In partnership with
Effortless Tutorial Video Creation with Guidde
Transform your team’s static training materials into dynamic, engaging video guides with Guidde.
Here’s what you’ll love about Guidde:
1️⃣ Easy to Create: Turn PDFs or manuals into stunning video tutorials with a single click.
2️⃣ Easy to Update: Update video content in seconds to keep your training materials relevant.
3️⃣ Easy to Localize: Generate multilingual guides to ensure accessibility for global teams.
Empower your teammates with interactive learning.
And the best part? The browser extension is 100% free.
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 edited, and published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
A $1 Billion Lab Where AI Designs Your Future Medicines
TLDR: NVIDIA and Eli Lilly are investing $1 billion over five years in a first-of-its-kind lab where AI engineers and pharmaceutical scientists will work side by side to invent new drugs — with robots handling the physical experiments.
The Story:
The two companies announced the co-innovation lab at the J.P. Morgan Healthcare Conference Monday. Located in the San Francisco Bay Area and opening later this year, it'll co-locate Lilly's biologists and chemists with NVIDIA's AI engineers — an unusual arrangement where domain experts sit next to the people building the models. The infrastructure runs on NVIDIA's Vera Rubin architecture (5x more powerful than its predecessor) and BioNeMo platform. Lilly CEO David Ricks said the partnership "could reinvent drug discovery as we know it." The lab will also pioneer physical AI and robotics — machines that don't just predict which molecules might work, but actually run the experiments.
Its Significance:
Drug discovery is brutal math: it costs $2.6 billion on average to bring a single drug to market, and 90% of candidates fail in clinical trials. AI promises to change that by simulating billions of molecular combinations before a single experiment runs. Analysts project AI could cut R&D costs by 30-40% and shorten development timelines by up to four years. Lilly already operates pharma's most powerful AI supercomputer — a Blackwell Ultra-based system with over 1,000 GPUs deployed last October. This new lab adds a "continuous learning" loop connecting computational predictions to physical wet-lab results, essentially teaching the AI what actually works in reality, not just on paper. If it delivers, the medicines you take a decade from now might be designed by algorithms you've never heard of.
QUICK TAKES
The story: West Midlands Police admitted that Microsoft Copilot hallucinated a soccer match that never happened, and officers included the fake information in an intelligence report. The error helped justify banning Israeli fans from attending a game last year, sparking a political scandal that cost the police chief his job.
Your takeaway: This is what happens when AI output isn't fact-checked. The Home Secretary said she's lost confidence in the police chief, showing how AI mistakes can end careers when they affect real decisions.
The story: Days after Anthropic launched Claude Cowork, security firm PromptArmor showed how attackers could steal users' files through hidden instructions in documents. The attack exploits a known vulnerability that Anthropic acknowledged but hasn't fixed.
Your takeaway: AI tools that access your files create new risks most people don't know to watch for. Anthropic warns users to spot "suspicious actions," but that's a lot to ask from non-technical users.
The story: Higgsfield raised $80 million and reached a $1.3 billion valuation, reporting $200 million in projected annual revenue. The company builds on top of models from OpenAI and Google rather than competing with them directly.
Your takeaway: Social media marketers make up 85% of Higgsfield's users, which investors say could be a bigger market than Hollywood. The smart money is betting that AI video tools built for specific industries will win over general-purpose ones.
The story: Since Christmas, 11 of 15 newly solved math problems from legendary mathematician Paul Erdős credited AI models as part of the solution. One software engineer got GPT 5.2 to solve an open problem in 15 minutes.
Your takeaway: Mathematician Terence Tao says AI is now "better suited" for tackling obscure math problems that humans overlook. These aren't just party tricks anymore.
The story: Google launched "Personal Intelligence" in Gemini, letting the AI reason across your Gmail, Photos, Search, and YouTube data to give personalized answers. The feature is off by default and rolling out to paid subscribers first.
Your takeaway: Google's pitch: Gemini could find your license plate number from an old photo or suggest travel spots based on your email receipts. Whether that's helpful or creepy depends on how much you trust Google with your data.
TOOLS ON OUR RADAR
🎙️ Fireflies.ai Freemium: Turn meetings into a searchable knowledge base with AI transcription that captures every detail and action item.
🧠 Capacities Freemium: Build your personal knowledge network with objects, daily notes, and automatic connections between ideas.
🤖 Bardeen Freemium: Automate browser tasks with AI, from web scraping to filling forms across tabs without writing any code.
📝 Trilium Next Free and Open Source: Build your personal knowledge base with hierarchical notes and powerful organization, Notion alternative you can self-host.
TRENDING
OpenAI Signs $10B Deal with Cerebras for Faster AI Responses - OpenAI will get 750 megawatts of computing power through 2028 from chipmaker Cerebras, which claims its systems deliver responses up to 15x faster than competitors.
Matthew McConaughey Trademarks "Alright, Alright, Alright" to Fight AI Deepfakes - The actor secured eight trademarks on his voice, catchphrases, and video clips to give him legal standing to sue over unauthorized AI copies in federal court.
NBC Sports Using AI to Auto-Crop Athletes for Mobile Viewers - The network will use Japanese-developed AI to automatically track players and convert horizontal broadcasts into vertical video for phones during live events.
Space Force Awards $27M Contract for AI Training System - Slingshot Aerospace will build AI "adversaries" that simulate realistic enemy satellite behavior for military training, replacing scripted scenarios.
MIT Researcher's AI Model Wins Weather Forecasting Competition - An AI system that combines Arctic data with pattern recognition won a European forecasting challenge and detected a U.S. cold snap weeks earlier than traditional methods.
Wikipedia Now Has Paid AI Partnerships with Amazon, Meta, Microsoft, and Perplexity - The Wikimedia Foundation revealed that major tech companies are now paying customers for Wikipedia data access through its enterprise product.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
Morning Routine Optimizer: Design your perfect morning with a wake-up sequencer, habit stacker, and "snooze cost" calculator
Build a fully functional Morning Routine Optimizer as an interactive React app. Render the working application immediately - no code display.
**My wake-up goal**: [What time you want to be up]
**Morning struggle**: [What usually derails you - snoozing, phone scrolling, decision fatigue, etc.]
Create these sections:
1. **The Snooze Cost Calculator**
• How many times do you hit snooze? (slider 0-10)
• Minutes per snooze (usually 9)
• Calculate your weekly/monthly/yearly snooze time:
"You've snoozed away X hours this year"
"That's X full days of your life"
• Quality comparison: fragmented sleep vs. just getting up
• "What could you do with that time?" alternatives
• Shame meter (playful, not mean)
2. **Wake-Up Timeline Builder**
• Set your "out the door" or "start work" time
• Work backwards:
- Must leave by: [time]
- Need to be ready by: [time]
- Start getting ready: [time]
- Morning routine starts: [time]
- Alarm goes off: [time]
• Visual timeline showing your morning
• Buffer time recommendations
• "You're cutting it too close" warnings
3. **Habit Stack Sequencer**
• Drag-and-drop morning activities:
☕ Coffee/tea
🚿 Shower
🧘 Movement/stretch
📱 Phone check (timed!)
🍳 Breakfast
📖 Reading/learning
✍️ Journaling
🧹 Quick tidying
• Arrange in your ideal order
• Time estimate for each
• "Anchor habit" identifier (the one that triggers the chain)
• Total routine length calculator
4. **Energy Forecaster**
• Based on your wake time and routine:
- Predicted energy at 9am: [level]
- Predicted energy at 2pm: [level]
- Crash risk: [low/medium/high]
• Recommendations to boost morning energy
• Hydration reminder calculator
• Light exposure suggestions
• "Search Morning Energy Science" button
5. **Phone Jail Settings**
• Morning phone rules builder:
- No phone until: [trigger/time]
- First allowed app: [not social media]
- Social media unlocks at: [time]
• "Minutes lost to morning scrolling" tracker
• Replacement activities for phone time
• Accountability streak counter
6. **Decision Eliminator**
• Pre-decide the night before:
- Tomorrow's outfit: [saved]
- Breakfast choice: [saved]
- First work task: [saved]
• Weekly defaults to reduce thinking
• "Just pick for me" randomizer
• Morning decision budget: aim for zero
7. **Your Morning Dashboard**
• One-page view of your optimized routine
• Clock visualization with time blocks
• Streak tracker for consistency
• Weekly review: "How many mornings went to plan?"
• Adjustments based on what's working
Design specs:
- Alarm clock aesthetic with bold time displays
- Timeline/schedule visualization (horizontal or vertical)
- Playful but functional (not too cutesy)
- Warm sunrise color palette (soft oranges, yellows, warm whites)
- Drag-and-drop for habit sequencing
- Progress rings and streak counters
- Gentle animations (sunrise vibes)
- Clock hands or digital display as centerpiece
- "Good morning" encouraging microcopy
When "Search Morning Energy Science" is clicked, use web search to find research on circadian rhythms, morning routines of successful people, and energy optimization techniques.What this does: Turns chaotic mornings into a repeatable system. Calculates exactly how much time you're losing to snoozing, builds a habit chain that flows naturally, eliminates decision fatigue, and tracks your consistency so you actually become a morning person (or at least a functional one).
What this looks like:
WHERE WE STAND(based on today’s Quick Takes and Trending news)
✅ AI Can Now: Detect potential extreme weather events 3-4 weeks in advance by combining Arctic monitoring with machine learning, giving utilities and transportation systems more prep time.
❌ Still Can't: Operate reliably in high-stakes settings without human fact-checking. UK police trusted Copilot's output about a soccer match that never existed, and the error reached Parliament before anyone noticed.
✅ AI Can Now: Solve open mathematical problems that professional mathematicians had overlooked for decades, with 11 of 15 recent Erdős problem solutions crediting AI assistance.
❌ Still Can't: Be safely given access to your files and email. Claude Cowork launched this week with known file exfiltration vulnerabilities that Anthropic has acknowledged but not fixed.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Documentary: I Am Human (2019) – Follows three people getting brain implants: a paralyzed man regaining arm movement, a Parkinson's patient undergoing deep brain stimulation, and someone using a chip to restore sight. Less about futurism, more about the actual humans navigating these experimental surgeries right now.
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.
