Microsoft’s 50-day AI Skills Fest
Plus: Microsoft’s 50-day AI Skills Fest
Read time: 5 minutes
Have you ever noticed the difference between AI-generated videos and the real animated ones? The next big leap in AI avatars is here – combining a face and voice to create a fully interactive talking character.
What are on FIRE 🔥
IN PARTNERSHIP WITH GREAT LEARNING
✨ From Zero to AI Hero in Hours With Great Learning Academy (Get 50% off Today)
Think AI is too complicated to start? That thought is the only thing holding you back.
Let’s change that – today.
Whether you’re a complete beginner or just unsure where to begin, Great Learning Academy’s Premium Courses are built to take you from “I have no idea” to “I got this!” — fast and stress-free.
🎯 What You’ll Get:
-
Comprehensive Curriculum: Crafted by top industry experts to take you from beginner to job-ready with confidence.
-
Real-World Projects: Work on guided, industry-relevant projects that build your portfolio and catch recruiters’ eyes.
-
24/7 AI Mentor Support: Stuck somewhere? Your AI mentor is available anytime to guide you — no delays, no confusion.
-
AI-led Mock Interviews: Practice smart. Get personalized feedback from AI-powered mock interviews that prep you for real opportunities.
-
Interactive Coding Exercises: Master coding with hands-on exercises, real-time AI hints, and instant feedback to help you keep moving forward.
This isn’t just another online course. This is your shortcut to mastering AI – with the right tools, expert support, and job-ready skills that matter.
Limited Time Offer: Courses start at just $40- Now at 50% off
Bonus: Free Preview — No Commitments, Just Click and Explore
AI INSIGHTS
🤖 AI Avatars Escape The Uncanny Valley
AI avatars are stepping up their game! They’re no longer just static images or voiceovers. These avatars can now move, talk, and show emotions in real-time, making them feel more human than ever.
Key Takeaways:
1. The New Big Players in AI Avatars
OmniHuman-1 (Bytedance, February 2025) and Hedra’s Character-3 (March 2025) are leading the charge.
→ These avatars offer full-body movement and talking faces, making them more realistic than anything we’ve seen before.
2. Creating realistic AI avatars involves several elements:
-
Face: The avatar’s face must be expressive and move naturally, matching the speech.
-
Voice: The voice must be appropriate for the character; companies like ElevenLabs provide diverse voice libraries.
-
Lip sync: Syncing lip movements with speech is critical. Tools like Sync and models like MoCha and OmniHuman focus on improving this.
-
Body: Avatars now have full-body motion, though scaling this feature is still in early stages.
-
Background: The avatar’s environment must also match lighting, depth, and interactions (e.g., avatars should be able to pick up products in ads).
-
Real-time interaction: Low-latency streaming is essential for real-time avatars that can join live conversations or video calls.
3. How Are People Using AI Avatars?
-
For Consumers: You can now make personalized content like music videos, AI streamers, or even interactive stories with talking avatars!
-
For SMBs: AI avatars for advertising and product promotions. You can create engaging videos without needing a full production crew.
-
For Enterprises: Companies are using avatars to create training videos, localized content for global teams, and even personalized content with an executive’s likeness.
4. What’s Still a Work in Progress?
-
Consistency & Flexibility: Avatars currently have a fixed look, but the goal is more flexibility to change appearance and behavior.
-
Facial & Body Movements: Some avatars still feel robotic, but companies like Captions and Argil are improving facial expressions and body gestures.
-
Real-World Interactions: While avatars can’t interact with objects yet, this capability is on the horizon.
5. Real-Time Avatars Are On the Horizon
AI avatars could soon be helping with everything from AI doctor consultations to assisting customers in real-time. While there’s still some work to do, platforms like Tolan are making real-time interaction a reality.
It’s exciting to think about how these AI avatars could change how we engage with the digital world.
Why it matters: The magic happens when their faces and bodies sync with their speech. It’s not just the mouth that moves – it’s the entire face and body working together to create a truly lifelike character. And the coolest part? Some of these avatars can even express emotions based on the text you give them.
TODAY IN AI
AI HIGHLIGHTS
👀 You can now show Gemini what you’re talking about by sharing your phone’s camera or screen during a convo — thanks to the new Project Astra capabilities in Gemini Live.
🕵️♂️ Meta Llama 4’s #2 spot on LMArena got cheating calls. It used an experimental version to boost scores over OpenAI’s GPT-4o. LMArena called it out and changed its rules.
🎉 Microsoft just launched a 50-day AI Skills Fest. It’s packed with live events, in-person training, hackathons, and a shot at breaking a Guinness World Record.
🤖 AI is now reading your resume – seriously. Companies are using AI tools to screen, rank, and shortlist candidates. These bots don’t just look for keywords – they can actually infer skills and understand your resume like a recruiter would.
🗣️ Amazon launched Nova Sonic, a new voice model that listens and speaks like a real human. It runs on Bedrock and powers Alexa+, handling live conversations with smooth pauses and natural flow. It’s fast, accurate, and Amazon says it’s 80% cheaper than GPT-4o.
🔔 The new ElevenLabs MCP server gives Claude and Cursor full access to ElevenLabs’ AI voice platform. Just type your prompt – your AI can now talk, generate audio, and even make calls for you.
🚀 LinkedIn’s co-founder asked Replit AI Agent to clone LinkedIn. The AI built a real prototype just from a 2-word prompt. Even Google says 25% of its new code is AI-made.
💰 AI Daily Fundraising: Rescale has raised $115M to expand its AI-driven digital engineering platform, helping industries like aerospace, automotive, and life sciences integrate advanced computing and AI for faster product development and innovation.
AI SOURCES FROM AI FIRE
NEW EMPOWERED AI TOOLS
-
🎙️ The #1 Way to Launch, Host & Grow Your Podcast. Want millions to hear you? We’ve got you — on Spotify, Apple, Amazon, Youtube, Audible & 15+ platforms. Try Hubhopper* For Free. Start for Free.
-
⚡ Runway releases Gen-4 Turbo, creating stunning 10-second videos in just 30 seconds.
-
🔊 OuteTTS-1.0 generates realistic multilingual voice clones in 20 languages.
-
🎓 EverTutor is an AI-powered voice tutor for personalized interactive learning.
-
👨💻 Databutton is an AI developer that builds, codes, and deploys your apps starting at $20/month.
* indicates a promoted tool, if any
AI QUICK HITS
-
📰 US Publishers Push Big Tech to Pay for News Content (Link)
-
🦾 IBM Releases A New Mainframe Z17, Built For Generative AI And AI Agents (Link)
-
🎥 Google Reimagines ‘The Wizard of Oz’ in Fully Immersive 3D With AI (Link)
-
🤖 Dr. Oz Wants AI Avatars to Replace Doctors in CMS (Link)
-
🚀 Mira Murati’s AI Startup Lands Top Ex-OpenAI Advisors (Link)
AI CHART
As mentioned yesterday, we’ll dive deeper into the chapter that received the most votes from the poll. Based on your responses, Chapter 1: AI Research & Development is where the most interest lies. Here are 8 key insights from this chapter to give you a clearer view of the current AI landscape.
-
AI Research Growth
Over the past decade, the number of AI publications has exploded.→ In 2023 alone, AI-related research in computer science jumped from 102,000 papers in 2013 to 242,000, marking a 19.7% increase from the year before.
→ What’s more, AI publications now make up 41.8% of all computer science papers, almost doubling from 21.6% in 2013!
-
Who’s Leading the Pack?
-
China is dominating the global scene, contributing 23.2% of all AI publications and 22.6% of citations in 2023.
-
Meanwhile, the U.S. continues to lead in producing high-impact, influential research, with institutions there still contributing to the most cited AI papers.
-
-
AI Patent Growth
It’s not just research; AI innovation is driving patenting too. Since 2010, the number of AI patents has surged from 3,833 to a massive 122,511 in 2023. China leads the pack, holding 69.7% of these patents! -
U.S. Models Continue to Dominate
In 2024, U.S. institutions produced 40 notable AI models, far outpacing China’s 15 and Europe’s 3. The U.S. has consistently been the leader in AI model development, a trend that’s only growing stronger. -
AI Models: Bigger, Better, and More Expensive
AI models are getting huge—and not just in size. The computational resources required to train these models double approximately every 5 months, and datasets grow every 8 months. All that power comes with a cost, and we’re seeing a massive environmental impact.→ For instance, training GPT-4 (2023) emitted 5,184 tons of carbon, which is about 300 times the carbon footprint of the average American!
-
Lowering the Cost of AI
On the bright side, the cost of using AI models is dropping drastically. The price to query a model like GPT-3.5 (64.8) dropped from $20 per million tokens in 2022 to just $0.07 by 2024—an insane 280-fold reduction in just 18 months! -
AI Hardware Getting Faster & Greener
The machines driving these models are also improving. AI hardware is getting 43% faster each year, while energy efficiency is increasing by 40% annually. This is a huge win for making AI more sustainable. -
The open-source AI community is growing fast.
→ In 2024, U.S. projects on platforms like GitHub led the way, with 23.4% of all AI-related open-source projects.
→ India and Europe also contribute significantly, at 19.9% and 19.5%, respectively.
AI JOBS
We read your emails, comments, and poll replies daily
How would you rate today’s newsletter?Your feedback helps us create the best newsletter possible
|
Hit reply and say Hello – we’d love to hear from you!
Like what you’re reading? Forward it to friends, and they can sign up here.
Cheers,
The AI Fire Team
Leave a Reply