An untidy history of AI across four books
10 hours ago
- #Technology Hype
- #Machine Learning
- #Artificial Intelligence
- AI's history traces back to ancient technologies like the abacus, evolving into modern AI research post-WWII with the symbolic paradigm.
- Machine learning emerged as a competing approach, overcoming early limitations with increased data and computing power, notably through GPUs.
- Neural networks gained prominence after a 2011 ImageNet competition success, leading to widespread applications in social media, search engines, and e-commerce.
- OpenAI, founded in 2015, released ChatGPT in 2022, rapidly making generative AI ubiquitous and sparking significant commercial and cultural impact.
- AI hype often distorts its capabilities, conflating generative AI (content creation) with predictive AI (forecasting), despite their distinct limitations and risks.
- Critics like Narayanan and Kapoor warn against AI's overhyped potential, emphasizing its current unreliability and the dangers of misapplication.
- Public confusion about AI is exacerbated by misleading marketing, from Hollywood narratives to commercial products falsely labeled as AI-powered.
- High-profile figures like Harari, Kurzweil, and Kissinger contribute to AI mystique with speculative, often inaccurate predictions about its future impact.
- Existential risks of AI, such as those compared to nuclear weapons, are criticized as distractions from more immediate, tangible harms like bias and misuse.
- The AI industry's focus on futuristic, utopian visions often overlooks practical, incremental improvements that could address real-world inefficiencies.