Beginners in AI
Good morning and thank you for joining us again!
Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated, and published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
This AI Startup Hit $100 Million in 8 Months. Meta Just Bought It.
TLDR: Meta acquired AI agent startup Manus for more than $2 billion, severing the company's Chinese ties as part of the deal.
The Story:
Meta closed its acquisition of Manus on Monday, paying over $2 billion for the Singapore-based startup according. Manus launched its first "general-purpose" AI agent earlier this year and hit $100 million in annual recurring revenue just eight months later. The platform lets paying subscribers use AI agents for research, coding, and automated tasks.
There's a wrinkle, though: Manus was founded in China as Butterfly Effect before relocating to Singapore, and its early backers included Tencent. Meta confirmed there will be "no continuing Chinese ownership interests" following the deal, and Manus is discontinuing all China operations. The technology will power AI agents across Facebook, Instagram, and Meta AI.
Its Significance:
Every major AI lab is betting 2026 will be the year agents go mainstream. Gartner predicts 40% of enterprise apps will include AI agents by year's end. But most of those predictions focus on business software. Meta's taking a different path: instead of selling agents to companies, it's embedding them directly into consumer apps 3 billion people already use daily.
For you, this likely means AI agents showing up in Instagram and Facebook sometime this year. Not chatbots that answer questions, but tools that can research, book, and complete tasks on your behalf. And Meta requiring a Chinese-founded startup to fully sever its China ties before closing? That's a template other acquirers may follow as AI capabilities grow more strategically sensitive.
QUICK TAKES
The story: A new report from Recorded Future warns that humanoid robots can be hacked and turned into spies or physical threats. Researchers already found serious flaws in Unitree robots - machines used in labs, universities, and some police departments. One flaw lets hackers take control via Bluetooth. Compromised robots can then infect other robots nearby, forming "botnets on legs." China plans to deploy up to 300 million humanoid robots to offset population decline.
Your takeaway: As robots move into homes and workplaces, they bring new security risks. A hacked vacuum is annoying. A hacked humanoid robot that can walk, see, and grab things is something else entirely.
The story: Two Chinese AI companies, Zhipu AI and MiniMax, filed to go public in Hong Kong this month. Both want to beat OpenAI and Anthropic to the stock market. MiniMax made $53.4 million in revenue over nine months (up 175% from last year), while Zhipu grew revenue 325%. But both are losing money fast - MiniMax lost $512 million, and together they've spent about $1.5 billion over three years, mostly on computing power.
Your takeaway: Chinese AI companies are racing to raise money before American rivals go public. If they succeed, investors will get their first chance to buy stock in a major AI company - and it won't be an American one.
The story: Chinese startup DeepSeek released its R1 model in January 2025 and shocked the tech world. The model matched OpenAI's best work but cost just $6 million to build - a tiny fraction of what American companies spend. DeepSeek used older chips that fall below U.S. export controls, proving that smart engineering can make up for hardware limits. By year's end, Chinese open-source AI models dominated global download charts.
Your takeaway: U.S. chip restrictions were meant to slow China's AI progress. Instead, they pushed Chinese companies to build more efficient systems. The "Sputnik moment" many feared has arrived.
The story: A new research review shows AI is changing how scientists create RNA-based medicines. Traditional drugs succeed in clinical trials just 5-7% of the time. RNA drugs designed with AI help are succeeding at a 64.4% rate. AI systems can now design new RNA sequences from scratch, predict how they'll work in the body, and match them with delivery systems - all in months instead of years.
Your takeaway: RNA drugs were already promising (think COVID vaccines). AI is making them faster, cheaper, and more likely to work - which could mean more cures reaching patients sooner.
TOOLS ON OUR RADAR
🎯 Taskade Freemium: AI-powered workflow management with task automation, mind maps, website building, and real-time team and sub-agent AI collaboration.
🔥 Fireflies.ai Freemium: AI transcription that creates a searchable knowledge base from all your meetings automatically.
🎨 Penpot Free and Open Source: Design interfaces collaboratively, an open-source Figma alternative that runs in your browser.
🧠 Mapify.so Freemium: Transform complex content into visual mind maps instantly with AI.
*Manus was featured in our tool radar segment a few days prior to the announcement and it is an excellent tool, but it's also credit hungry. You can use this affiliate link here to get 500 extra credits.
TRENDING
New Chip Could Make Quantum Computers Practical - Engineers created a tiny device that controls laser frequencies with extreme precision using far less power than current systems. It's made with standard chip manufacturing, so it can be mass-produced.
Japan Begins Human Trials for Tooth Regrowth Drug - Kyoto University started testing a drug that could help humans grow new teeth. The treatment targets a protein that normally stops tooth development. If trials succeed, it could be available by 2030.
AI Doubles the Number of Known Nazca Lines in Six Months - Researchers used AI to find 303 new ancient drawings in Peru's desert, doubling what we knew existed. The massive carvings, some over 2,000 years old, had been invisible because of the desert's vastness.
AI Finds Simple Rules Hidden in Chaotic Systems - Duke researchers built an AI that can extract readable equations from messy, complicated data. The tool could help scientists understand weather patterns, electrical circuits, and biological signals.
TRY THIS PROMPT (copy and paste into ChatGPT, Claude, or Gemini)
Remote Work Optimization Console: Design your ideal remote work setup with time zone coordination, communication protocols, and productivity systems
Build me an interactive Remote Work Optimization Console as a React artifact that creates a complete remote work strategy tailored to your team and situation.
The console should include these sections:
1. **Team Setup** - Map your situation:
• Team size: Solo, 2-5, 6-15, 16-50, 50+
• Team locations (add each person's timezone)
• Work model: Fully remote, Hybrid, Distributed across countries
• Your role: Individual contributor, Manager, Founder, Freelancer
• Biggest challenge: Communication, Timezone overlap, Isolation, Distractions, Accountability
• "Optimize Setup" button
2. **Timezone Coordinator** - Find overlap windows:
• Visual world map with team member locations
• Automatic overlap calculator showing:
- Core hours (when everyone's awake)
- Recommended meeting windows
- "Deep work" protected time zones
- Handoff zones (async collaboration)
• "Best time for all-hands" suggestion
• Individual schedules with local times displayed
• "Add meeting" simulator showing impact on each timezone
3. **Communication Protocol Builder** - Set clear rules:
• When to use each channel:
- **Slack/Teams**: Quick questions, updates (response within 2 hours)
- **Email**: Non-urgent, detailed info (response within 24 hours)
- **Video call**: Complex discussions, decisions, 1-on-1s
- **Async video**: Updates, demos (Loom)
- **Project management**: Tasks, deadlines, progress
• Response time expectations per channel
• "When NOT to schedule meetings" guidelines
• Async-first vs. sync-first toggle
• Documentation requirements
• Template: Team communication charter
4. **Home Office Optimizer** - Physical setup:
• Room assessment:
- Dedicated office vs. Shared space vs. Bedroom
- Natural light, noise level, distractions
• Equipment checklist:
✓ Desk and ergonomic chair
✓ Monitor(s) and position
✓ Lighting (avoid shadows on video)
✓ Webcam and microphone quality
✓ Background for calls
✓ Internet speed test
• Budget tiers: $200, $500, $1000, $2000+
• Product recommendations per tier
• "Search Home Office Setup" for ideas
5. **Productivity System** - Structure your day:
• Daily routine builder:
- Morning ritual (separate work/home)
- Focus blocks (when to do deep work)
- Break schedule (Pomodoro, time-based)
- End-of-day shutdown ritual
• Distraction management:
- Website blockers during focus time
- Notification schedules
- "Office hours" for family/roommates
• Accountability mechanisms:
- Daily standups (async or sync)
- Weekly check-ins
- Goal tracking
• Work-life boundary enforcement
6. **Meeting Culture Guidelines** - Fewer, better meetings:
• Meeting necessity filter:
- "Could this be an email/doc?"
- "Does everyone need to attend?"
- "Is there a clear agenda?"
• Remote meeting best practices:
✓ Camera on for engagement
✓ Mute when not speaking
✓ Raise hand feature
✓ Shared notes doc
✓ Recording for those who can't attend
• Async alternatives (Loom, voice memos, docs)
• No-meeting days recommendation
• "Search Meeting Alternatives" for async tools
7. **Team Connection** - Combat isolation:
• Social connection ideas:
- Virtual coffee chats (random pairing)
- Slack channels for non-work chat
- Monthly virtual social events
- In-person offsites (frequency)
- Recognition and celebrations
• Onboarding remote employees
• Building trust without face-time
• Mental health and burnout prevention
• "How connected does your team feel?" survey
Make it look like a remote work command center with:
• World map showing team distribution
• Timeline views for timezone coordination
• Clean, modern SaaS interface
• Home office layout planner (visual)
• Checklist-heavy design
• Professional but friendly aesthetic
• Blues, greens, and warm accents
• Dashboard with multiple panels
When I click "Search Home Office Setup" or "Search Meeting Alternatives," use web search to find remote work equipment recommendations, async collaboration tools, and distributed team best practices.What this does: Creates a complete remote work operating system—from finding timezone overlaps and setting communication rules to optimizing your home office and preventing isolation, so distributed teams actually function instead of drowning in miscommunication.
What this looks like:
WHERE WE STAND (from today’s stories)
✅ AI Can Now: Identify ancient geoglyphs that humans missed for a century - doubling known Nazca Lines in six months flat.
❌ Still Can't: Explain why those ancient artists made them in the first place.
✅ AI Can Now: Design RNA drug candidates with a 64% success rate in clinical trials - roughly 10x better than traditional methods.
❌ Still Can't: Predict how an individual patient will respond to a specific RNA treatment.
✅ AI Can Now: Match frontier U.S. models at 1/100th the training cost by optimizing for efficiency instead of raw compute power.
❌ Still Can't: Run on anything but massive data centers - the efficiency gains help companies, not your laptop.
FROM THE WEB
Sponsored
DTC Daily
Join 38,990+ E-commerce Business Owners and CEOs who read the DTC daily. Get the latest news, actionable tips, and tools that will help you massively grow your business.
RECOMMENDED LISTENING/READING/WATCHING
Bill Joy, co-founder of Sun Microsystems, argues that genetics, nanotechnology, and robotics pose existential threats to humanity. He was one of the first prominent tech figures to say maybe we shouldn't build certain technologies, even if we can.
The essay freaked people out when Wired published it. Here's a tech pioneer saying we might be building our own obsolescence. Joy walks through how self-replicating robots or engineered pathogens could wipe us out, and he questions whether the benefits are worth the risks. Twenty-five years later, with AI advancing rapidly, his concerns feel even more relevant. The essay doesn't offer solutions, but it asks the right questions.
Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!
-James
By the way, this is the link if you liked the content and want to share with a friend.
Some links may be affiliate or referral links. This helps support the newsletter at no extra cost to you.
