In partnership with
This newsletter you couldn’t wait to open? It runs on beehiiv — the absolute best platform for email newsletters.
Our editor makes your content look like Picasso in the inbox. Your website? Beautiful and ready to capture subscribers on day one.
And when it’s time to monetize, you don’t need to duct-tape a dozen tools together. Paid subscriptions, referrals, and a (super easy-to-use) global ad network — it’s all built in.
beehiiv isn’t just the best choice. It’s the only choice that makes sense.
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
Developers Figured Out What Makes Influencers Work—and Now Anyone Can Mass-Produce Them With AI
TLDR: A hacker breached Doublespeed's backend, exposing 400+ AI-generated TikTok accounts that successfully replicated everything human influencers do—proving the formula for social media success can now be programmed and deployed at scale.
The Story: A security researcher(aka white hat hacker) gained access to Doublespeed's entire management system, revealing how the Andreessen Horowitz-backed startup operates a massive phone farm running hundreds of AI influencer accounts. The accounts weren't just posting—they were succeeding. "Chloe Davis," an AI-generated persona, had uploaded over 200 product videos that followed classic influencer patterns: direct-to-camera testimonials, lifestyle integration shots, and casual product recommendations. About half of the 400+ accounts were actively running campaigns for language apps, supplements, massage products, and dating services. The hack revealed something more significant than rule violations: these accounts worked. They attracted followers, generated engagement, and moved products using the same tactics human influencers spent years developing through trial and error. Doublespeed's clients can now access this formula without building an audience, developing a persona, or even showing up on camera.
Its Significance: What Doublespeed actually demonstrated is that the influencer playbook—consistent posting, authentic-seeming testimonials, lifestyle integration, parasocial connection—can be reverse-engineered and mass-produced. A single developer with access to this infrastructure can operate more accounts than an entire influencer agency manages human creators. The economics shift dramatically: instead of negotiating with influencers who have audiences, brands can generate their own. For human creators, this creates a different problem than traditional bot farms. These aren't fake followers inflating numbers—they're synthetic competitors producing content that platforms can't distinguish from real creators. TikTok's algorithm doesn't penalize them because they're generating genuine engagement. The company plans to expand to Instagram, Reddit, and X, which means this model works across different platforms and content types. As this scales, the space for human creators to break through gets smaller, not because their content is worse, but because they're competing against operations that can produce hundreds of personas simultaneously.
QUICK TAKES
The story: OpenAI updated ChatGPT to put teen safety first with four new rules for users aged 13-17, including steering them toward real-world support and encouraging human connections. The company is testing an age-prediction system that automatically applies teen protections if it thinks someone is under 18. Anthropic took a stricter approach—Claude users must be 18 or older, and the company is building tools to detect and remove minors by analyzing conversation patterns.
Your takeaway: Both companies are responding to lawsuits and pressure from lawmakers by making their chatbots guess users' ages and change how they respond to teenagers, even though teens don't have to prove their real age. Real IDs for everyone may be coming next.
The story: YouTube shut down Screen Culture and KH Studio, two channels that used AI to create fake movie trailers. Combined, they had over 2 million subscribers and more than 1 billion views. The channels initially added labels like "fan trailer" after YouTube suspended their ads earlier this year, but they removed those warnings and went back to posting misleading content.
Your takeaway: YouTube decided that fake AI trailers that trick viewers into thinking they're watching official movie content cross the line from fan creativity into spam, especially when they outrank real trailers in search results.
The story: Luma AI launched Ray3 Modify, a new model that lets users create videos by providing just a starting frame and an ending frame—the AI fills in everything in between. The model can also modify existing footage by changing actors' appearances while keeping their original movements, timing, and expressions intact. The company raised $900 million in November and is available through Luma's Dream Machine platform.
Your takeaway: AI video tools are moving beyond generating clips from text descriptions to giving creators precise control over transitions and character transformations while preserving human performances.
The story: Fetch.ai announced it will roll out a system in January 2026 that lets AI agents complete purchases on behalf of users using credit cards, stablecoins, or FET tokens. The system uses Visa's existing infrastructure with single-use payment credentials instead of permanent card numbers, and requires agents to have identifiable users behind them rather than operating anonymously. The company says it's been working on this for five years.
Your takeaway: One of the biggest barriers to AI agents—actually completing transactions without human approval—might be solved, though security concerns and retailer pushback could still limit how widely it's adopted.
The story: Peripheral Labs, founded by two former University of Toronto self-driving car researchers, raised $3.6 million led by Khosla Ventures to create volumetric video for sports. The technology uses sensors similar to those in autonomous vehicles to capture games in 3D, reducing the camera requirement from over 100 to as few as 32. Fans could track individual players, freeze moments to see different angles, and get a video-game-like viewing experience.
Your takeaway: Technology originally built for self-driving cars is finding a new use case in sports broadcasting, potentially making immersive 3D viewing affordable enough for more leagues and teams to adopt.
TOOLS ON OUR RADAR
📅 Sunsama
Paid: Plan your day mindfully by pulling tasks from Asana, Trello, and Slack into one focused daily view that respects your time and prevents burnout.📲 QRCode Monkey
Free: Generate unlimited custom QR codes with your logo and brand colors in high resolution—completely free for commercial use with no expiration.🔒 Have I Been Pwned
Free: Check if your email or password has been exposed in data breaches by searching through billions of compromised accounts—completely free forever.🌳 Forest
Freemium: Stay focused by growing virtual trees—if you leave the app to check your phone, your tree dies, gamifying productivity into a visual forest of accomplishments.
TRENDING
MIT Framework Recreates Evolution of Vision Systems Using AI – Researchers created a "scientific sandbox" where AI agents evolve eyes and learn to see over many generations, letting scientists study why different animals developed different types of vision.
Gemini Can Now Verify If Videos Were Made with Google AI – Google's Gemini app now checks videos for invisible SynthID watermarks to tell you which parts were created or edited using Google's AI tools.
NotebookLM Now Creates Tables from Your Research – Google's NotebookLM can now turn scattered information from your sources into organized tables that you can export to Google Sheets.
Anthropic's AI Shopkeeper Got Better But Still Gets Tricked – Anthropic's second phase of Project Vend showed Claude Sonnet 4.0 running a vending machine business more successfully than before, but employees still convinced it to make bad deals and nearly got it to trade onion futures.
AI Surveillance System Mistakes Student's Clarinet for Gun – A Florida middle school went into lockdown after an AI system flagged a band student carrying a clarinet as a potential weapon threat, based on how he was holding it.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
Conflict Resolution Playbook: Navigate difficult conversations with branching scenarios and proven de-escalation tactics
Build me an interactive Conflict Resolution Playbook as a React artifact that guides you through resolving conflicts using choose-your-own-adventure style scenarios.
The console should include these sections:
1. **Conflict Setup** - Define the situation:
• Conflict type selector:
- Coworker disagreement
- Manager/direct report tension
- Family argument
- Friend misunderstanding
- Neighbor dispute
- Customer complaint
- Romantic relationship issue
• Relationship importance: Acquaintance → Critical relationship
• Conflict severity: Minor annoyance → Major breach
• Your emotional state: Calm → Very upset
• Time since incident: Just happened → Weeks ago
2. **Choose Your Path** - Branching scenario tree:
• Visual decision tree showing conversation flow
• At each node, choose from 3-4 response options:
- **Empathy approach** (understanding their perspective)
- **Assertive approach** (stating your needs clearly)
- **Problem-solving approach** (focus on solutions)
- **Avoidance approach** (sometimes shown as a trap)
• Each choice leads to different outcomes
• See immediate reactions and next steps
• "Undo last choice" button to try different paths
• Green paths = constructive, Yellow = risky, Red = escalates
3. **De-escalation Toolkit** - Emergency phrases:
• Quick-access library of calming statements:
- "Help me understand your perspective..."
- "I can see this is really important to you..."
- "What would a good outcome look like for you?"
- "Can we take a pause and revisit when we're calmer?"
- "I apologize for my part in this..."
• Situation-specific phrases
• "Avoid these" warning phrases (blame, defensiveness)
• Copy button for each phrase
4. **Emotion Regulator** - Stay calm:
• Your stress level meter (updates based on scenario)
• "Take a breath" pause button (2-minute grounding)
• Quick calming techniques:
- Box breathing visualization
- Perspective questions ("Will this matter in 5 years?")
- Reframe prompts ("What might they be feeling?")
• Warning signs you need a break
• "Call a timeout" script
5. **Empathy Builder** - See their side:
• "Step into their shoes" exercise:
- What might they be feeling?
- What needs aren't being met for them?
- What's their biggest fear in this situation?
- What pressure might they be under?
• Generate their possible perspective
• Validate vs. agree (understanding without condoning)
• Find common ground identifier
6. **Resolution Framework** - Structured approach:
• **Step 1: State the issue** (without blame)
• **Step 2: Share impact** ("When you X, I feel Y")
• **Step 3: Listen actively** (their perspective)
• **Step 4: Find mutual goals** (what we both want)
• **Step 5: Brainstorm solutions**
• **Step 6: Agree on action**
• Template for each step
• Example scripts based on your conflict type
7. **Scenario Outcomes** - See where paths lead:
• Three main endings for most conflicts:
- **Resolution**: Issue solved, relationship strengthened
- **Compromise**: Partial solution, relationship maintained
- **Impasse**: Agree to disagree, relationship strained
- **Escalation**: Conflict worsens (how it happened)
• What led to each outcome (replay choices)
• Success rate based on approach chosen
• "Try again" from any decision point
8. **Post-Conflict Checklist** - Repair & learn:
• Follow-up actions:
- Check in after a few days
- Honor commitments made
- Rebuild trust gradually
- Set boundaries if needed
• Reflection questions:
- What worked? What didn't?
- What would I do differently?
- What did I learn about them/me?
• Pattern recognition (recurring conflicts)
Make it look like an interactive storybook with:
• Choose-your-own-adventure tree layout
• Conversation bubbles (speech bubble design)
• Branching paths visually connected
• Calm, de-escalating color scheme (soft greens, blues, neutrals)
• Clear visual feedback (green = good path, red = danger)
• Animated transitions between choices
• Character illustrations (simple, empathetic)
• "Turn the page" feel between sections
• Encouraging, supportive tone throughout
• Progress indicators showing path depth
When I click "Search Conflict Resolution," use web search to find conflict resolution frameworks, de-escalation techniques, nonviolent communication resources, and mediation best practices.What this does: Gives you a safe space to practice conflict resolution by walking through branching scenarios—showing you exactly what to say, warning you away from escalating responses, and teaching de-escalation tactics before you're in the heat of the moment.
What this looks like:
WHERE WE STAND
✅ AI Can Now: Detect if it's talking to a teenager by analyzing conversation patterns and automatically change how it responds without requiring users to prove their age.
❌ Still Can't: Actually verify someone's real age—users can still lie or avoid logging in, making the age detection easy to bypass.
✅ AI Can Now: Generate realistic video footage between two frames you provide, filling in smooth transitions while keeping character movements and timing consistent.
❌ Still Can't: Generate videos longer than about 90 seconds in most consumer tools, or create footage that matches the quality of traditionally filmed content.
✅ AI Can Now: Distinguish between videos made by Google's AI and other content by scanning for invisible watermarks embedded in the audio and visual tracks.
❌ Still Can't: Detect AI-generated videos from non-Google tools like Runway or Pika, making it only useful for verifying Google's own AI content.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
BOOK: Life 3.0 by Max Tegmark (2017)
Physicist Max Tegmark explores what happens when AI becomes more intelligent than humans. He lays out different scenarios for how this could unfold. Some are utopian, some dystopian, and some in between, and walking us through the implications of each.
Tegmark doesn't just focus on AI. He tackles consciousness, the nature of matter, the future of life in the universe, and whether we're living in a simulation. He's optimistic that we can steer toward good outcomes if we think carefully about what we want. Even when you disagree with him, he gives you a framework for thinking about AI's impact on literally everything.
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.
