ArtificialIntelligence.How

Checked August 2026

A luminous stream of ideas moving through an abstract AI system and emerging as creative work

The useful guide to artificial intelligence

AI changes weekly.
Start here anyway.

Current models, real costs, complete guides, local hardware, useful projects, and the communities where people are doing the work.

Compare frontier models →Browse complete guides

Interactive AI laboratories

Put the model—and the agent around it—under glass.

Inspect live transformer matrices, then trace context, prompt injection, agent loops, tools, skills, plugins, multi-agent topology, OpenClaw, and Hermes Agent through observable system state.

Enter the AI Lab →Transformer microscope

Choose a section

Find what you need to learn or decide

Four sculptural AI systems transforming the same input into different outputs on a comparison benchModelsChoose a current frontier model

Compare OpenAI, Anthropic, Google, and xAI by access, cost, context, and actual workload fit.

Explore section →A person thoughtfully using AI alongside books, notes, and project materials in a daylight studioGuidesLearn by finishing useful work

Complete guides for first use, APIs, model evaluations, document Q&A, and safe agents.

Explore section →A compact AI computer and a larger GPU workstation arranged on a warm, practical studio deskLocal AIRun capable models on your hardware

Understand memory, software, privacy, DGX Spark, Mac Studio, RTX desktops, and professional workstations.

Explore section →A creative workshop filled with prototypes, maps, materials, and connected project ideasIdeasBuild something with an outcome

Practical work and private local projects with a small first version and a visible success test.

Explore section →AI builders collaborating around a live hardware demonstration at an independent conferenceCommunityFind conferences and user groups

Research events, builder conferences, practitioner communities, open-model groups, and local chapters.

Explore section →

Possibilities, not hype

What will you do with it?

The useful question is not what AI can do in theory. It is what you can finish, understand, test, or share this week.

A person thoughtfully using AI alongside books, notes, and project materials in a daylight studioUnderstand your materialTurn a private document collection into answers you can verifyStart here →A creative workshop filled with prototypes, maps, materials, and connected project ideasMake something usefulPrototype a tool, research a question, or organize a body of workStart here →AI builders collaborating around a live hardware demonstration at an independent conferenceFind your peopleMeet builders, researchers, operators, and curious practitionersStart here →

Current reference pages

Start with what changed

Model and hardware pages are dated, linked to primary sources, and written to answer a specific search or purchase question.

ModelsGPT-5.6 Sol, Terra, and Luna: access, pricing, and which one to use

A practical guide to the GPT-5.6 model family, including Sol, Terra, Luna, API pricing, context limits, reasoning settings, access, and workload selection.

Read 9 minute guide →ModelsClaude Fable 5, Opus 5, and Sonnet 5: what each model is for

Compare Anthropic's Claude Fable 5, Opus 5, and Sonnet 5 models by access, price, context, latency, coding ability, and production use.

Read 8 minute guide →ModelsGemini 3.1 Pro and Gemini 3.6 Flash for multimodal and agentic work

A practical guide to Gemini 3.1 Pro Preview and Gemini 3.6 Flash, including multimodal capabilities, API access, 1M context, pricing tiers, and workload choice.

Read 8 minute guide →ModelsGrok 4.5: pricing, access, search tools, and practical use cases

Learn when to use Grok 4.5, how to access it through Grok and the xAI API, what token and long-context pricing costs, and where search tools change the workflow.

Read 7 minute guide →Local AINVIDIA DGX Spark for local AI: what it runs and how to set it up

A complete NVIDIA DGX Spark guide covering the GB10 chip, 128 GB unified memory, current price, model fit, first boot, Ollama, LM Studio, llama.cpp, networking, and benchmarks.

Read 16 minute guide →GuidesHow to start using AI and get a result you can trust

A complete beginner's guide to using ChatGPT, Claude, Gemini, or Grok for a real task, including prompt structure, source checking, privacy, and a practical first exercise.

Read 10 minute guide →

A practical standard

Use AI deliberately, not automatically

  • Start with a real task and a visible success test.
  • Use the least expensive model that reliably passes.
  • Keep private data inside an approved boundary.
  • Check current prices, access, and specifications at the original source.