- Prompt Dojo
- Posts
- 🥋 What If Every Image Could Tell Its Story?
🥋 What If Every Image Could Tell Its Story?
Learn about the latest, trending AI video tools

Welcome back to Prompt Dojo.
This week, we'll explore self-criticism decomposition prompting that can improve AI accuracy up to 90%, compare Google's new Veo 3 video generator with Midjourney's V1 Video Model, and cover OpenAI's infrastructure diversification and its implications for your AI workflows. Ready to level up? Let's begin!
Power Prompt
Self-Criticism Decomposition
Purpose: Dramatically improve AI output quality by having the model break down problems AND critique its own reasoning before delivering final answers.
TASK: [Your specific request]
DECOMPOSITION PHASE:
"First, break this task into 3-5 logical sub-problems. For each sub-problem, identify:
1. What specific information is needed
2. What assumptions might be made
3. What could go wrong in this step"
SOLUTION PHASE:
"Now solve each sub-problem step by step, showing your reasoning."
SELF-CRITICISM PHASE:
"Before giving your final answer, critique your solution:
- What are the weakest points in your reasoning?
- What alternative approaches might work better?
- What would an expert in this field challenge about your answer?"
FINAL ANSWER:
"Based on this self-analysis, provide your refined final answer."
Best models: Claude Sonnet 4 for complex reasoning, ChatGPT Plus for business analysis, Gemini Advanced for creative problems
Source: Sander Schulhoff and The Prompt Report - a comprehensive study co-authored with OpenAI, Microsoft, Google, Princeton, and Stanford analyzing 1,500+ academic papers.
Try this prompt and reply with your results!
Fresh off the Mat News

China's Humanoid Robots Debut Autonomous Soccer League - Four teams of humanoid robots faced off in fully autonomous 3-on-3 soccer matches powered entirely by artificial intelligence in Beijing, touted as a first in China and a preview for the upcoming World Humanoid Robot Games.
OpenAI Partners with Google for AI Infrastructure - OpenAI has recently begun renting Google's artificial intelligence chips to power ChatGPT and its other products, marking the first time OpenAI has used non-Nvidia chips.
Trump Administration Prepares AI Energy Executive Orders - The Trump administration is readying a package of executive actions aimed at boosting energy supply to power the U.S. artificial intelligence infrastructure.
Impact: For users, this means more stable AI services, competitive pricing, and potentially better performance as companies optimize across different chip architectures. The humanoid robotics progress also hints at rapidly advancing embodied AI capabilities. Could Robotic athletes become a thing in the future?
Training Ground
Google Veo 3 vs Midjourney V1 Video for AI Video Generation
For this week we have a matchup prepared for you. We're comparing Google's new Veo 3 and Midjourney's V1 Video Model for AI-generated video content. Both create stunning videos from prompts, but which one delivers better results for creators?
Google Veo 3 - Example
Strengths: Native audio generation with dialogue and sound effects, 8-second videos with potential for longer clips, superior physics and realism
Best for: Complete audiovisual content, realistic human motion, lip-syncing dialogue, cinematic storytelling with sound
Weaknesses: Expensive, limited to 8 seconds currently, strict safety restrictions
Price: AI Pro $19.99/month (3 Veo 3 Fast videos daily, then reverts to Veo 2) .$249.99/month Google AI Ultra plan or Google Cloud Vertex AI enterprise pricing
Midjourney V1 Video Model
Strengths: Exceptional artistic quality, otherworldly aesthetic, extends to 21 seconds, works with any Midjourney image
Best for: Artistic animations, creative visual content, extending existing Midjourney artwork, stylized motion
Weaknesses: No audio generation, web-only access, 8x cost of image generation, requires Midjourney subscription
Price: $10/month basic Midjourney plan plus 8x video generation costs
Takeaway: Midjourney wins for artistic quality, longer duration, and budget-conscious creators. Veo 3 wins for complete audiovisual realism and professional use but is really expensive. If interest, experiment with the Pro plan first to give it a try.

Taiki’s Toolkit- Tools to Explore
Midjourney - Creates 5-second videos extendable to 21 seconds with cinematic quality that feels otherworldly rather than hyperrealistic.
Pika Labs - Advanced keyframe transition system enabling smooth 1-10 second video generation.
Luma - Large-scale video generative model with ultra-realistic details and coherent motion.
Reply