcphttps://whatisai.fyi/mcpPublished the full Model Context Protocol explainer. It explains how AI applications connect to external tools and data through a shared protocol.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gpushttps://whatisai.fyi/gpusPublished a new explainer. A GPU is a chip built to run many similar calculations at the same time, which suits the matrix maths inside neural networks.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/flopshttps://whatisai.fyi/flopsPublished a new explainer. FLOPs count the small arithmetic steps, like one multiply or one add on decimal numbers, that a computer does to train an AI model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-statehttps://whatisai.fyi/agent-statePublished a new explainer. Agent state is the running record an AI agent keeps while it works, so a task can pause, resume, or recover from a crash.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/prompt-injectionhttps://whatisai.fyi/prompt-injectionPublished a new explainer. Prompt injection is text that sneaks new instructions into what an AI model reads, so the model behaves in unintended ways.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/cpushttps://whatisai.fyi/cpusPublished a new explainer. A CPU is the chip that runs a computer's software, carrying out instructions with just a few cores backed by lots of cache memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/tpushttps://whatisai.fyi/tpusPublished a new explainer. TPUs are chips Google designed to speed up the maths of machine learning, rented out through Google Cloud.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/npushttps://whatisai.fyi/npusPublished a new explainer. An NPU is a part of a chip built to speed up AI models at low power.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ai-acceleratorshttps://whatisai.fyi/ai-acceleratorsPublished a new explainer. AI accelerators are chips built to speed up the maths inside AI models, which is largely multiply-add sums.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/guardrailshttps://whatisai.fyi/guardrailsPublished a new explainer. Guardrails are checks placed around an AI model that inspect what goes in and what comes out, and can block, change or check inputs and replies that are unsafe, off-topic or break the app's rules.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/evalshttps://whatisai.fyi/evalsPublished a new explainer. An eval checks an AI system's work. You feed it an input and use a set of rules to score what comes back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-evaluationhttps://whatisai.fyi/llm-evaluationPublished a new explainer. LLM evaluation means testing an AI feature with structured tests, so you can check how accurate and reliable it is even though its answers vary.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-evaluationhttps://whatisai.fyi/agent-evaluationPublished a new explainer. Agent evaluation tests an AI agent on set tasks, checking both its final result and the steps it took, over several tries.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-orchestrationhttps://whatisai.fyi/agent-orchestrationPublished a new explainer. Agent orchestration is how an app with several AI agents decides which agent works, in what order, and who picks what happens next.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/amazon-bedrockhttps://whatisai.fyi/amazon-bedrockPublished a new explainer. Amazon Bedrock is an AWS service that lets developers call AI models from several companies through one set of APIs, with extra tools for search and safety.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/benchmarkshttps://whatisai.fyi/benchmarksPublished a new explainer. A benchmark is a fixed set of test tasks with a scoring rule, so different AI models can be compared on the same exam.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/claude-codehttps://whatisai.fyi/claude-codePublished a new explainer. Claude Code is Anthropic's coding assistant that reads a project, edits files and runs commands for you, asking before risky steps.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/context-compactionhttps://whatisai.fyi/context-compactionPublished a new explainer. Context compaction swaps the older part of a long AI conversation for a short summary, so the chat can keep going inside the model's memory limit.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/contextual-retrievalhttps://whatisai.fyi/contextual-retrievalPublished a new explainer. Contextual retrieval has a model write a short note that places each chunk in its document, then indexes the note with the chunk so search can find it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/crewaihttps://whatisai.fyi/crewaiPublished a new explainer. CrewAI is an open-source toolkit, written in Python, for building teams of AI agents, each with a role, that work through a list of tasks together.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/cursorhttps://whatisai.fyi/cursorPublished a new explainer. Cursor is a coding tool with an AI agent that can read your project, edit files and run commands for you.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/data-privacyhttps://whatisai.fyi/data-privacyPublished a new explainer. Data privacy is about what an AI company may do with your chats, such as training models, and the settings that let you limit it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/feature-engineeringhttps://whatisai.fyi/feature-engineeringPublished a new explainer. Feature engineering turns raw data, like prices and colour names, into the lists of numbers a model can learn from.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/github-copilothttps://whatisai.fyi/github-copilotPublished a new explainer. GitHub Copilot is an AI helper for programmers that offers the next lines of code while you write, explains code when you ask, and can take on jobs you give it.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/google-adkhttps://whatisai.fyi/google-adkPublished a new explainer. Google ADK is an open-source toolkit from Google for writing AI agents in code, giving them tools, a record of each chat and helper agents.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/graphraghttps://whatisai.fyi/graphragPublished a new explainer. GraphRAG turns your documents into a map of people, places and links, groups and summarises it, then answers questions from that map.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/haystackhttps://whatisai.fyi/haystackPublished a new explainer. An open-source Python toolkit from deepset that builds AI apps (search, RAG, agents) by linking small parts into a pipeline.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hnswhttps://whatisai.fyi/hnswPublished a new explainer. HNSW finds the stored vectors closest to a query quickly by searching a stack of linked graphs, from a sparse top layer down to a dense bottom one.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/human-in-the-loophttps://whatisai.fyi/human-in-the-loopPublished a new explainer. With this setup, a person checks an AI system's work at chosen points and can approve it, change it, reject it or step in.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/jailbreakshttps://whatisai.fyi/jailbreaksPublished a new explainer. A jailbreak is a prompt written to trick an AI model into ignoring its safety training and producing something it would normally refuse.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/knowledge-graphshttps://whatisai.fyi/knowledge-graphsPublished a new explainer. A knowledge graph stores facts as labelled links between real-world things, so software can follow the links to answer questions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/langgraphhttps://whatisai.fyi/langgraphPublished a new explainer. LangGraph is an open-source framework for building AI agents as graphs, where steps share one state that can be saved, paused and resumed.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llamaindexhttps://whatisai.fyi/llamaindexPublished a new explainer. LlamaIndex is an open-source toolkit that brings your own files and data to a language model when you ask a question, so apps and agents can answer from them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llmopshttps://whatisai.fyi/llmopsPublished a new explainer. The day-to-day work of running an app built on a language model, from managing prompts and testing answers to watching speed and cost.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/lost-in-the-middlehttps://whatisai.fyi/lost-in-the-middlePublished a new explainer. Language models often use facts near the opening or closing of a long input better than facts buried in the middle.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/metadata-filteringhttps://whatisai.fyi/metadata-filteringPublished a new explainer. Metadata filtering limits a vector search to records whose labels, such as category or year, match rules you set.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/microsoft-foundryhttps://whatisai.fyi/microsoft-foundryPublished a new explainer. Microsoft Foundry is Microsoft's Azure platform for building AI apps and agents, with a large model catalogue, tools, testing and safety controls in one place.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/ollamahttps://whatisai.fyi/ollamaPublished a new explainer. Ollama is an open-source program that downloads open AI models and runs them on your own computer, with a simple command and a local API.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pytorchhttps://whatisai.fyi/pytorchPublished a new explainer. PyTorch is a free Python library for building and training neural networks, with fast maths on GPUs and automatic gradients.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/rate-limitshttps://whatisai.fyi/rate-limitsPublished a new explainer. A rate limit caps how much you can use an AI service in a set stretch of time, such as how many requests you send each minute.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/reacthttps://whatisai.fyi/reactPublished a new explainer. ReAct is a prompting method in which a language model alternates written thoughts with tool actions, reading each result before it picks the next step.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/red-teaminghttps://whatisai.fyi/red-teamingPublished a new explainer. Red teaming means deliberately testing an AI system the way an attacker would, to find its weak spots.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sandboxing-agentshttps://whatisai.fyi/sandboxing-agentsPublished a new explainer. Sandboxing an agent means running its commands inside a fence, set up by the operating system, that limits which files and websites they can reach.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sub-agentshttps://whatisai.fyi/sub-agentsPublished a new explainer. A sub-agent is a helper AI that a main agent sends off to do one side task in its own workspace, then report back a short result.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/supervised-fine-tuninghttps://whatisai.fyi/supervised-fine-tuningPublished a new explainer. Supervised fine-tuning keeps training an already-trained model on example questions paired with good answers, so it learns to answer that way.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/token-costshttps://whatisai.fyi/token-costsPublished a new explainer. AI model APIs charge by the token, with separate prices for the text you send in and the text the model writes back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/lorahttps://whatisai.fyi/loraPublished a new explainer. LoRA adapts a large model to a new task by freezing its weights and training two thin matrices whose product is added to them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pefthttps://whatisai.fyi/peftPublished a new explainer. PEFT is a free Hugging Face library that teaches a big AI model a new task by training a small set of extra weights instead of the whole model.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vertex-aihttps://whatisai.fyi/vertex-aiPublished a new explainer. Vertex AI is Google Cloud’s platform for choosing AI models, tuning them on your own examples and running models and agents for real users.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/on-device-aihttps://whatisai.fyi/on-device-aiPublished a new explainer. On-device AI runs the model on your own phone, laptop or browser, so your data does not have to go to a server.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-as-a-judgehttps://whatisai.fyi/llm-as-a-judgePublished a new explainer. LLM-as-a-judge means asking a strong language model to grade answers to open-ended questions.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/summarizationhttps://whatisai.fyi/summarizationPublished a new explainer. Summarization is when a model turns a long text into a shorter version that keeps the important information.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-observabilityhttps://whatisai.fyi/llm-observabilityPublished a new explainer. LLM observability means recording what an AI app does on each request, such as the prompt, each step, the time taken and the tokens used.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/prompt-cachinghttps://whatisai.fyi/prompt-cachingPublished a new explainer. Prompt caching saves the work a model did on the unchanged start of a prompt, so the next request with the same start is faster and cheaper.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/pydantic-aihttps://whatisai.fyi/pydantic-aiPublished a new explainer. Pydantic AI is a Python framework for building AI agents whose tool inputs and final answers are checked against types you define.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/model-servinghttps://whatisai.fyi/model-servingPublished a new explainer. Model serving means running a trained model on a server so apps can send it requests over the network and get answers back.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/kv-cachehttps://whatisai.fyi/kv-cachePublished a new explainer. A KV cache lets a language model save work from earlier tokens, so each new token does not redo the same maths.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/openai-agents-sdkhttps://whatisai.fyi/openai-agents-sdkPublished a new explainer. OpenAI's Agents SDK is a small open-source toolkit from OpenAI for building AI agents that use tools, pass work to each other and can record each run as a trace.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llm-gatewayshttps://whatisai.fyi/llm-gatewaysPublished a new explainer. An LLM gateway is one server that sits between your apps and AI model providers, adding limits, caching, fallbacks and logs to model calls.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/json-modehttps://whatisai.fyi/json-modePublished a new explainer. JSON mode is an API setting that asks a language model to reply in JSON, text a program can parse (read into data), and on some providers, such as OpenAI, enforces valid syntax, without fixing which keys or types appear.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/context-engineeringhttps://whatisai.fyi/context-engineeringPublished a new explainer. Context engineering is choosing, trimming and updating the tokens a model sees at each step, so an AI agent works from a small, useful context.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/amazon-bedrock-agentcorehttps://whatisai.fyi/amazon-bedrock-agentcorePublished a new explainer. A set of AWS services, called Amazon Bedrock AgentCore, for running AI agents in the cloud, with hosting, memory, sign-in, tool access and monitoring.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-memoryhttps://whatisai.fyi/agent-memoryPublished a new explainer. Agent memory is how an AI agent saves useful information outside the model and loads the right pieces back into its prompt later.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/claude-agent-sdkhttps://whatisai.fyi/claude-agent-sdkPublished a new explainer. Anthropic's Claude Agent SDK is a Python and TypeScript library for building agents that run on the same loop, tools and context handling as Claude Code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/batch-inferencehttps://whatisai.fyi/batch-inferencePublished a new explainer. Batch inference sends many AI requests as one job that runs in the background, trading an instant reply for a lower price.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/a2a-protocolhttps://whatisai.fyi/a2a-protocolPublished a new explainer. A2A is an open protocol for AI agents: one agent can find another, hand it a task and track the result, even when different companies built them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/benchmark-contaminationhttps://whatisai.fyi/benchmark-contaminationPublished a new explainer. Benchmark contamination happens when test questions, or close copies of them, end up in a model's training data, so its score can look better than its real skill.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/langchainhttps://whatisai.fyi/langchainPublished a new explainer. LangChain is an open-source framework for building apps and agents on top of large language models, with one standard way to talk to many AI providers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agentic-workflowshttps://whatisai.fyi/agentic-workflowsPublished a new explainer. An agentic workflow runs a task as several language model and tool steps, joined by a path that your code fixes in advance.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/aws-inferentiahttps://whatisai.fyi/aws-inferentiaPublished a new explainer. AWS Inferentia is a family of computer chips that Amazon designed to run trained AI models cheaply and quickly in its cloud.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/azure-openaihttps://whatisai.fyi/azure-openaiPublished a new explainer. Azure OpenAI lets companies use OpenAI's models through Microsoft's Azure cloud, with Azure billing, safety filters and data controls.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/quantizationhttps://whatisai.fyi/quantizationPublished a new explainer. Quantization stores a model's numbers with fewer bits, such as 8-bit integers instead of 32-bit decimals, so it needs less memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agent-planninghttps://whatisai.fyi/agent-planningPublished a new explainer. Agent planning is how an AI agent splits a big task into smaller steps first, then carries out those steps one by one.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/aws-trainiumhttps://whatisai.fyi/aws-trainiumPublished a new explainer. AWS Trainium is a computer chip Amazon designed for training and running AI models, rented through special Amazon EC2 cloud servers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/multi-agent-systemshttps://whatisai.fyi/multi-agent-systemsPublished a new explainer. A multi-agent system is a group of AI agents, each a language model using tools in its own loop, that work together on one job.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/approximate-nearest-neighbor-searchhttps://whatisai.fyi/approximate-nearest-neighbor-searchPublished a new explainer. Approximate nearest neighbor search finds stored vectors that are very close to a query quickly, by checking only a promising part of the collection and accepting a few misses.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/inference-optimizationhttps://whatisai.fyi/inference-optimizationPublished a new explainer. Inference optimization is a set of methods, like caching, quantization and speculative decoding, that let a trained language model answer with less work or less memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/strands-agentshttps://whatisai.fyi/strands-agentsPublished a new explainer. Strands Agents is a free AWS toolkit, released as open source under Apache 2.0, for making AI agents with Python or TypeScript, where the model itself plans the steps and picks the tools.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/voice-agentshttps://whatisai.fyi/voice-agentsPublished a new explainer. A voice agent is an AI app you talk to out loud; it listens, works out what you want, can use tools, and answers in speech.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/agnohttps://whatisai.fyi/agnoPublished a new explainer. Agno is open-source software for building AI agents and running them as a live service. You can build single agents, teams of agents and step-by-step workflows.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/indirect-prompt-injectionhttps://whatisai.fyi/indirect-prompt-injectionPublished a new explainer. Indirect prompt injection is when an attacker hides instructions in content an AI app reads later, such as an email or web page, instead of typing them in.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/code-interpretershttps://whatisai.fyi/code-interpretersPublished a new explainer. A code interpreter lets an AI model write a small program, run it in a sealed-off sandbox, and use the result in its answer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/autogenhttps://whatisai.fyi/autogenPublished a new explainer. AutoGen is an open-source Microsoft framework for building apps where several AI agents talk to each other, and sometimes to people, to finish a task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hugging-facehttps://whatisai.fyi/hugging-facePublished a new explainer. Hugging Face runs the Hub, a website where people share AI models, datasets and small demo apps so others can find, download and reuse them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/comfyuihttps://whatisai.fyi/comfyuiPublished a new explainer. ComfyUI is a free, open-source app where you build AI image, video and audio generators by wiring boxes called nodes together.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mlopshttps://whatisai.fyi/mlopsPublished a new explainer. MLOps is a way of working that makes building and releasing machine learning models simpler and more automatic.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/qlorahttps://whatisai.fyi/qloraPublished a new explainer. QLoRA fine-tunes a large language model by storing its frozen weights in 4 bits and training only a small add-on that sits beside them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/fine-tuninghttps://whatisai.fyi/fine-tuningPublished a new explainer. Fine-tuning takes a model that is already trained and keeps training it on a smaller set of examples for one task.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hydehttps://whatisai.fyi/hydePublished a new explainer. HyDE has a language model write a made-up answer to your question, then searches for real documents that look like that answer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/model-routinghttps://whatisai.fyi/model-routingPublished a new explainer. Model routing sends each prompt to the model that suits it, so simple requests go to cheaper models and hard ones to stronger models.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/open-webuihttps://whatisai.fyi/open-webuiPublished a new explainer. Open WebUI, a free chat app you host yourself, lets you talk to AI models running on your own computer or in the cloud, from one web page.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/sglanghttps://whatisai.fyi/sglangPublished a new explainer. SGLang is open-source server software that runs large language models on GPUs and answers many requests quickly, partly by reusing work shared between prompts.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/mastrahttps://whatisai.fyi/mastraPublished a new explainer. Mastra is an open-source TypeScript toolkit for making AI agents, step-by-step workflows and the apps that use them.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/self-reflectionhttps://whatisai.fyi/self-reflectionPublished a new explainer. Self-reflection is when an AI model checks its own first answer, writes feedback on it, and then tries again using that feedback.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/llama-cpphttps://whatisai.fyi/llama-cppPublished a new explainer. llama.cpp is an open-source program that runs large language models on your own computer.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/hugging-face-transformershttps://whatisai.fyi/hugging-face-transformersPublished a new explainer. Transformers is a free, open-source Python library that lets you download a ready-trained AI model by name and run or train it with a short piece of code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/self-hosting-llmshttps://whatisai.fyi/self-hosting-llmsPublished a new explainer. Self-hosting means running a language model on a computer you control, so your prompts go to your own machine instead of a company's service.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/smolagentshttps://whatisai.fyi/smolagentsPublished a new explainer. Smolagents is a small open-source Python library from Hugging Face for building AI agents, including ones that act by writing short pieces of Python code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/amazon-sagemaker-aihttps://whatisai.fyi/amazon-sagemaker-aiPublished a new explainer. SageMaker AI, from AWS, lets you build, train and run machine learning models without managing your own servers.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/unslothhttps://whatisai.fyi/unslothPublished a new explainer. Unsloth is an open-source tool for training and running AI language models on your own computer, built to fine-tune them faster and with less graphics-card memory.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/semantic-kernelhttps://whatisai.fyi/semantic-kernelPublished a new explainer. Semantic Kernel is an open-source Microsoft toolkit that connects AI models to your own code, so a model can ask your functions to do real work.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/dspyhttps://whatisai.fyi/dspyPublished a new explainer. DSPy is a Python framework where you describe an AI job by naming its inputs and outputs, and it tunes the prompts for you.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vercel-ai-sdkhttps://whatisai.fyi/vercel-ai-sdkPublished a new explainer. The Vercel AI SDK is an open-source TypeScript toolkit, free to use, that lets apps talk to AI models from many companies through the same code.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/vllmhttps://whatisai.fyi/vllmPublished a new explainer. vLLM is open-source software that runs large language models on a server and answers many requests at once, using GPU memory carefully so fewer bytes go to waste.Mon, 28 Sep 2026 00:00:00 GMThttps://whatisai.fyi/gemma-4https://whatisai.fyi/gemma-4Gemma 4 is Google’s open-weight, natively multimodal model family, designed for advanced reasoning and agentic workflows across edge devices and personal computers. Verified against Google AI for Developers's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/jevhttps://whatisai.fyi/jevJev is TypeSafe AI’s decision model, returning typed choices, scores, or probabilities for software automation instead of generating open-ended natural-language responses. Verified against TypeSafe AI's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/gpt-6-astrahttps://whatisai.fyi/gpt-6-astraGPT-6 Astra is OpenAI’s 2026 frontier model for advanced reasoning, coding, computer use, scientific work, and long-running agentic tasks across ChatGPT and the API. Verified against OpenAI's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/claude-opus-5-5https://whatisai.fyi/claude-opus-5-5Claude Opus 5.5 is Anthropic’s 2026 Opus model, positioned for demanding coding, research, and professional work with stronger efficiency than the previous Opus generation. Verified against Anthropic's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/claude-fable-5-1https://whatisai.fyi/claude-fable-5-1Claude Fable 5.1 is Anthropic’s generally available model for ambitious coding, knowledge work, research, and long-running agents, with configurable effort and broad platform availability. Verified against Anthropic's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/gemini-3-5https://whatisai.fyi/gemini-3-5Gemini 3.5 is Google’s 2026 multimodal model generation, built for stronger reasoning, coding, agentic workflows, and understanding across text, images, audio, video, and tools. Verified against Google's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/grok-4-7https://whatisai.fyi/grok-4-7Grok 4.7 is xAI’s 2026 model release for reasoning, coding, information synthesis, and tool-using tasks across the Grok product and developer platform. Verified against xAI's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/chatgpt-workhttps://whatisai.fyi/chatgpt-workChatGPT Work is OpenAI’s agent for longer projects, able to work across connected apps and files and create finished documents, spreadsheets, presentations, reports, and sites. Verified against OpenAI's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/gemini-enterprisehttps://whatisai.fyi/gemini-enterpriseGemini Enterprise is Google Cloud’s workplace AI platform for searching company information, using enterprise data and applications, and building governed agents for business workflows. Verified against Google's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/musehttps://whatisai.fyi/museMuse is Meta’s personal AI agent for carrying out tasks and longer-term goals across connected apps, using a dedicated secure virtual machine and asking for approval before sensitive actions. Verified against Meta's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/muse-spark-1-1https://whatisai.fyi/muse-spark-1-1Muse Spark 1.1 is Meta’s multimodal reasoning model for agentic tasks, with tool and computer use, coding, long-context work, and multi-agent orchestration through Meta AI and the Meta Model API. Verified against Meta AI's official announcement.Sun, 27 Sep 2026 12:00:00 GMThttps://whatisai.fyi/updates/#2026-09-27https://whatisai.fyi/updates/#2026-09-27Created the catalogue, removed the low-value Code repos category, and added the current Gemma 4, Jev and Laya model records. The catalogue should help readers understand AI, not inflate its count with repository names.Sun, 27 Sep 2026 00:00:00 GMThttps://whatisai.fyi/updates/#2026-09-27https://whatisai.fyi/updates/#2026-09-27Removed superseded model versions and discontinued or redundant products, normalized current product names, and added seven current model and product references. The catalogue should lead with what readers encounter now while retaining only historically useful landmark terms.Sun, 27 Sep 2026 00:00:00 GMT