Very Briefly on AI Agents
One of the most significant breakthroughs in history has probably been generative AI. For two years, its impact, usage patterns, and life-changing features have rapidly started to affect every field from education to law. Economist Daron Acemoğlu thinks its contribution to the global economy in the next decade will only be around 0.93-1.56%, while Goldman Sachs claims it will go up to 7%.
One of the most significant breakthroughs in history has probably been generative AI. For two years, its impact, usage patterns, and life-changing features have rapidly started to affect every field from education to law. Economist Daron Acemoğlu thinks its contribution to the global economy in the next decade will only be around 0.93-1.56%, while Goldman Sachs claims it will go up to 7%. [i] What I'm curious about is what indirect factors they consider when making these assumptions. Meanwhile, one of the most researched topics among technologists recently is artificial intelligence agents (AI agents). If I were to explain AI agents in my own language, they are tiny, super-intelligent, decision-making assistants placed inside our electronic devices (phones, computers, tablets, consoles…) that are extensions of our arms. AI agents understand you, react, learn… and can even make decisions [ii] . Their biggest features are that they operate completely independently of the user and "make decisions." They can organize your travels, make hotel reservations, schedule your appointments, analyze your credit options, or write your speeches for you. In short, they organize your data and can independently keep up with the pace of a digital world. We are still very early in the journey, but still, AIAs have existed for a long time. But previously, they mostly provided power and competitive advantage to AI researchers and especially large companies. The development of generative AI has greatly contributed to the faster creation of AIAs. With the influence of fierce competition, they are increasingly being offered directly to the user at a faster pace. 2024 also seems to be the year for startups developing AIAs. As of the first half of the year, Y Combinator supports 12, and Abstract Ventures supports 5 investments [iii] . I'm not even counting the big AI companies. As we said, AIAs also "learn"... they learn, and they are developed. In at most 10 years, everyone will have super-intelligent "assistants," for better or worse. If we also consider the interaction between agents, judging how possible it is to calculate their impact today is, in my opinion, very debatable. I am sure that the impact of AIAs will be very positive, especially in areas such as health and education. In fact, there is a very high probability that the market will expand significantly, especially in cybersecurity. But I am very curious about how governance and legal systems, which are struggling to adapt to today's conditions, will be able to cope with this issue. If they cannot cope, what will fill the gap? [i] https://www.project-syndicate.org/commentary/ai-productivity-boom-forecasts-countered-by-theory-and-data-by-daron-acemoglu-2024-05 [ii] https://www.technologyreview.com/2024/07/05/1094711/what-are-ai-agents/ [iii] https://www.cbinsights.com/reports/CB-Insights_Future-Workforce-AI-Agents.pdf?utm_medium=email&_hsenc=p2ANqtz--fY_p7nlsQPuKEoGCNQ7GwkFVY-KBRQq8vXiiumghY4u0pvOr7nsWWXbJxnowpaBAB0nMEh0nEI3c7bsG1UKElmeouqANmKDtd8DceGgag0PXpK1Q&_hsmi=228539474&utm_content=228539474&utm_source=hs_automation