- There's An AI For That
- Posts
- AI Turns Images Into Smooth 3D Videos
AI Turns Images Into Smooth 3D Videos
Agents, MCP, 3D videos, AI + Minecraft, home robots, Google's comeback, and more...
There's An AI For That
March 21, 2025
Hey AI Enthusiast!
Welcome back to the world’s #1 AI newsletter.
Here’s a glimpse into what we have today:
Meta AI is coming to Europe.
Humanoid robots in homes this year.
Build your own custom AI agents fast.
Inside Google’s frenzied AI comeback.
Turn any image into a smooth 3D video.
Assess AI by challenging it with Minecraft.
Stop receiving our newsletter here
Breaking News
The latest developments in AI
Meta AI is rolling out its AI chatbot in Europe for key platforms after nearly a year of regulatory delays, but limiting it to text-only functions.
Norwegian startup 1X is preparing to deploy and test its humanoid robot, Neo Gamma, in hundreds to thousands of homes in 2025.
xAI has launched an API for generating images, currently featuring the "grok-2-image-1212" model, and users can request up to 10 JPGs per input.
Today’s Sponsor
Agent.ai - Work Smarter, Not Harder
Build custom AI agents fast with Agent.ai—no-code, low-code or Python.
Handle text, vision, or mixed-media tasks.
12+ foundation models for solutions that accelerate decision-making and automate tasks.
Integrate webhooks and APIs to support everything from quick demos to fully scaled initiatives without vendor lock-in.
Create one agent or a team of agents.
Agent.ai’s future-proof platform will evolve with your technical abilities and your business goals.
Start today!
Coming in Hot
AI Tools of the Day
💬 Promptimize AI - Turn simple prompts into smart, detailed instructions that generate significantly better responses from AI chatbots like ChatGPT.
💼 PrometAI - Transform your business idea into a structured plan with AI-powered templates—no blank pages, just a clear path to success.
📔 Automateed - Generate and publish entire ebooks with AI, complete with images, content, and even audiobook versions.
🎙️ BusyScribe - Convert WhatsApp, Telegram, and Messenger voice messages into text instantly for easier communication in any situation.
📧 one.email - Cut through email clutter with an AI-driven inbox that simplifies emails, finds what matters, and keeps you focused.
📚 remio - Auto-capture, organize, and retrieve key details from messy information with an AI-powered personal knowledge hub.
📚 Imagine Explainers - Make learning more visual and accessible by creating and sharing explainer videos that break down complex topics.
📊 ProductScope.AI - Design and market any product with AI-generated photos, videos, insights, and optimized copy in minutes, not weeks.
🎨 OctoComics - Craft original character-driven stories and serialized webtoons with AI-assisted scripts, scenes, and immersive visual storytelling.
🎸 Shoots by Diktatorial - Create stage-ready photos and videos from your selfies in seconds, giving your music the spotlight it deserves.
🦾 Submit your AI tool to the #1 AI tools marketplace.
Mind-Blowing AI
Stable Virtual Camera is a new AI tool that’ll let you turn any image into a smooth 3D video, with full control over camera movement.
AI-generated video has been improving fast, but this takes things further by generating depth-accurate perspectives that actually feel natural. Instead of relying on frame-by-frame reconstructions, it builds out new angles from the original image, making it look like the scene was shot with an actual camera. It’s still in research preview, but it’s already showing what’s possible when AI and 3D video start to merge. 🎥🚀
Reach out to over 1.5 million AI enthusiasts today. Become a sponsor
AI Finds
Beyond the Feed
A look at Google’s turbulent two-year sprint to catch up with OpenAI reveals a story of intense pressure, bold risks, and eventual triumph.
This deep dive explores the Model Context Protocol (MCP), an emerging standard that could revolutionize how AI agents interact with external tools.
Minecraft Benchmark (MC-Bench) is a platform where users challenge AI models in Minecraft building contests, offering a fun way to assess generative AI.
“Watch the Skies” is set to become the first feature film dubbed entirely with AI and will hit 100 AMC theaters across major U.S. markets on May 9.
MIT and NVIDIA’s HART is a hybrid image generator that produces high-quality images 9x faster than state-of-the-art models.
By TAAFT
AI Tools Starter Pack
If you’re taking your first steps into AI or looking to upgrade your essential toolkit, our “AI Tools Starter Pack” has you covered.
We've cut through the noise to bring you a focused collection of reliable, practical tools that you’ll want to start using today if you haven’t already.
No matter if you’re a creator, developer, entrepreneur, or simply looking to enhance your workflow, these are the tools that have proven themselves time and again.
Prompt of the Day
Extended Response Sequencer
Click here to copy the full prompt easily!
Simply hover over the prompt after you land on the page and hit “Copy.” Then, paste it into your favorite chatbot and answer the three questions at the bottom of the prompt.
From now on, I require very long responses and a total approximate length that I will specify at the beginning of each question (in words or characters). I understand that you have a word limit per response.
Before starting to generate the response, once you understand my question and the desired total length, analyze the requested total length and estimate how many parts (approximately) you will need to divide the response into to deliver the complete content, respecting your single response length limit.
Communicate to the user the estimated number of parts you will divide the response into before starting to answer. For example: "Understood. For the requested total length, I estimate that the response will be delivered in [number] parts."
Once the estimation is communicated, begin generating the first part of the response. When you approach your single response length limit, stop when you reach the limit.
At the point where you stop, clearly indicate the end of that part of the response with a HIGHLY VISIBLE END OF PART MARKER. Use the following format for the marker:
Use code with caution.
****************** END OF PART [NUMBER] ******************
Where [NUMBER] is replaced by the part number (1, 2, 3, etc.). This marker, with lines of asterisks and uppercase text, ensures high visibility even in plain text.
Wait for me to write the word "continue" or "next" (or any similar indication that I want you to continue) before generating the next part of the response.
When I indicate "continue" or "next", continue the response exactly where you left off in the previous part, without repeating information and maintaining the coherence of the text. Again, when you reach the length limit of the next part, stop, indicate the HIGHLY VISIBLE END OF PART MARKER (with the specified format), and wait for my instruction to continue.
Repeat this process of segmentation and continuation until you have provided the complete and extensive response that is required, covering the approximate total length that I indicated at the beginning.
Interaction Example:
User: Explain in detail the history of the Roman Empire. Approximate total length desired: 2000 words.
ChatGPT: Understood. For the requested total length (2000 words), I estimate that the response will be delivered in 4 parts. Starting with Part 1: [Generates the first part of the response that reaches the length limit]...
Use code with caution.
****************** END OF PART 1 ******************
User: continue
ChatGPT: ♻️ [Continues the response exactly where it left off in Part 1, until reaching the length limit again]...
Use code with caution.
****************** END OF PART 2 ******************
User: next
ChatGPT: ♻️ [Continues from Part 2]... [and so on, with similar markers for each part, until completing approximately 2000 words]
Do you understand these instructions for long segmented responses, with specified total length, part prediction, and HIGHLY VISIBLE END OF PART MARKER? Confirm that you have understood by responding: "Understood. Before responding, I will estimate the number of parts based on the total length requested. Then, I will deliver the response in parts, marking each one with the HIGHLY VISIBLE END OF PART MARKER (with asterisks and 'END OF PART [number]') and waiting for the instruction 'continue' to proceed."Source: Alberto Dotras
And if you have a favorite prompt, share it with us. We may feature it in an upcoming newsletter!
Feedback
What'd you think of today's newsletter?Vote below to let us know how we're doing. |
Too frequent? Update your preference here
Hated it and want to stop receiving them? Unsubscribe here
And that’s a wrap! We hope these AI insights bring you a smooth transition from a busy week into a relaxing weekend. Don’t forget to share your favorite finds with friends, family, and colleagues. Wishing you a fantastic Friday! 🦾
Signing off,
— There’s An AI For That