I tested ChatGPT's new image model and ran it through real business use cases. Real estate flyers pulled from a live Zillow link. Meta ads from a website URL. Product packaging. Brand mood boards. App store mockups. Logo exploration sheets. A 30-day content calendar. Travel itineraries. Recipe cards. A six-panel comic strip. Each one used to be a $200 to $2,000 line item with a designer. The bottleneck for small businesses was never creativity. It was the gap between an idea and a finished asset. That gap is closing fast. Full 31-minute walkthrough is up on YouTube. https://www.youtube.com/watch?v=zd5jt7hVe5s&t=8s The full 18-prompt pack is now in the Classroom: Free Downloads → 18 ChatGPT Image Prompts. PDF download, free for members.

New guides within the classroom section 🎯 START HERE — 10 lessons. Orientation, the recommended tool stack, and how to get your bearings in your first hour. Built so you stop guessing where to start. 🛠️ The Foundations — 12 lessons. Image prompt anatomy, video prompt anatomy, and every major tool worth learning right now: Nano Banana 2, ChatGPT Image, Freepik / Magnific, Seedance 2.0, Veo 3.1, Kling 3.0, ElevenLabs, Suno. Plus the consistent character method, a system for organizing AI work, and your first Claude Code agent. This is the meat of the creative AI craft. 💰 AI for Money — 12 lessons. Land your first client, price your work, structure proposals and contracts, build the personal brand that converts, plus the 90-day plan from "I have AI skills" to your first $10k month. If you're new here: Open START HERE. Pick one course to focus on this week. Post your first generation in #General before the weekend. If you've been here a while: A lot changed. Click into any course you haven't touched yet. The Foundations is where most returning members will get the biggest jump. AI for Money if you're trying to turn this into income. The space moved fast while I was heads-down. Tools, models, opportunities. The Classroom now has the structure to keep up. More coming this month: the AI North Star is back, monthly. Live Q&A Wednesdays 7:30pm ET. If you've shipped anything since we last talked, drop it in #General. Tag me. I want to see what you've built. — West

I’ve had some people ask me what the difference is between Freepik and Higgsfield. One of the main differences is Freepik’s node-based model called "Spaces." Here’s a quick video I made going through some Freepik Spaces use cases and why I recommend getting familiar with this tool. Check it out.

### ByteDance Suspends Seedance 2.0 Global Launch: Copyright Disputes Win Round One ByteDance hit pause on the global rollout of Seedance 2.0—the model that was reshaping video production at $0.42 per shot. Reason: cease-and-desist letters from Disney, Paramount Skydance, and other major Hollywood studios over use of copyrighted materials. Why it matters: This is the first real casualty in the AI video arms race. Seedance 2.0 was the most capable, lowest-cost tool on the market. Its suspension changes the competitive math entirely. Other platforms will face the same copyright pressure if they don't solve this now. This isn't about technical capability anymore—it's about legal liability. If you're building with Seedance anywhere, the model could disappear from under you. Watch the others (Runway, Pika, Google Veo) for how they handle training data transparency over the next 30 days. <https://www.reuters.com/technology/bytedance-suspends-launch-video-ai-model-after-copyright-disputes-information-2026-03-14/> --- ## The Insights ### Study: AI Makes Humans More Creative (Not Just Faster) Swansea University tested 800+ participants on design tasks using AI-supported tools. Finding: people with access to AI-generated design galleries spent *more* time on the task, produced *better* designs, and felt *more* involved. The kicker: **including intentionally flawed AI suggestions helped humans explore further and break out of early assumptions.** Diversity in output—even bad ideas—sparked creative risk-taking. Why it matters: The narrative that AI replaces creativity is backwards. AI as a collaborator that shows you possibilities you wouldn't explore alone changes how you approach your craft. You're not getting faster at the same thing. You're accessing ideas at a different scale. <https://www.sciencedaily.com/releases/2026/03/260315004355.htm

### Pika Labs Drops AI Selves: Your Digital Clone Is Live Pika didn't just release a feature. They released your replacement. AI Selves lets you upload reference material and Pika generates a persistent AI video avatar that moves, speaks, and reacts like you. This isn't deepfake territory—it's a fully controllable digital extension of your identity that can be deployed at scale. The real implication: You can now generate personalized video content featuring *yourself* in cinematic scenes, product demos, explainers, and social content. Minutes instead of hours. No actors. No cameras. Just your AI double working around the clock. The UGC and short-form ad space is already buzzing. An AI system that creates, tests, and scales personalized ads featuring your face 24/7 is no longer science fiction. Why it matters: Solo creators and small teams just got a $10K production capability in their pocket. <https://www.vo3ai.com/blog/pika-ai-selves-launch-shakes-up-ai-video-personalization-google-and-grok-fire-ba-2026-03-07> ### Grok Video Model Quietly Took the #1 Spot While Pika and Google traded headlines, xAI's Grok image-to-video model climbed to the top of the AI video leaderboard. Score: 1406. Beating Google Veo, Runway, Pika—everyone. This matters because Grok entered the game late. Jumping to #1 in weeks means xAI is throwing serious compute at this. If they drop API access to the broader market, the competitive landscape shifts again. Why it matters: Best-in-class performance matters. Grok being #1 signals where the money is flowing in AI infrastructure. ### Midjourney V8: Speed Distillation Phase Complete David Holz confirmed on March 4 that V8 is in final distillation—tuned for pure speed. New server cluster coming end of March to handle the load. Why it matters: V8 should be available any day now. This is the image generation upgrade you've been waiting on. Mark your calendar.