# My a AI: Ako ešte zmení svet, a ako jej prispôsobiť ten náš | Marek Urban | TEDxTurcianskeTeplice

Source: https://www.youtube.com/watch?v=jUq3jc68FtY
Recap page: https://rapidrecap.app/video/jUq3jc68FtY
Generated: 2026-01-26T18:07:12.51+00:00

---
## Quick Overview

Marek Urban argues that generative AI is rapidly advancing past human capabilities in cognitive tasks, necessitating a societal shift where human value is redefined away from professional identity towards intrinsically rewarding activities like play, as the projected arrival of Artificial General Intelligence (AGI) and subsequent Superintelligence will render most current human work obsolete.

**Key Points:**
- The adoption rate of ChatGPT is historic; it grew from 500 million to 600 million users in one week, with unofficial numbers reaching 1 billion, contrasting with Twitter's 600 million users.
- Generative AI significantly reduces mental effort required for problem-solving while simultaneously increasing performance across cognitive levels, including memory, understanding, application, analysis, evaluation, and creativity.
- Current second-generation reasoning models, like GPT-4o, are already achieving performance comparable to the top 2% of the human population on IQ tests and outperform human doctors in diagnostic accuracy in real-world hospital settings.
- The speaker notes that GPT is currently three levels ahead of the Slovak education system, stating, "it can solve problems that take a human 60 minutes" and this capability doubles every seven months.
- The speaker predicts the arrival of Artificial General Intelligence (AGI) by 2027 (expert estimates) or 2031 (market predictions), followed three years later by Superintelligence, which will outperform humans in all tasks.
- To cope with the obsolescence of work identity post-Superintelligence, philosopher Nick Bostrom suggests the solution lies in embracing 'play'—activities chosen for joy and curiosity, not necessity or obligation—as the source of future human value.
- AI models are increasingly capable of tool use, including Googling information, coding in Python, analyzing data, and managing files, mirroring intelligent behavior seen in animals using tools.

**Context:** Marek Urban, presenting at TEDxTurcianskeTeplice, discusses the current state and future trajectory of generative Artificial Intelligence, emphasizing that researchers aim to proactively influence the world shaped by AI rather than merely reacting to models released by entities like OpenAI. He grounds his predictions in current adoption statistics, cognitive research showing AI's impact on mental effort and performance benchmarks using revised Bloom's taxonomy.

## Detailed Analysis

Urban begins by highlighting the explosive growth of generative AI, citing that ChatGPT reached close to a billion users in a single week, and notes that 50.7% of surveyed university students in the Czech Republic use it several times a week because 'thinking hurts' and AI reduces mental effort while boosting performance. He maps this progress onto the revised Bloom's taxonomy of cognitive goals, stating that first-generation models handled memory and understanding, second-generation models (like GPT-4o) handle application and analysis, and current reasoning models are strong at evaluation, though creativity remains challenging due to the need for original yet useful solutions relevant to a specific social context. Crucially, Urban reveals that the newest models perform at the level of the top 2% of humans on IQ tests and demonstrate superior diagnostic accuracy compared to human doctors in practical hospital scenarios. He forecasts AGI arrival by 2027–2031, followed shortly by Superintelligence, which will surpass human capability in every domain. This technological shift makes current professional identities unsustainable, leading to a non-normative retirement crisis. Urban concludes by advocating for a solution proposed by Nick Bostrom: shifting human focus to 'play'—activities pursued for intrinsic joy and curiosity—as the only viable source of value when necessity-driven work ceases to exist.

### Generative AI Adoption and Impact

- ChatGPT reached 1 billion users unofficially within a week
- 50.7% of surveyed university students use GPT multiple times weekly
- AI reduces mental effort while increasing performance across cognitive tasks

### Cognitive Capabilities Benchmarking

- Current models match top 2% human IQ performance
- AI significantly outperforms human doctors in diagnostic accuracy in real-world settings
- Models demonstrate strong capability in evaluation and analysis, approaching creativity

### Timeline for Advanced AI

- AGI predicted by 2027-2031, followed three years later by Superintelligence
- Problem-solving duration solvable by AI doubles every seven months

### Future Societal Implications

- Superintelligence removes the basis for human competition in most tasks
- This creates a non-normative 'retirement' crisis as professional identity becomes obsolete

### Proposed Human Response

- Nick Bostrom suggests 'play'—activities done for joy and curiosity, not necessity—as the future source of human value
- Society must rediscover who they are outside of their professions