feat: update demos

frontend/public/demo/threads/3823e443-4e2b-4679-b496-a9506eae462b/user-data/outputs/fei-fei-li-podcast-timeline.md

@@ -0,0 +1,176 @@
+# Dr. Fei-Fei Li: Recent Podcast Appearances Timeline (Last 6 Months)
+
+## Overview
+Dr. Fei-Fei Li, often called the "Godmother of AI," has been actively appearing on major podcasts discussing the future of artificial intelligence, spatial intelligence, human-centered AI, and her work at World Labs. This timeline compiles key highlights from her recent podcast appearances from August 2025 to January 2026.
+
+---
+
+## Timeline of Recent Podcast Appearances
+
+### January 15, 2025 - **Possible Podcast** (with Reid Hoffman and Aria Finger)
+**Episode:** "Fei-Fei Li on spatial intelligence and human-centered AI"
+
+**Key Highlights:**
+- **Spatial Intelligence as Next Frontier:** Emphasized that spatial intelligence represents the next major evolution beyond large language models (LLMs)
+- **Human-Centered AI Philosophy:** Discussed the importance of building AI that amplifies human potential rather than replacing humans
+- **Regulatory Guardrails:** Addressed the need for thoughtful regulation and governance frameworks for AI development
+- **World Labs Mission:** Explained her current role as co-founder and CEO of World Labs, focusing on spatial intelligence technology
+- **ImageNet Legacy:** Reflected on how ImageNet revolutionized computer vision and sparked the deep learning revolution
+
+**Notable Quote:** "Humans are capable of creating God-like technology so that we can improve our medieval institutions and raise above our paleolithic emotions."
+
+---
+
+### August 15, 2025 - **Firing Line (PBS)**
+**Episode:** "Fei-Fei Li on ethical AI development"
+
+**Key Highlights:**
+- **Ethical AI Development:** Discussed the challenges and responsibilities in developing AI ethically
+- **Societal Impact:** Addressed how AI will transform various sectors including healthcare, education, and employment
+- **Policy Recommendations:** Provided insights on what policy frameworks are needed for responsible AI deployment
+- **Global Collaboration:** Emphasized the need for international cooperation on AI standards and safety
+
+---
+
+### November 16, 2025 - **Lenny's Podcast**
+**Episode:** "The Godmother of AI on jobs, robots & why world models are next"
+
+**Key Highlights:**
+- **World Models Introduction:** Explained why world models and spatial intelligence represent the next frontier beyond LLMs
+- **AI Won't Replace Humans:** Argued that AI won't replace humans but will require us to take responsibility for ourselves
+- **Marble Applications:** Revealed surprising applications of World Labs' Marble product, from movie production to psychological research
+- **Robotics Challenges:** Discussed why robotics faces unique challenges compared with language models
+- **Historical Context:** Shared rarely told history of AI development, including that just nine years ago, calling yourself an AI company was "basically a death sentence"
+- **Participation for All:** Explained how anyone can participate in AI regardless of their role or background
+
+**Key Discussion Points:**
+1. How ImageNet helped spark the current AI explosion
+2. The "bitter lesson" in AI and robotics
+3. Applications of Marble in creative industries and therapy
+4. Human-centered AI initiatives at Stanford
+
+---
+
+### November 25, 2025 - **Masters of Scale Summit**
+**Episode:** "The 'Godmother of AI' on the next phase of AI" (with Reid Hoffman)
+
+**Key Highlights:**
+- **Fearless Approach:** Discussed why scientists and entrepreneurs need to be fearless in the face of an uncertain AI future
+- **Spatial Intelligence & World Modeling:** Detailed the next phase of AI focusing on spatial understanding
+- **Trust Building:** Explained how leaders should build societal trust in AI products and companies
+- **Human Agency:** Emphasized that trust cannot be outsourced to machines and must remain fundamentally human
+- **Entrepreneurial Responsibility:** Argued that entrepreneurs should care about trust from day one of AI development
+
+**Chapter Topics Covered:**
+- The next phase of AI: spatial intelligence & world modeling
+- What spatial intelligence has done for humans
+- Whether AI is over-hyped
+- How to build society trust in AI
+- Why we need to be "fearless" with AI
+
+---
+
+### December 9, 2025 - **The Tim Ferriss Show** (#839)
+**Episode:** "Dr. Fei-Fei Li, The Godmother of AI — Asking Audacious Questions, Civilizational Technology, and Finding Your North Star"
+
+**Key Highlights:**
+- **Civilizational Technology:** Defined AI as a "civilizational technology" that will have profound economic, social, cultural, and political impacts
+- **Personal Journey:** Shared her immigrant story from Chengdu to New Jersey, and her family's seven years running a dry cleaning shop while she attended Princeton
+- **ImageNet Creation:** Detailed the creation of ImageNet and how it birthed modern AI, including innovative use of Amazon Mechanical Turk for data labeling
+- **Spatial Intelligence Vision:** Explained why she founded World Labs to focus on spatial intelligence as the next frontier
+- **Educational Philosophy:** Proposed rethinking evaluation by showing students AI's "B-minus" work and challenging them to beat it
+- **Human-Centered Focus:** Emphasized that "people are at the heart of everything" in AI development
+
+**Notable Quotes:**
+- "Really, at the end of the day, people are at the heart of everything. People made AI, people will be using AI, people will be impacted by AI, and people should have a say in AI."
+- "AI is absolutely a civilizational technology... it'll have—or [is] already having—a profound impact in the economic, social, cultural, political, downstream effects of our society."
+- "What is your North Star?"
+
+**Key Topics Discussed:**
+- From fighter jets to physics to asking "What is intelligence?"
+- The epiphany everyone missed: Big data as the hidden hypothesis
+- Against the single-genius myth: Science as non-linear lineage
+- Quality control puzzles in AI training data
+- Medieval French towns on a budget: How World Labs serves high school theater
+- Flight simulators for robots and strawberry field therapy for OCD
+
+---
+
+### June 16, 2025 - **Y Combinator Startup Podcast**
+**Episode:** "Fei-Fei Li - Spatial Intelligence is the Next Frontier in AI"
+
+**Key Highlights:**
+- **Startup Perspective:** Provided insights for AI startups on navigating the current landscape
+- **Technical Deep Dive:** Offered detailed explanations of spatial intelligence technologies
+- **Entrepreneurial Advice:** Shared lessons from transitioning from academia to entrepreneurship
+- **Market Opportunities:** Identified emerging opportunities in spatial AI applications
+
+---
+
+## Common Themes Across Recent Appearances
+
+### 1. **Spatial Intelligence as the Next Frontier**
+- Repeated emphasis that spatial intelligence represents the next major evolution beyond language models
+- World Labs' focus on creating AI that understands and interacts with the physical world
+- Applications ranging from robotics and autonomous systems to creative industries and therapy
+
+### 2. **Human-Centered AI Philosophy**
+- Consistent message that AI should augment rather than replace human capabilities
+- Emphasis on maintaining human agency and responsibility in AI systems
+- Focus on building trust and ethical frameworks
+
+### 3. **Educational Transformation**
+- Advocacy for integrating AI into education to enhance learning
+- Proposal to use AI as a benchmark for student improvement
+- Emphasis on making AI accessible to people from all backgrounds
+
+### 4. **Historical Perspective**
+- Frequent references to ImageNet's role in sparking the deep learning revolution
+- Context about how rapidly the AI landscape has changed
+- Emphasis on collaborative, non-linear progress in scientific advancement
+
+### 5. **Entrepreneurial Vision**
+- Insights on building AI companies in the current environment
+- Balance between technological innovation and responsible development
+- Focus on practical applications that solve real-world problems
+
+---
+
+## Key Insights and Predictions
+
+### **Near-Term Developments (1-3 years):**
+- Rapid advancement in spatial intelligence and world modeling technologies
+- Increased integration of AI in education and creative industries
+- Growing focus on AI ethics and governance frameworks
+- Expansion of practical applications in healthcare, therapy, and accessibility
+
+### **Medium-Term Vision (3-5 years):**
+- More sophisticated human-AI collaboration systems
+- Breakthroughs in robotics enabled by spatial intelligence
+- Transformation of how we teach and learn with AI assistance
+- Development of new industries centered around spatial AI
+
+### **Long-Term Philosophy:**
+- AI as a "civilizational technology" that requires thoughtful stewardship
+- Emphasis on maintaining human values and agency in technological progress
+- Vision of technology that helps humanity "raise above our paleolithic emotions"
+
+---
+
+## Conclusion
+
+Dr. Fei-Fei Li's recent podcast appearances reveal a consistent vision: AI should be developed as a human-centered technology that enhances our capabilities rather than replacing them. Her focus has shifted from the foundational work of ImageNet to the next frontier of spatial intelligence through World Labs. Across all conversations, she emphasizes the importance of ethics, education, and maintaining human agency in the age of artificial intelligence.
+
+The timeline shows her evolving role from academic researcher to entrepreneur while maintaining her core philosophy that technology should serve humanity's best interests. Her message remains one of cautious optimism, emphasizing both the tremendous potential and significant responsibilities of developing transformative AI technologies.
+
+---
+
+## Sources
+1. The Tim Ferriss Show (December 9, 2025)
+2. Lenny's Podcast (November 16, 2025)
+3. Masters of Scale Summit (November 25, 2025)
+4. Possible Podcast (January 15, 2025)
+5. Firing Line, PBS (August 15, 2025)
+6. Y Combinator Startup Podcast (June 16, 2025)
+
+*Compiled on January 25, 2026*

frontend/public/demo/threads/d3e5adaf-084c-4dd5-9d29-94f1d6bccd98/user-data/outputs/diana_hu_research.md

@@ -0,0 +1,332 @@
+# Diana Hu: Technical Startup Founder Advice - Comprehensive Research
+
+## Video Overview
+**Title:** Tips For Technical Startup Founders | Startup School  
+**Speaker:** Diana Hu, Y Combinator Group Partner  
+**Date:** April 21, 2023  
+**Length:** 28 minutes  
+**YouTube URL:** https://www.youtube.com/watch?v=rP7bpYsfa6Q
+
+## Speaker Background
+
+### Education
+- **BS and MS in Electrical and Computer Engineering** from Carnegie Mellon University
+- Focus on **computer vision and machine learning**
+- Originally from Chile
+
+### Career Path
+1. **Co-founder & CTO of Escher Reality** (YC S17)
+   - Startup building augmented reality SDK for game developers
+   - Company acquired by Niantic (makers of Pokémon Go) in February 2018
+
+2. **Director of Engineering at Niantic**
+   - Headed AR platform after acquisition
+   - Responsible for scaling AR infrastructure to millions of users
+
+3. **Group Partner at Y Combinator** (Current)
+   - Has conducted **over 1,700 office hours** across 5 batches
+   - Advises top YC alumni companies
+   - Specializes in technical founder guidance
+
+### Key Achievements
+- Successfully built and sold AR startup to Niantic
+- Scaled systems from prototype to millions of users
+- Extensive experience mentoring technical founders
+
+## Escher Reality Acquisition
+- **Founded:** 2016
+- **Y Combinator Batch:** Summer 2017 (S17)
+- **Product:** Augmented Reality backend/SDK for cross-platform mobile AR
+- **Acquisition:** February 1, 2018 by Niantic
+- **Terms:** Undisclosed, but both co-founders (Ross Finman and Diana Hu) joined Niantic
+- **Technology:** Persistent, cross-platform, multi-user AR experiences
+- **Impact:** Accelerated Niantic's work on planet-scale AR platform
+
+## Video Content Analysis
+
+### Three Stages of Technical Founder Journey
+
+#### Stage 1: Ideating (0:00-8:30)
+**Goal:** Build a prototype as soon as possible (matter of days)
+
+**Key Principles:**
+- Build something to show/demo to users
+- Doesn't have to work fully
+- CEO co-founder should be finding users to show prototype
+
+**Examples:**
+1. **Optimizely** (YC W10)
+   - Built prototype in couple of days
+   - JavaScript file on S3 for A/B testing
+   - Manual execution via Chrome console
+
+2. **Escher Reality** (Diana's company)
+   - Computer vision algorithms on phones
+   - Demo completed in few weeks
+   - Visual demo easier than explaining
+
+3. **Remora** (YC W21)
+   - Carbon capture for semi-trucks
+   - Used 3D renderings to show promise
+   - Enough to get users excited despite hard tech
+
+**Common Mistakes:**
+- Overbuilding at this stage
+- Not talking/listening to users soon enough
+- Getting too attached to initial ideas
+
+#### Stage 2: Building MVP (8:30-19:43)
+**Goal:** Build to launch quickly (weeks, not months)
+
+**Key Principles:**
+
+1. **Do Things That Don't Scale** (Paul Graham)
+   - Manual onboarding (editing database directly)
+   - Founders processing requests manually
+   - Example: Stripe founders filling bank forms manually
+
+2. **Create 90/10 Solution** (Paul Buchheit)
+   - Get 90% of value with 10% of effort
+   - Restrict product to limited dimensions
+   - Push features to post-launch
+
+3. **Choose Tech for Iteration Speed**
+   - Balance product needs with personal expertise
+   - Use third-party frameworks and APIs
+   - Don't build from scratch
+
+**Examples:**
+1. **DoorDash** (originally Palo Alto Delivery)
+   - Static HTML with PDF menus
+   - Google Forms for orders
+   - "Find My Friends" to track deliveries
+   - Built in one afternoon
+   - Focused only on Palo Alto initially
+
+2. **WayUp** (YC 2015)
+   - CTO JJ chose Django/Python over Ruby/Rails
+   - Prioritized iteration speed over popular choice
+   - Simple stack: Postgres, Python, Heroku
+
+3. **Justin TV/Twitch**
+   - Four founders (three technical)
+   - Each tackled different parts: video streaming, database, web
+   - Hired "misfits" overlooked by Google
+
+**Tech Stack Philosophy:**
+- "If you build a company and it works, tech choices don't matter as much"
+- Facebook: PHP → HipHop transpiler
+- JavaScript: V8 engine optimization
+- Choose what you're dangerous enough with
+
+#### Stage 3: Launch Stage (19:43-26:51)
+**Goal:** Iterate towards product-market fit
+
+**Key Principles:**
+
+1. **Quickly Iterate with Hard and Soft Data**
+   - Set up simple analytics dashboard (Google Analytics, Amplitude, Mixpanel)
+   - Keep talking to users
+   - Marry data with user insights
+
+2. **Continuously Launch**
+   - Example: Segment launched 5 times in one month
+   - Each launch added features based on user feedback
+   - Weekly launches to maintain momentum
+
+3. **Balance Building vs Fixing**
+   - Tech debt is totally fine early on
+   - "Feel the heat of your tech burning"
+   - Fix only what prevents product-market fit
+
+**Examples:**
+1. **WePay** (YC company)
+   - Started as B2C payments (Venmo-like)
+   - Analytics showed features unused
+   - User interviews revealed GoFundMe needed API
+   - Pivoted to API product
+
+2. **Pokémon Go Launch**
+   - Massive scaling issues on day 1
+   - Load balancer problems caused DDoS-like situation
+   - Didn't kill the company (made $1B+ revenue)
+   - "Breaking because of too much demand is a good thing"
+
+3. **Segment**
+   - December 2012: First launch on Hacker News
+   - Weekly launches adding features
+   - Started with Google Analytics, Mixpanel, Intercom support
+   - Added Node, PHP, WordPress support based on feedback
+
+### Role Evolution Post Product-Market Fit
+- **2-5 engineers:** 70% coding time
+- **5-10 engineers:** <50% coding time
+- **Beyond 10 engineers:** Little to no coding time
+- Decision point: Architect role vs People/VP role
+
+## Key Concepts Deep Dive
+
+### 90/10 Solution (Paul Buchheit)
+- Find ways to get 90% of the value with 10% of the effort
+- Available 90% solution now is better than 100% solution later
+- Restrict product dimensions: geography, user type, data type, functionality
+
+### Technical Debt in Startups
+- **Early stage:** Embrace technical debt
+- **Post product-market fit:** Address scaling issues
+- **Philosophy:** "Tech debt is totally fine - feel the heat of your tech burning"
+- Only fix what prevents reaching product-market fit
+
+### MVP Principles
+1. **Speed over perfection:** Launch in weeks, not months
+2. **Manual processes:** Founders do unscalable work
+3. **Limited scope:** Constrain to prove core value
+4. **Iterative validation:** Launch, learn, iterate
+
+## Companies Mentioned (with Context)
+
+### Optimizely (YC W10)
+- A/B testing platform
+- Prototype: JavaScript file on S3, manual execution
+- Founders: Pete Koomen and Dan Siroker
+- Dan previously headed analytics for Obama campaign
+
+### Remora (YC W21)
+- Carbon capture device for semi-trucks
+- Prototype: 3D renderings to demonstrate concept
+- Captures 80%+ of truck emissions
+- Can make trucks carbon-negative with biofuels
+
+### Justin TV/Twitch
+- Live streaming platform → gaming focus
+- Founders: Justin Kan, Emmett Shear, Michael Seibel, Kyle Vogt
+- MVP built by 4 founders (3 technical)
+- Hired overlooked engineers from Google
+
+### Stripe
+- Payment processing API
+- Early days: Founders manually processed payments
+- Filled bank forms manually for each transaction
+- Classic "do things that don't scale" example
+
+### DoorDash
+- Originally "Palo Alto Delivery"
+- Static HTML with PDF menus
+- Google Forms for orders
+- "Find My Friends" for delivery tracking
+- Focused on suburbs vs metro areas (competitive advantage)
+
+### WayUp (YC 2015)
+- Job board for college students
+- CTO JJ chose Django/Python over Ruby/Rails
+- Prioritized iteration speed over popular choice
+- Simple, effective tech stack
+
+### WePay (YC company)
+- Started as B2C payments (Venmo competitor)
+- Pivoted to API after user discovery
+- GoFundMe became key customer
+- Example of data + user interviews driving pivot
+
+### Segment
+- Analytics infrastructure
+- Multiple launches in short timeframe
+- Started with limited integrations
+- Added features based on user requests
+- Acquired by Twilio for $3.2B
+
+### Algolia
+- Search API mentioned as YC success
+- Part of Diana's network of advised companies
+
+## Actionable Advice for Technical Founders
+
+### Immediate Actions (Week 1)
+1. **Build clickable prototype** (Figma, InVision) in 1-3 days
+2. **Find 10 potential users** to show prototype
+3. **Use existing tools** rather than building from scratch
+4. **Embrace ugly code** - it's temporary
+
+### Tech Stack Selection
+1. **Choose familiarity over trendiness**
+2. **Use third-party services** for non-core functions
+3. **Keep infrastructure simple** (Heroku, Firebase, AWS)
+4. **Only build what's unique** to your value proposition
+
+### Hiring Strategy
+1. **Don't hire too early** (slows you down)
+2. **Founders must build** to gain product insights
+3. **Look for "misfits"** - overlooked talent
+4. **Post product-market fit:** Scale team strategically
+
+### Launch Strategy
+1. **Launch multiple times** (weekly iterations)
+2. **Combine analytics with user interviews**
+3. **Balance feature development with bug fixes**
+4. **Accept technical debt** until product-market fit
+
+### Mindset Shifts
+1. **From perfectionist to pragmatist**
+2. **From specialist to generalist** (do whatever it takes)
+3. **From employee to owner** (no task beneath you)
+4. **From certainty to comfort with ambiguity**
+
+## Diana's Personal Insights
+
+### From Her Experience
+- "Technical founder is committed to the success of your company"
+- "Do whatever it takes to get it to work"
+- "Your product will evolve - if someone else builds it, you miss key learnings"
+- "The only tech choices that matter are tied to customer promises"
+
+### Common Traps to Avoid
+1. **"What would Google do?"** - Building like a big company too early
+2. **Hiring to move faster** - Actually slows you down initially
+3. **Over-fixing vs building** - Focus on product-market fit first
+4. **Building features without user insights** - Keep talking to users
+
+## Resources & References
+
+### YC Resources
+- Y Combinator Library: "Tips for technical startup founders"
+- Paul Graham Essay: "Do Things That Don't Scale"
+- Paul Buchheit Concept: "90/10 Solution"
+- Startup School: Technical founder track
+
+### Tools Mentioned
+- **Prototyping:** Figma, InVision
+- **Analytics:** Google Analytics, Amplitude, Mixpanel
+- **Infrastructure:** Heroku, Firebase, AWS, GCP
+- **Authentication:** Auth0
+- **Payments:** Stripe
+- **Landing Pages:** Webflow
+
+### Further Reading
+1. Paul Graham essays (paulgraham.com)
+2. Y Combinator Startup School materials
+3. Case studies: Stripe, DoorDash, Segment early days
+4. Technical debt management in startups
+
+## Key Takeaways
+
+### For Technical Founders
+1. **Speed is your superpower** - Move faster than established companies
+2. **Embrace imperfection** - Good enough beats perfect when speed matters
+3. **Stay close to users** - Insights come from conversations, not just data
+4. **Tech debt is a feature, not a bug** - Early stage startups should accumulate it
+
+### For Startup Strategy
+1. **Constrained focus** leads to better unit economics (DoorDash example)
+2. **Manual processes** create customer intimacy and learning
+3. **Continuous launching** builds momentum and feedback loops
+4. **Break things at scale** is a good problem to have
+
+### For Team Building
+1. **Founders build first** - Critical for product insights
+2. **Hire for adaptability** over pedigree
+3. **Evolve role with growth** - Coding time decreases with team size
+4. **Culture emerges** from early team composition
+
+---
+
+*Research compiled from YouTube transcript, web searches, and Y Combinator resources. Last updated: January 25, 2026*