Swarming Agents, Synthetic Videos & Microsoft CEOs Talk to AI. Plus Buy For me from Amazon

Date: 6-Apr-2025

Hey AI enthusiast,

This week’s edition feels like AI science fiction—but it’s very real and moving fast. For startup founders, solopreneurs, and tech enthusiasts, like you.

Inside this Edition:

• TikTok’s parent company just dropped Dream Actor MI1, an AI that can play any role on screen.

• Lindy’s AI swarms are changing how agents work—think collaborative bots solving problems together.

• And Amazon’s “Buy for me” Agent shows how shopping is about to get… smarter.

We also look at how 3 Microsoft CEOs speak to Copilot, how deepfakes are being fought back, and the newest creative tools from Runway ML and Midjourney.

Plus as usual, there is stuff that you can learn and use!

If you’re building or just trying to stay ahead with AI news—this issue has what you need.

Best,

Renjit

PS: Founders: In case you want to prototype and test your AI automation idea:

In this edition:

• Amazon’s AI Agent: Shopping Beyond Its Own Shelves

• Animating Humans: DreamActor-M1 from TikTok’s Par …

• Convergence AI’s Parallel Agents: AI That Multitas …

• Lindy AI Introduces AI Swarms

• Learn AI: Guidde generative AI platform that helps …

• India is the largest market for Open AI

• Using AI to fight Deepfakes and getting funded in …

• More useful + fun AI news scoured for you from X.c …

  • Create a full video using MidJourney V7

  • Three Generations of Microsoft CEOS take Copilot for a spin:

  • Runway Gen-4 dropped this week, a video AI that nails character consistency and lifelike motion.

  • Good News from the world of AI:

Amazon’s AI Agent: Shopping Beyond Its Own Shelves

Imagine searching for a unique gadget, only to find it’s not available on Amazon. In the past, this meant scouring other websites yourself. Now, Amazon is testing a feature called “Buy for Me,” designed to handle this task for you.

Here’s how it works:

• Seamless Integration: When Amazon doesn’t stock a product, “Buy for Me” displays options from third-party sites directly within the Amazon app. You can select and purchase these items without leaving the platform.

• AI-Powered Transactions: Utilizing Amazon’s Nova AI models and Anthropic’s Claude, the agent autonomously visits external websites, fills in your shipping and payment details, and completes the purchase on your behalf.

• Security Measures: Amazon ensures that your billing information is securely encrypted during these transactions, maintaining your privacy.

So What?

While this innovation offers convenience, it also raises questions:

• Trust Factor: Relying on AI to make purchases requires confidence in its accuracy and security. Users must trust that the agent will select the correct items and handle their information responsibly.

• Customer Service: For returns or issues, customers will need to interact directly with the third-party seller, which could complicate the support process.

This development reflects a broader trend where companies like OpenAI and Google are exploring AI agents capable of performing tasks such as booking flights and hotels.

As AI continues to evolve, features like “Buy for Me” could redefine online shopping, offering both opportunities and challenges for consumers and businesses alike.

Animating Humans: DreamActor-M1 from TikTok’s Parent Company

Imagine bringing a still photo to life, making it move just like in real life. That’s the promise of DreamActor-M1, a new framework for human image animation. Released by Bytedance (parent company of TikTok).

• Fine-Tuned Movements

DreamActor-M1 captures subtle facial expressions and body motions, ensuring animations look natural and true to the original image.

• Scales with Ease

Whether it’s a close-up portrait or a full-body shot, this framework adapts seamlessly, maintaining quality across different image sizes.

• Consistent Over Time

By integrating motion patterns from sequential frames with visual references, DreamActor-M1 ensures that animations remain coherent, even during complex movements.

• Audio-Driven Lip Syncing

Feed it audio, and watch as it produces lip-syncing animations in multiple languages, making characters speak naturally.

• Handles Unseen Poses

Even if the reference image doesn’t show certain poses, DreamActor-M1 can generate consistent animations, filling in the gaps with ease.

In essence, DreamActor-M1 offers a comprehensive solution for animating human images, blending control, adaptability, and realism. Check out the paper and the output videos here» Link

So what?

Easier for the millions of creators on TikTok to release those dance videos. More content means more Advertising dollars for Bytedance. Potentially paving the way for Bytedance to enter the market of animation and AI powered movies. Oh yes, this is coming sooner that we think!

Convergence AI’s Parallel Agents: AI That Multitasks Like a Team

Most AI agents work solo. Convergence AI changes that with Parallel Agents—a system that splits your task into parts and runs multiple agents at once, each in its own browser.

How it works:

• A planning agent breaks down your prompt

• Proxy launches multiple agents in parallel

• Each handles a specific role in real time

You watch as they scroll, click, and extract—like an AI task force in action.

When finished:

• Proxy gives you a clean, shareable report

• View it in-app or download as a PDF

Try it free for your daily work » Link

So what?

AI agents can now browse websites, pull data, and even complete purchases on their own. But early results show a clear gap between what’s promised and what’s actually delivered, says Venture Beat.

OpenAI’s browser agent, Operator, made waves with demos like ordering pizza or buying event tickets. Still, it’s unclear how developers and businesses will use these tools in real-world settings.

“The killer app hasn’t shown up yet,” says Sam Witteveen, cofounder of Red Dragon, which builds AI agent tools. He believes the value will come from automating tasks most people find tedious—like hunting for the lowest product price or finding the best hotel deal online.

Lindy AI Introduces AI Swarms

Agent swarms let Lindy AI agents duplicate themselves and do 100s of things at once
Integration supremacy: They claim to be the #1 agent in the world with the most integrations, with 5,000+ integrations and 4,000+ web scrapers.

So what?

Businesses and Solopreneurs can take now advantage of multiple AI agents to increase their productivity. Lindy AI is a tool that we use in our AI Automation business to solve problems for our clients, so it is something we can vouch for.

Interested in discussing how we can increase sales and reduce your operating expenses?

Speak to us (book time with us here).

Learn AI: Guidde generative AI platform that helps your team share know-how 10x faster. 🚀

Stop creating training videos manually. Are you used to spend hours creating training videos manually?

I discovered Guidde—a simple tool that automatically transforms any document into a polished video walkthrough with just one click.

Easy for companies to update thier documentation, the videos update too—no extra work needed.

Your team will save countless hours while creating content that people actually prefer using, leading to better retention and performance.

Try Guidde Today»

India is the largest market for Open AI

Open AI released its 4o model with advanced image creation capabilities and the Internet went nuts. We wrote about it last week.

Their COO, Brad Lightcap wrote on X: “It’s been a wild first week for images in ChatGPT — over 130 million users have generated more than 700 million images since last Tuesday.

India is now our fastest-growing ChatGPT market 💪🇮🇳”

I understand from research that Open AI has approached an ARR of $30BN and has 450-500 Million users, with a 5% paid conversion. That is phenomenal!

Using AI to fight Deepfakes and getting funded in the process

Artificial intelligence cybersecurity company Adaptive Security announced last week that it closed a $43 million funding round in March, co-led by OpenAI and Andreessen Horowitz.

• Adaptive Security uses advanced AI to spot deepfakes in videos and audio.

• The company was founded by experts in cybersecurity and artificial intelligence.

• OpenAI’s funding will help Adaptive Security improve its technology and reach more clients.

• The startup plans to work with media companies, social platforms, and governments to fight misinformation.

• Adaptive Security’s tools can be used in real-time, offering quick responses to deepfake content.

So what? With AI comes the risk of misinformation and Deepfakes. Companies like Adaptive use AI to fight AI deepfakes. Read more» Link

More useful + fun AI news scoured for you from X.com

1. Create a full video using MidJourney V7

2. New Robot video from Brett Adcock 😀 

3. Three Generations of Microsoft CEOS take Copilot for a spin:

4. Runway Gen-4 dropped this week, a video AI that nails character consistency and lifelike motion.

   Early tests show it outshines other top models in detail and fluidity.

5. Good News from the world of AI:

UC Berkeley and UCSF have developed an AI that turns brain signals into speech in just one second. This project aims to offer natural-sounding communication for people who’ve lost their voice.

If you want to interact with me on social media, follow me on LinkedIn and Twitter.