AI summary by TwinMind
Will AI Make Us the Last Generation to Read and Write? | Victor Riparbelli | TED
TED
16m 24s
Opening Thesis: End of Reading and Writing
- Provocative claim: grandchildren will be last generation to read/write
- Humanity pursuing better ways to convey ideas beyond text
- Dawn of AI-enabled communication era
- Future generations will replace text with audio, video, immersive tech
- Reading/writing will become historical artifacts like papyrus, hieroglyphs, cave paintings
Personal Background and Journey
- Speaker's love for reading, Copenhagen library visits, science fiction books
- Internet discovery at age 10 opened new world of ideas, music, people
- Observed how technology changed content distribution and content itself
- Music passion: software instruments, sampling, drum machines created new genres
- Started first business at 13 in World of Warcraft (50-person dragon-slaying team)
- Compared online gaming leadership to startup management
Founding Synthesia
- 2016: Discovered "Face to Face" research paper by Professor Matthias Niessner
- Neural networks producing photorealistic video felt like magic
- Predicted Hollywood film creation from bedroom within 10 years (took 3 years)
- Founded Synthesia with Matthias, Lourdes, Steffen as cofounders
- Initial vision: make everyone a Hollywood director
- Evolved to enabling all content (text messages to corporate training) in video/audio
History and Limitations of Text
- Text as original compression method for human communication
- Pre-1500 BC: only oral information sharing
- 1500 BC: first alphabet invented, simplified complex writing systems
- 1440: Gutenberg printing press enabled mass production of written content
- Mid-20th century: reading became accessible to masses, literacy spread
- Today: text ubiquitous, impossible to live without reading/writing
Text's Imperfections
- Efficient and scalable but very lossy compression method
- Lacks nuance: tone, body language, context, location
- Ambiguous interpretation ("good news" example)
- Emojis attempted improvement but still imperfect (smiley face debate)
- Visual vs text comparison: image understood in seconds vs 30 seconds reading description
- Even 30 pages of description wouldn't match actual image perception
- Time dimension (video) compounds the problem exponentially
Evolution Toward Richer Communication
- Innovation trajectory: radio, TV, internet, VR, social media, AI
- 2024 evidence: people prefer watching/listening
- TikTok: fastest growing social network AND search engine
- Voice notes on WhatsApp, dating apps
- Product videos for shopping
- Personal admission: starts learning on YouTube/TikTok/podcasts before books
- Only reads 200-page books when deeply invested
Cultural Attitudes and Guilt
- Guilt around consuming video/audio instead of books
- Commentary: "Young people can't focus, need dopamine hits"
- Same parental concerns for 200 years
- Provocative counter-theory: tired of overly dense, slow information
- Current generation may learn faster because of technology, not despite it
- Question: is problem us or text?
- Still read extensively from multiple sources daily
- Psychological attribution of higher value to written word
Economic Barriers to Video Content
- Cost in time and money prevents video adoption
- Current choice: speed/scale of text vs accuracy/engagement of video
- Economic incentive: only "important enough" content gets video treatment
- Entertainment: "Fast and Furious 278" vs avant-garde student films
- Corporate: Super Bowl ads are video, fire safety documents are text
AI Revolution and Synthesia's Role
- AI changes equation: enables speed, scale, accuracy, engagement simultaneously
- Photorealistic content creation digitally
- Computers learning world appearance, replicating/remixing with detail
- New creativity wave driven by individuals, not Hollywood
- Synthesia focus: AI avatars (digital humans)
- Current reach: millions of daily interactions
- Applications: teaching, onboarding, health guidance, product sales
- 130+ languages supported
- Avatars becoming indistinguishable from reality
AI Avatar Demo
- "Alex" personal avatar demonstration
- Entirely AI-generated video
- Natural expression and sound quality
- Technical note: still learning when to stop speaking
Democratization of Content Creation
- Keyboards/computers made everyone writers
- AI will make everyone directors with Hollywood-grade video
- No training required
- Synthesia: 1M+ people turned into video creators
- People create videos independently without management approval
Research Findings
- UCL London collaboration study (400 participants, ongoing)
- Results planned for publication next year
- Initial finding: 77% prefer learning through video vs text
Future Media Formats
- Current AI content as "bridge genre" - old formats with new tech
- Today's AI videos: linear, beginning/end, emulating physical cameras
- Future combination: AI video + reasoning systems (language models)
- Interactive and personalized media
- Thinking, narrating, personalizing content
Educational Applications
- Music theory assistant knowing skill level and taste
- Celebrity teachers for math in personalized contexts (soccer, sci-fi)
- Education turbocharged with AI systems
Entertainment Evolution
- Viewer-shaped content responding to world around them
- Interactive films, never-ending TV series
- Social media feeds as endless personalized stories
- TikTok as example of endless personalized content stream
Future Communication Modes
- AR, VR, brain-computer interfaces integration
- Media appearing like real life
- Thought-based message transmission
- Inbox of friends'/relatives' AI avatars instead of text
- Fire safety videos as "Spielberg-esque masterpieces"
Ethical and Philosophical Questions
- Do we care if content is AI-generated?
- Comparison to special effects, green screen usage
- Actors paid to endorse unknown products
- Real vs generated actors
- Harry Potter vs Daniel Radcliffe fandom
- Mickey Mouse and Pixar characters emotional connection
- Trust in AI agents vs human call center representatives
- Entertainment from computer-computer interactions
- Computer sports/chess competitions
