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 published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
How High Schoolers Built Real Startups Before Getting Driver's Licenses
TLDR: Teenagers are launching AI-powered startups and making millions before they can get a driver's license, with 15-year-old Nick Dobroshinsky's stock research tool reaching 50,000 monthly users and 17-year-old Raghav Arora's GetASAP raising $3.45 million to deliver supplies to retailers across Southeast Asia.
The Story: Dobroshinsky built BeyondSPX using AI-generated code despite knowing only "about 10 lines of code" himself. He lets ChatGPT handle the technical work while he uses Reddit bots for marketing. Arora started GetASAP at 14, and three years later runs a company with 48 employees that supplies 40,000 outlets including 18 supermarket chains. They deliver products from Unilever to Ben & Jerry's within eight hours. The company now moves 250 shipping containers every year and got one of the highest valuations in Y Combinator's Summer 2025 batch. Sydney's Alby Churven jumped into AI coding at 14 after a Y Combinator video went viral. Tools like ChatGPT collapsed the old barriers. Teens don't need teams, coding skills, or expensive tech to build real companies anymore.
Its Significance: This isn't just teenage hustle. It's a shift in who can build technology companies. When a high schooler can use AI to write code, study markets, and run marketing, the advantage moves from technical skills to speed and knowing what customers want. Venture capitalist Kulveer Taggar sees "an impatience and a mentality of 'I can't wait X amount of years to do it. The opportunity is now.'" If AI tools lower barriers this much for teenagers, they're doing the same for everyone else. The real competition isn't between AI companies. It's between people who move fast with AI tools and those who wait for permission. A Pew Research study found most teens now use generative AI tools daily, not just for homework but for creative and work projects. These teens didn't ask if they were ready. They just started building.
QUICK TAKES
The story: OpenAI said its Atlas AI browser will always face security risks from "prompt injection" attacks, where hidden instructions on websites trick AI into doing harmful things. The UK's National Cyber Security Centre agrees these attacks may never be completely stopped. OpenAI is using new defenses including an AI system that acts as an attacker to find weak spots, but admits the problem is similar to scams on the web—something that will always exist.
Your takeaway: This raises serious questions about whether AI agents can safely browse the internet and handle sensitive tasks like email and banking without human oversight.
The story: Alphabet bought Intersect Power, a company that builds data centers and power plants together, to avoid waiting for local power grids to catch up with AI's massive energy needs. The deal includes projects that will produce 10.8 gigawatts of power by 2028—more than 20 times what the Hoover Dam produces. Alphabet already owned part of the company from a funding round last year.
Your takeaway: Energy is becoming AI's biggest bottleneck, forcing tech giants to build their own power sources instead of relying on existing infrastructure.
The story: Al Jazeera announced "The Core," a complete rebuild of its news operation that puts AI at every stage of journalism—from gathering information to delivering content. The system uses Google's Gemini AI to suggest story angles, generate summaries, translate content, and analyze data. The news network will use six connected AI systems including an AI language model trained on Al Jazeera's archives and tools that generate video content.
Your takeaway: Major news organizations are transforming their entire operations around AI, not just adding AI features, showing how deeply AI is reshaping traditional industries.
The story: The Air Force is ending NIPRGPT, its ChatGPT-like AI tool, on December 31 to make way for the Pentagon's new GenAI.mil system. More than 700,000 military personnel used NIPRGPT since it launched in June 2024, but the Army blocked it from its networks over security concerns. The Air Force says NIPRGPT helped shape the Pentagon's overall AI strategy and provided valuable data on how troops use AI tools.
Your takeaway: Even successful AI tools get replaced quickly as organizations consolidate systems and learn from early experiments, showing how fast the AI landscape is changing.
The story: OpenAI CEO Sam Altman said the next major AI breakthrough will come from systems that remember everything about users over time, not from better reasoning abilities. He said AI is already good at thinking through problems but lacks long-term memory—each conversation starts fresh. Altman described current AI memory as being in the "GPT-2 era" and said meaningful progress on persistent memory could arrive by 2026.
Your takeaway: The focus is shifting from making AI smarter to making it remember more, which could transform AI assistants but raises major privacy concerns about storing personal data.
TOOLS ON OUR RADAR
🎧 Endel
Freemium: Improve focus with personalized soundscapes generated from your biometric data—uses your heart rate, time of day, and weather to create adaptive audio that helps you concentrate or relax.📊 Linear
Freemium: Track issues and sprints with the fastest project management tool designed for modern software teams—keyboard-first interface and beautiful design that developers actually love using.🎨 Splat
Freemium: Turn any photo into a coloring page instantly—perfect for parents, teachers, and anyone who wants to transform memories into creative activities for kids.🔐 KeePassXC
Free and Open Source: Manage passwords offline with strong encryption and no cloud dependency—your password vault stays on your device where hackers can't reach it.👓 Ray-Ban Meta Smart Glasses Gen 2
Paid: Take photos and videos hands-free with built-in 12MP camera and Meta AI. Make calls, listen to music, and capture POV content without pulling out your phone.
TRENDING
Instacart Ends AI Pricing Experiments After Customer Complaints — Instacart stopped using AI to show different prices to different customers for the same items after consumer groups said it inflated grocery bills. Tests showed the same dozen eggs ranged from $3.99 to $4.79 depending on who was shopping.
Wearable AI Device Detects Early Signs of Frailty in Older Adults — University of Arizona researchers created a soft sleeve worn on the thigh that uses AI to detect frailty warning signs by monitoring leg movement. The device analyzes data on the spot and sends only results via Bluetooth, cutting data transmission by 99 percent.
Zoom Launches AI Assistant That Works in Your Browser — Zoom released AI Companion 3.0, which brings its AI assistant to web browsers and adds automation features. The system can search through past meetings and connected apps like Google Drive, and includes a drag-and-drop tool for building AI workflows.
AI Security System Stops Package Thieves in Real Time — Deep Sentinel uses AI-powered cameras paired with live guards to catch package thieves. When the AI spots suspicious behavior, it can speak to the person through a speaker or alert a guard who warns them the police are being called.
TRY THIS PROMPT (copy and paste into ChatGPT, Claude, or Gemini)
Pitch Deck Outliner: Structure investor-ready presentations with proven narrative flow and objection handling
**💡 Try This Prompt: Pitch Deck Outliner**
**Structure investor-ready presentations with proven narrative flow and objection handling**
**Copy and paste into Claude:**
```
Build me an interactive Pitch Deck Outliner as a React artifact that helps founders and sales teams create compelling presentations with the right structure and content.
The console should include these sections:
1. **Pitch Setup** - Define your context:
• Pitch type selector:
- Seed funding
- Series A/B/C
- Sales pitch (B2B)
- Partnership pitch
- Board presentation
- Demo day (3 min)
• Target audience: VCs, Angels, Customers, Partners, Board
• Time allotted: 3 min, 10 min, 20 min, 45 min
• Stage: Pre-revenue, Early traction, Scaling, Established
• "Load Template" button
2. **Slide Sequence Builder** - Optimal flow:
• Visual slide sorter (drag to reorder)
• Recommended slide order for your pitch type:
1. Cover (Company name + one-liner)
2. Problem (pain point you're solving)
3. Solution (your product/approach)
4. Why Now (timing/market forces)
5. Market Size (TAM/SAM/SOM)
6. Product (demo or screenshots)
7. Traction (metrics that matter)
8. Business Model (how you make money)
9. Competition (landscape + positioning)
10. Go-to-Market (customer acquisition)
11. Team (why you'll win)
12. Financials (projections)
13. Ask ($ amount + use of funds)
• Optional slides: Vision, Partnerships, Press, Risks
• Slide count recommendation (10-15 for investors)
• "Add slide" / "Remove slide" controls
3. **Slide Content Guide** - What to include:
• For each slide, expand to see:
- **Purpose**: What this slide accomplishes
- **Key elements**: Must-have content
- **Common mistakes**: What to avoid
- **Best practices**: Pro tips
- **Example**: Visual reference
• Content checklist per slide
• "Too much text" warning (6 words per line max)
• Completeness indicator (essential items filled?)
4. **Narrative Arc Builder** - Story flow:
• Tension curve showing emotional journey:
- Opening hook (grab attention)
- Rising tension (problem gets worse)
- Turning point (your solution)
- Climax (traction/proof)
- Resolution (the ask)
• Slide placement on curve
• Pacing suggestions (time per slide)
• Transition guidance between slides
• "Make it compelling" storytelling tips
5. **Objection Handler** - Anticipate pushback:
• Common objections by slide:
- Problem: "Is this really a problem?"
- Market: "Market too small/crowded"
- Product: "Why will people switch?"
- Traction: "Not enough growth"
- Team: "Missing key roles"
- Financials: "Unrealistic projections"
• Recommended responses for each
• Where to address proactively in deck
• Backup slides to have ready
• "Search Similar Pitches" for examples
6. **Metrics Dashboard** - What to highlight:
• Traction metrics by stage:
- **Pre-revenue**: Users, engagement, waitlist
- **Early**: MRR, growth rate, retention
- **Scaling**: ARR, LTV:CAC, gross margin
• Which metrics matter most for your pitch
• How to visualize them (graph types)
• Vanity vs. meaningful metrics
• "Red flag" metrics to avoid or explain
7. **Competitive Positioning** - Stand out:
• 2x2 matrix builder for competition slide:
- X-axis: Choose dimension (price, features, speed)
- Y-axis: Choose dimension (quality, ease, scale)
- Plot competitors
- Show your unique position
• Alternative: Feature comparison table
• "Why us?" differentiation statement
• Defensibility factors (moat, network effects)
8. **Financial Story** - Numbers that convince:
• Projection builder (3-5 years):
- Revenue assumptions
- Cost structure
- Key drivers
- Growth trajectory
• Funding ask calculator:
- Amount needed
- Use of funds breakdown
- Runway extension
- Milestones to hit
• Unit economics display
• "Does this make sense?" sanity check
9. **Design Checklist** - Presentation polish:
• Visual hierarchy review:
✓ One message per slide
✓ Large, readable fonts
✓ High contrast
✓ Consistent branding
✓ Minimal text
✓ Professional images (no clip art)
• Color scheme suggestion
• Template recommendations
• "Search Pitch Deck Examples" for inspiration
10. **Practice Mode** - Rehearse delivery:
• Slide-by-slide talking points
• Time per slide tracker
• "What will they ask here?" prompts
• Smooth transitions written out
• Backup slide triggers (when to show)
• Export as speaker notes
Make it look like a professional slide editor with:
• Slide sorter view (thumbnail grid)
• Clean, business-focused design
• Preview pane showing slide layout
• Professional color scheme (navy, gray, accent colors)
• Drag-and-drop slide reordering
• Progress indicators (deck completeness)
• Timeline view for narrative arc
• Checklist styling throughout
• Export-ready formatting
When I click "Search Similar Pitches" or "Search Pitch Deck Examples," use web search to find successful pitch decks in your industry, pitch deck best practices, and investor feedback on common mistakes.
```
**What this does:** Guides you through building an investor-ready pitch deck with proven structure, anticipates objections, helps you choose the right metrics to highlight, and ensures your narrative arc keeps audiences engaged—so you don't waste your 15 minutes with investors.What this does: Guides you through building an investor-ready pitch deck with proven structure, anticipates objections, helps you choose the right metrics to highlight, and ensures your narrative arc keeps audiences engaged—so you don't waste your 15 minutes with investors.
What this looks like:
WHERE WE STAND (based on today’s stories)
✅ AI Can Now: Detect security threats before they're used in attacks by having AI systems act as attackers to find their own weaknesses.
❌ Still Can't: Completely stop prompt injection attacks that trick AI browsers into following hidden malicious instructions, even with the best defenses.
✅ AI Can Now: Remember conversations and preferences within a single chat session to provide context during that interaction.
❌ Still Can't: Build true long-term memory that recalls everything across a user's entire lifetime of interactions the way humans naturally remember.
✅ AI Can Now: Run powerful language models on regular laptops and personal devices instead of requiring cloud servers and constant internet access.
❌ Still Can't: Match the speed and capabilities of cloud-based AI when running the largest, most advanced models on local hardware alone.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
DeepMind's team walks through how they built AlphaFold, the AI that cracked protein folding—a problem that stumped scientists for half a century. John Jumper, who won the Nobel Prize for this breakthrough, explains the technical challenges and what it means for drug discovery and biology. It's AI doing something genuinely useful for science, told by the people who contributed to the field.
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 links. This helps support the newsletter at no extra cost to you.
