Adam Gurski Creative

AI 101

The practical companion: context patterns, reusable briefs, current models, tools, sources, and a human-review loop.

01 · Quick start

Run one useful loop.

Choose a task small enough to inspect and useful enough to matter. The model can do the draft work; you still own the outcome.

Source

Bring the real document, data, example, or project.

Brief

Name the role, task, context, and output format.

Build

Ask for an artifact you can examine—not just an answer.

Check

Verify facts, fit, risk, and what changed before using it.

02 · Context

Three levels of control.

Move up only when the task benefits from more durable, inspectable context. “Memory” is never a substitute for checking what the system actually has.

Level 1

Chat in a web app

Fast, temporary, and ideal for bounded questions. You supply the relevant context in the conversation.

Level 2

Project

A curated context boundary with files and instructions for a sustained topic or body of work.

Level 3

Codex or an agentic harness

Source, tools, tests, and durable records enable multi-step execution you can inspect and verify.

Aim for inspectable, user-controlled context—not “perfect memory.” Remove information the task does not need, especially sensitive material.

03 · Prompting

RTCF turns a request into a brief.

It is a starting structure, not a magic incantation. Clear context and an inspectable output matter more than ornamental prompt language.

R · Role

Who should do the thinking?

T · Task

What concrete outcome is needed?

C · Context

What source material and constraints apply?

F · Format

What should arrive, and how will it be checked?

04 · Reusable prompts

Copy, adapt, inspect.

Replace generic nouns with your actual material. Never paste confidential information into a system without confirming that its access and data policy fit the work.

Snake: vague request versus RTCF brief
Role: You are a senior front-end game designer.
Task: Build a polished, playable Snake game in one HTML file.
Context: It will be shown live to a nontechnical audience, must work offline, and must be easy to inspect.
Format: Return one complete HTML file with keyboard and touch controls, score, restart, accessible contrast, and a short verification checklist.
Deep research
Research this decision using current primary sources. Separate confirmed facts from inference, cite every material claim, identify unresolved questions, and end with a one-page decision brief.
Agentic build
Inspect the existing project first. Write acceptance tests for the requested behavior, implement the smallest complete change, run the tests, inspect the result in a browser, and leave a concise handoff with changed files and remaining risks.
Five-level explanation
Explain this topic at five levels: a child, a high-school student, a working professional, a domain specialist, and an executive deciding what to do next. Keep the core idea consistent while changing vocabulary and depth.
Restoration
Assess this damaged or incomplete source. List what is observable, what is uncertain, and what must not be invented. Then propose a reversible restoration plan and create a first pass that preserves provenance.
On-brand creation
Study the supplied brand references before creating anything. Extract the typography, color, spacing, voice, and composition rules; then produce the requested asset and include a short brand-alignment checklist.
05 · Model directory

Pick for access, fit, and control.

Models change quickly. Use these official links to verify availability, terms, and model cards before making a consequential choice.

Major hosted choices

Provider / modelUseful whenAccessOpenness / licenseOfficial source
OpenAI · GPT-5.6General knowledge work, coding, research, and agentsChatGPT, Codex, APIProprietary · service termsOpenAI
Anthropic · Claude Fable 5Long-running complex knowledge and coding workClaude, platform, cloud marketplacesProprietary · service termsAnthropic Fable
Anthropic · Claude Opus 5High-capability analysis and professional workClaude and platformProprietary · service termsAnthropic Opus
Anthropic · Claude Sonnet 5Balanced everyday writing, analysis, and codingClaude and platformProprietary · service termsAnthropic Sonnet
Google · Gemini 3.7 FlashFast multimodal, long-context, coding, and agentsGemini, AI Studio, Vertex AI, APIProprietary · service termsGoogle AI
xAI · Grok 4.6Long-running agents, research, coding, and visualsGrok, partners, APIProprietary · service termsxAI
DeepSeek · DeepSeek V4 Pro / FlashReasoning-effort-controlled agent workflowsApp, web, APIHosted release · service termsDeepSeek
Alibaba · Qwen 3.8Hosted production work with long context and toolsQwen Cloud and productsHosted; related weights belowQwen
Moonshot · Kimi K3Long-context multimodal knowledge and coding workKimi app, web, APIProprietary · service termsMoonshot AI
Mistral · Mistral Medium 3.5Multimodal agentic and coding workAPI and downloadable weightsOpen-weight · Modified MIT v26.04Mistral docs

Leading open / local families

FamilyUseful whenAccessOpenness / licenseOfficial source
Qwen 3.8 Flash-NextLocal or self-hosted multimodal experimentationWeights and common runtimesOpen-weight · Qwen Community License 1.0; conditions applyQwen model card
Gemma 4On-device and self-hosted agentic appsWeights, AI Edge, Colab, Vertex AIOpen model · Apache 2.0Google
Mistral open modelsGeneral, coding, vision, audio, and safety workloadsWeights and Mistral hostingOpen-weight · Apache 2.0 or Modified MIT by modelMistral licensing
NVIDIA NemotronCustomizable agent models and enterprise stacksComponents, NIM, partner runtimesOpen models · NVIDIA Nemotron Open Model License; check cardNVIDIA
OpenAI gpt-ossLocal and self-hosted text reasoning with tools120B / 20B weights and third-party runtimesOpen-weight · Apache 2.0 plus usage policyOpenAI open models
Meta Llama / MuseHosted Meta assistant plus downloadable local optionsMeta AI/API; Llama and Muse Glimmer weightsMixed · proprietary, Llama Community, or Apache 2.0 by releaseMeta Muse Glimmer

“Open-source,” “open-weight,” and “proprietary” are not interchangeable. Verify the specific model card and license for the exact version you intend to use.

06 · Tools and environments

Match the environment to the task.

Start in chat, move to a project when context recurs, and use an agentic harness when the work must inspect files, use tools, and verify an artifact.

07 · Session videos

See the context ladder.

Use this as a companion explanation of the move from a single chat to curated projects and tool-enabled agentic work.

Three levels of AI context and memory controlChat → project → Codex or an agentic harness.
08 · Presentation sources

Follow the evidence.

These are attendee-safe sources used in the presentation. The extended model directory above links directly to provider announcements, documentation, or model cards.

09 · Human judgment

Use the 4Ds.

A short fluency and safety checklist for every consequential task.

Define

Outcome, audience, evidence, boundary.

Delegate

The work the model can actually perform.

Discern

Truth, taste, risk, downstream effects.

Document

Sources, decisions, changes, verification.

10 · Keep going

Bring one real task.

Source it, brief it, build it, check it.