Before I get started, I’ve been getting 100s of emails per week on cool startups you all are building with AI!
And while I was talking to Accel, I couldn’t help but show some of those emails to the GPs there.
They loved it so much, we’ve partnered to give you all a chance to raise money and get mentored by them if you’re building in AI!
Here’s the deet:
Atoms is Accel’s global pre-seed program for Indian-origin founders with irrational conviction.
✅ $1M+ in capital
✅ Backing from operators at OpenAI, Google, Stripe
✅ Tracks for AI + breakthrough tech (Leaptech)
✅ Flipkart, Freshworks, Zetwerk — all started here
⏳ Applications now live.
👉 Join the next global wave:
I had to record 10 days of video content but I was traveling constantly.
Recording quality content while jumping between airports, hotels, and meetings? Nearly impossible.
So I tried something different. I created 30+ professional videos without recording a single video. Yep, not even one!
Workation tips from the Maldives (never been there).
Productivity hacks from a cozy café in London (also never been there).
Tech reviews from a minimalist home office that looks nothing like my actual messy desk.
It was not just about making videos.
Some of the content I created was impossible to create professionally.
And I also had some fun with it :)
(A still from a video where I am teaching Quantum Mechanics in the Amazon rainforest)
Talking about use cases?
I made product videos, instagram reels, B-rolls, A-rolls, and so much more
All of it with AI.
Using just Flux Kontext and Veo 3.
And then taught our team how to do it to fast-track the process.
And tbh…the result looked better than some actual hollywood movie trailers I've seen.
Okay I know…you all want to try it.
Before I take you through the process, let's talk about this week's AI madness...
1. CodeLLM - Your Code Whisperer
An AI code editor that doesn't just autocomplete. It writes, debugs, and optimizes entire functions while you watch.
It's like having a senior developer sitting next to you, except this one never gets tired of explaining why your loop is inefficient.
Perfect for: Developers tired of switching between ChatGPT and their IDE, data scientists who want code that actually works, and anyone learning to code who needs a patient mentor.
2. Humanizer AI - The AI Detection Escape Artist
Does your email written by GPT sound like a sophisticated robot wrote it during a coffee break?
Well this tool is for you. It takes your AI-generated text and makes it sound like it came from an actual human brain.
Super useful for: Copywriters who use AI for inspiration but need it to pass the "human test," students working on projects, and marketers who want their content to feel authentic.
3. AudioCraft 2.0 - Your Personal Hans Zimmer
Meta's open-source toolkit that generates custom music, sound effects, and audio textures on demand.
Want epic battle music for your game? Relaxing café ambience for your podcast? It creates exactly what you need with emotional control that actually makes sense. Quick tutorial here.
Use case: Game developers who can't afford composers, sound designers building audio libraries, and content creators who are tired of hunting for royalty-free music that doesn't suck..
I was at a startup event last month where this founder pitched his AI salary negotiation app. Use case: Analyzes job market data and helps employees ask for raises using personalized scripts
The audience loved it. "Brilliant!" "So needed!" "When can I invest?"
I wanted to ask the hard questions: What happens when someone gets fired for using your app's aggressive negotiation tactics? What if companies start blacklisting employees who use salary negotiation apps?
Six months later, I heard he had to stop operations after burning through ₹50 lakhs. HR departments were actively warning against employees using his platform, and his legal team said the liability exposure was massive.
The problem: A business is not built on emotions. Friends, family, and sometimes even investors often tell you what you want to hear instead of what you need to hear.
Coming back to my 3-day content challenge...
In 3 minutes per video, I had:
Professional B-roll footage of me in any location
A-roll content with perfect lighting and backgrounds
Talking head videos that looked like expensive studio productions
Content that would have taken weeks to shoot traditionally
No setups, visas and location permits. And the best part? It looked completely natural.
Ready to multiply your content creation speed by 10x? Let's do this!
Let me walk you through creating something actually epic.
What we're making: A superhero movie trailer with you as the main character.
Step 1: Access Your Movie Studio
Go to Flux Kontext (available through various AI platforms). I used the Flux Playground.
Upload a clear photo of yourself (front-facing, good lighting works best)
Step 2: Design your perfect scene
Be specific about your location and purpose:
Example prompt I used: "Professional business person in a modern minimalist office, sitting at a sleek desk with a laptop, large window with city view in background, natural daylight, confident expression, wearing business casual attire, clean aesthetic, high-resolution photography, professional headshot quality"
Flux Kontext will create a photorealistic image with your face perfectly mapped onto the superhero
You can generate multiple shots for different scenes (close-ups, action shots, dramatic poses)
Step 3: Bring It to Life with Veo 3
Now comes the magical part. Turning your storyboard into an actual video:
Veo 3 is available through Gemini Pro.
Sign in through your google account. Gemini Pro is paid but you can get a one-month free subscription which you can cancel anytime.(so effectively free)
Start the trial and come back to Gemini Pro and choose the video tab
Step 4: Prompt Magic for our final showdown
Upload the picture you created through Flux Kontext here.
Give a detailed prompt to bring your character to life
Here’s the prompt I used:
Video prompt for Veo 3: "A professional content creator sitting at a modern desk in a bright office, looking directly at the camera with a friendly, confident expression. The person speaks, ‘Thank you for your love to our newsletter…cheers to staying ahead’, followed by a thumbs up gesture. The camera maintains a steady medium shot with professional lighting. The person appears engaged and authentic, with subtle head movements and natural blinking. Background shows a clean, organized workspace with soft natural lighting from a large window"
The results? Phenomenal.
Step 5: Scale Your Content Production
Generate multiple scenes for different video topics
Create seasonal content by changing backgrounds
Develop series content with consistent settings
Build a library of B-roll footage for future use
Bonus Step: Take it one step further with Voice Cloning
Veo 3 is good but adding your own voice to the video makes it much better. And how to do it?
Go to Eleven Labs, and choose Voice Cloning
In the ElevenLabs dashboard, select the “Voices” section on the left, then click “Add a new voice”.
From the modal, select “Instant Voice Clone”.
Follow the on-screen instructions to upload or record your audio.
Name and label your voice clone, confirm that you have the right and consent to clone the voice, then click “Save voice”.
Under the “Voices” section in the dashboard, select the “Personal” tab, then click on your voice clone to begin using it.
And that’s it! After using a voice clone, even your mom won’t be able to recognise that the video is made by AI.
What type of content do you need to create regularly?
Educational videos? Product demos? Personal brand content? Social media reels?
Describe your ideal content setup and send me relevant images.
I'll create a custom Flux Kontext + Veo 3 workflow that fits your specific needs.
First 10 replies get a complete prompt package with examples.
We're entering a world where "being somewhere" doesn't require... being somewhere. Where your digital presence can be more polished, more professional, and more inspiring than your physical reality.
If I create content of myself in the Maldives without ever visiting, am I lying to my audience? Or am I just being more efficient?
If the value I provide is the same or maybe even better because I had perfect lighting and no distractions…does the location of my physical body actually matter?
Is this the future of human connection? Are we all just going to become perfectly curated digital versions of ourselves?
And if those versions are better teachers, better communicators, and more helpful to our audiences... is that actually a problem?
Maybe the real question isn't whether it's authentic: it's whether it's honest about what it is.
Until next time,
Vaibhav🤝