← Course Map All Courses
Module 03 · 4 lessons · 6 quiz Qs

The Tools You Will Meet

Knowing the tool world stops you from chasing every new launch and helps you pick the two or three that fit your work.

Coastal cliffs and winding road over the sea

Photograph via Unsplash

Lesson 3.1 · Foundation Models for Conversation and Text

Three foundation models matter for most working professionals in 2026: ChatGPT (OpenAI), Claude (Anthropic), and Gemini (Google). A fourth, Llama (Meta), matters more if you host models on your own hardware. A fifth, Mistral, matters mostly to developers.

ChatGPT is the broadest. It has the largest user base, the most third-party integrations, and the deepest set of plugins and custom apps. It tends to be slightly better at image generation built-in, slightly more eager to do what you ask, and slightly more prone to hallucination on Caribbean-specific facts. Free tier is useful. Paid tier (around US$20 a month) gives you the better model and faster response.

Claude tends to write better long-form prose than the others, follows instructions more carefully when those instructions are complex, and is generally more honest about what it does not know. It has lower hallucination rates on technical questions. Caribbean creatives who write a lot, edit a lot, or work with detailed briefs often prefer Claude.

Gemini sits inside Google's product family. If your business works in Gmail, Docs, Sheets and Drive, Gemini integrates more deeply than the others.

The rule: pick one as your default for a month and use it for everything. Notice where it falls short, then add a second one to fill the gaps rather than to replace the first.

Lesson 3.2 · Image, Video, Audio, and the Specialist Tools

Image generators. Midjourney remains the most consistently good for stylised, atmospheric, hand-illustrated-feeling output. DALL-E (inside ChatGPT) is best for quick concepts. Stable Diffusion is the engine behind most “free, run-on-your-machine” alternatives. Flux is the new technical leader for photorealism. Ideogram handles text-in-images better than the others.

Video generators. Runway is the most established. Sora (OpenAI) leads on quality but is rationed. Veo (Google) is competitive. Kling is the rising challenger from Asia. As of 2026, all of them produce thirty to sixty seconds of usable footage. Long-form coherence (a single character keeping the same face across an entire scene) is still a struggle.

Audio and music. Suno and Udio handle text-to-music. ElevenLabs generates speech and clones voices from a few seconds of reference audio. Stable Audio and Riffusion handle sound design and loops.

Editing assistants. Adobe Firefly, Photoshop's generative fill, Premiere's generative tools, DaVinci Resolve's AI features, and Canva Magic. The professional software you already use is being rebuilt with AI inside.

Lesson 3.3 · Specialist and Emerging Tools

Perplexity. A search-augmented chat tool. Better than the foundation models for questions where current facts matter, because it cites its sources. Use this when you need recent information rather than reasoning.

NotebookLM. A Google tool that lets you upload documents and ask questions specifically about them. The model can only answer from the documents you provide, which dramatically reduces hallucination. Useful for digesting long contracts, briefs, or research papers.

Cursor and GitHub Copilot. Code assistants. If anyone in your business does any web development or scripting, these change the timeline of small projects from weeks to days.

HeyGen and Synthesia. AI avatars for video. Useful when you need a talking-head presenter and cannot get one in person.

Captions. Short-form video editing assistant, popular for Instagram Reels and TikTok production.

ChatGPT custom GPTs and Claude Projects. Both major foundation model providers now let you create a custom version of the assistant trained on your own knowledge base. For a working creative this means you can build a “studio assistant” that knows your prices, your contract templates, your client list, and your style. It is worth investing one Saturday in.

Lesson 3.4 · A Working Toolkit for the Average Caribbean Creative

If you are starting from zero in 2026, here is a defensible toolkit that costs about US$60 a month and covers most use cases.

Daily driver for text and brief drafting. Claude or ChatGPT, paid tier. About US$20 a month.

Research and current events. Perplexity, free tier is sufficient for most users. Paid tier is about US$20 a month if you do heavy research.

Image generation. Midjourney, US$10 to US$30 a month depending on how much you generate. If you do not generate images regularly, the free DALL-E inside ChatGPT is a starting point.

Video and audio (occasional). Pay-per-use credits on Runway, Suno or ElevenLabs. Around US$10 a month if you use them weekly.

Free additions. Google Gemini if you live in Google Workspace. Adobe Firefly if you have Adobe subscriptions already.

Measure your progress by what you produce with these tools rather than by how many you collect.

Module 03 Quiz · pass mark 5 of 7 points

Q3.1 · Which foundation model tends to be most honest about what it does not know and has lower hallucination rates on technical questions?
Q3.2 · A St Andrew filmmaker wants to test what a scene might look like before paying for a real shoot. Which family of tool fits this need best?
Q3.3 · True or false. Subscribing to all major AI tools at once is the most efficient way to find the right toolkit for your work.
Q3.4 · A producer wants answers grounded only in a forty-page client brief, with no extra invention. Which tool is purpose-built for this?
Q3.5 · (2 points · short answer) A Liguanea studio has US$60/month and needs text drafting, current-events research, image generation, and occasional video. Sketch a defensible toolkit choice.
Q3.6 · Which statement about long-form video generation in 2026 is most accurate?