# AI Field Notes: YouTube Channel Launch Kit

**Status:** proposed channel concept. Check channel name and handle availability before creating the account.

## Positioning

**Working name:** AI Field Notes  
**Tagline:** AI tools without the hype.  
**One-line promise:** Independent, practical explainers that help people choose and use AI models and tools with evidence, clear trade-offs, and no breathless claims.

**Audience:** curious builders, creators, students, and small teams who want to use AI but need help separating practical capability from marketing.

**Editorial stance:** test workflows; link primary sources; state versions and dates; show limitations; disclose sponsorships and affiliate links; distinguish observed results from opinion.

## Channel description

> AI Field Notes is an independent guide to the fast-moving AI landscape. Get practical model explainers, hands-on tool reviews, local AI walkthroughs, and creator workflows without the hype. Every episode aims to show what was tested, what it costs, what to verify, and who it may suit. Models and tools change quickly, so check linked sources for current versions, pricing, licensing, and requirements. Hosted by @joshuawayzwright.

## Brand direction

- **Voice:** curious, direct, skeptical but fair, and useful to newcomers without talking down to technical viewers.
- **Visuals:** field-note motif, clear labels, restrained lime/coral accents, readable model names, real interface capture, and source links on screen.
- **Thumbnail system:** one question or comparison, one recognizable product/model label, no more than a few words, and imagery from the actual workflow where possible.
- **Handle ideas:** `@AIFieldNotes`, `@AIFieldNotesYT`, or a personal-name variant. Availability is not checked.

## Repeatable formats

1. **Model field test:** a consistent task, setup details, observed outputs, limitations, and links to model cards.
2. **Release briefing:** what changed, who can access it, requirements, pricing, and what remains unknown.
3. **Workflow walkthrough:** an end-to-end task using real screens, with costs and failure points.
4. **Tool comparison:** defined criteria and a small, reproducible test rather than an unsupported universal ranking.
5. **Shorts:** one finding, one caveat, one source or next step in 30–60 seconds.

**Starter cadence:** one researched long-form video per week and two Shorts cut from that research. Reduce cadence rather than skip source checking.

## First ten episode briefs

1. How to choose a local AI model for your laptop
2. What model-card fields matter before you download a model
3. A repeatable way to compare two AI coding assistants
4. Running a small language model locally: setup, memory, and trade-offs
5. How to verify an AI model's license before using it
6. What happens when the internet goes away: an offline AI workflow
7. Image-generation model comparison using one prompt and clear criteria
8. AI voice tools: quality, consent, and commercial-use checks
9. Building a private document assistant with local tools
10. How to keep an AI workflow's recurring costs visible

For comparisons, document exact model versions, prompts, hardware, settings, date, and source links. Do not publish a result as a benchmark unless the test supports that claim.

## First-video outline

**Working title:** How to choose a local AI model for your laptop  
**Viewer takeaway:** leave with a short checklist for matching a model to hardware, task, and privacy needs.

- Cold open: the model name is only half the decision; hardware and use case matter too.
- Define the task: chat, coding, transcription, image generation, or embeddings.
- Show how to check memory requirements, quantization, format, and supported runtimes.
- Demonstrate one real download/run workflow and note setup friction.
- Compare local privacy and control with speed, quality, and hardware trade-offs.
- Close with a checklist and links to official model cards and licenses.

## Production workflow

1. Pick a viewer question and write the one-sentence takeaway.
2. Collect primary sources: model card, release notes, official docs, license, and pricing page.
3. Test the workflow yourself. Record environment, versions, prompts, settings, and failures.
4. Draft in Creator Studio; replace placeholders with verified evidence and citations.
5. Record narration or camera, then capture genuine screen footage. Use licensed music, visuals, and clips.
6. Edit for clarity, add readable citations, captions, chapters, and a truthful thumbnail/title.
7. Review all factual claims, links, disclosure language, and accessibility before upload.
8. Publish; then review impressions, click-through rate, retention, and useful viewer questions to choose the next test.

## Upload checklist

- Confirm title matches the actual test; avoid claiming a winner without evidence.
- Add a concise description, primary-source links, and chapters where useful.
- Use a custom thumbnail with legible text and accurate imagery.
- Add captions and review their technical names.
- Disclose paid placements, affiliate links, free access, and material relationships clearly.
- Check music, footage, logos, model licenses, and any voice or likeness permissions.
- Pin a question that invites corrections and source-backed follow-up.

## Launch tasks

- Check YouTube name/handle availability and create the channel using your own Google account.
- Add channel art, profile image, description, links, and contact details.
- Create playlists for model explainers, workflow walkthroughs, tool comparisons, and Shorts.
- Prepare three videos before announcing the channel so the first release has follow-ups.
- Submit your first few videos only after the evidence, license, and disclosure review.

This kit is a planning document, not a claim that a YouTube channel or handle has already been created.