The Illusion of Thought | Raphael Jude Ariho Bitarabeho | TEDxAga Khan HS Kampala Youth
Raphael Bitarabeho explores whether AI chatbots truly think or merely create an illusion of thought, using the Mad Libs game to demonstrate how neural networks work. He argues that AI's power comes not from processing data, but from embracing failure at scale, and introduces Court Vision, his AI application for physical activities, to illustrate how iteration and persistence drive transformative innovation.
Summary
Raphael Bitarabeho begins by addressing the paradox of 'rewriting tomorrow'—how we can change a future that hasn't been written yet. He frames this discussion through the lens of artificial intelligence, particularly AI chatbots like ChatGPT, Claude, and Gemini, questioning whether they actually think or produce an illusion of thought. He challenges the historical definition of intelligence as resembling human qualities like speech, logic, and reasoning, asking whether intelligence might exist in forms we don't recognize. To ground this abstract concept, he engages the audience in a Mad Libs game, demonstrating how early neural networks worked by randomly filling in knowledge without real understanding. He then explains the breakthrough 'Attention is all you need' paper, which revolutionized neural networks by teaching them to assign values to words based on relevance rather than generating random gibberish. Bitarabeho emphasizes that this illusion of thought is now a global concern, with AI systems creating hyperrealistic videos, writing code, replacing entry-level jobs, and prompting government regulation from leaders who barely understand the technology. He shares his personal project, Court Vision, an AI application integrating artificial intelligence into physical activities. Drawing parallels to Google DeepMind's journey from simple games to AlphaGo's championship victory, he explains Court Vision's training process: data collection, annotation, and training. His central argument is that AI doesn't surpass human intelligence through data processing or numerical computation, but through its ability to embrace failure at lightning speed. He concludes by extending this insight to human potential, arguing that individuals can also rewrite their futures by embracing failure, iterating boldly, and overcoming challenges.
Key Insights
- Early neural networks functioned like the Mad Libs game, randomly filling in blanks of knowledge with no real focus or understanding, until the 'Attention is all you need' paper taught models to assign values to words based on their relevance within sentences.
- AI chatbots do not actually think; what we witness is an illusion of thought created through advanced pattern recognition and language prediction mechanisms.
- AI does not surpass human intelligence because of its ability to process large amounts of data or crunch numbers, but because of its ability to embrace failure at a lightning fast speed.
- Google DeepMind's transformation from solving simple, highly focused problems like playing games to creating AlphaGo, which defeated the world Go champion, demonstrates that the difference between something simple and something world-changing is the ability to iterate, believe in vision, and be relentless.
- AI systems are now creating hyperrealistic videos, writing code better than junior developers, and replacing entry-level positions, while governments rush to regulate technology they barely understand.
Topics
Transcript
[0:06] Rewriting tomorrow. Taken on face value, this phrase seems quite tough to grapple with. It reminds me of a popular line from the film Back to the Future part 3, in which Doc Brown says, "Your future isn't written. No one's is. Your future is whatever you make it, so make it a good one." How can you rewrite something that is yet to be written? Good morning, ladies and gentlemen. My name is Rafael Bittar Abejo, and today I'd like to address this paradox by [0:37] referring to AI chatbots and demystifying the whole concept of artificial intelligence, and how they are literally rewriting our tomorrow from global innovation to education. As a young boy, I remember asking…
Full transcript available for MurmurCast members
Sign Up to AccessMore from TEDx Talks
The violence we don’t see | Dr. Kaitlin Devaney | TEDxU Chicago Illinois
Dr. Kaitlin Devaney expands the definition of violence beyond direct physical harm to include structural violence (built into systems that disadvantage certain groups) and cultural violence (narratives that normalize inequality). She argues these three forms of violence operate together, and addressing harm requires examining systems and questioning deficit-based narratives rather than only punishing visible criminal acts.
How to easily make big decisions | Neel Patel | TEDxU Chicago Illinois
Neel Patel presents a framework for making major life decisions by shifting focus from "what do I want to do" to "who do I want to be." He advocates for picking a decision and committing to make it the right one rather than endlessly seeking the perfect choice, illustrated through his personal journey from accounting student to college professor.
The 5 Lies You've Been Told About Breathing | Dr. Andy Sabatier, DPT | TEDxBend
Dr. Andy Sabatier debunks five common lies about breathing, emphasizing that functional breathing prioritizes diaphragmatic movement over chest breathing, proper CO2 retention over oxygen maximization, and intentional breathing practice that retrains automatic breathing patterns. By breathing functionally (nose inhalation, diaphragm first, slow exhalation), individuals can improve physical, mental, and emotional aspects of their lives.
The power of prioritizing your own health | Dr. Stephen Dahmer | TEDxBend
Dr. Stephen Dahmer argues that modern medicine confuses 'sick care' with 'health care' and calls for a health rebellion centered on daily practices like breathing, movement, connection, and purpose rather than medical interventions. He emphasizes that while it's not individuals' fault they inherited a disease-focused system, they bear responsibility for tending their own health through accessible, free practices.
What Science Tells Us About Childbirth That Your Doctor Won't | Assoc Prof Yvette Miller | TEDxQUT
Associate Professor Yvette Miller argues that modern obstetric practices treat childbirth as a medical emergency requiring intervention, when 90% of maternity guidelines lack strong scientific evidence and over 70% of pregnant women receive unnecessary medical interventions. She advocates for informed decision-making and questions the assumption that women's bodies are inherently unreliable.