Tutto quello che penso degli Agenti Ai
Summary
Marco Montemagno analyzes the shift from chatting with AI to delegating concrete tasks to AI agents, emphasizing that fear is an ineffective strategy compared to practical risk management. He argues that AI does not replace jobs but dismantles them piece by piece, requiring professionals to leverage their domain expertise to supervise outputs and shift from time-based to result-based pricing. Through a live demo using Hostinger, he illustrates how agents can autonomously build websites, while warning that the technology remains imperfect and requires human oversight to correct errors.
Topics discussed
Intro: The chaotic world of AI agents and robotics
Distinction between chatting and delegating tasks to AI
Why fear is not a strategy: analyzing risks pragmatically
The first time we can delegate work to non-humans
The importance of clear delegation and clarity of thought
Designing websites for AI agents, not just humans
Practical demo: Creating a site with Hostinger AI Builder
Specificity matters: Models, infrastructure, and pilots
AI as brilliant cooks with a habit of starting fires
The need for domain expertise to supervise AI output
How AI dismantles jobs piece by piece rather than replacing them
Shifting from time-based to result-based compensation
The irreplaceable value of human emotion and empathy
Solving AI problems requires more AI, not less
Local AI, cost paradoxes, and the Jevons paradox
Reviewing the generated website and publishing it
Conclusion: A pragmatic approach to working with AI agents
Listen ad-free on Castria