BEFORE YOU PRESS PLAY
What is discussed
AskMonty AI editorial noteSource consulted:
Montemagno encourages people to bring more than one perspective to important decisions. He starts with chatbots as a simple second opinion, then moves to comparing models, analysing benefits and risks, supplying context and explicitly criticising a plan. Asking only how to cut a cost, he notes, differs from considering its effect on quality. Agents can incorporate these practices into workflows through adversarial reviews, discussions among specialised roles and different reasoning methods. The aim is not an infallible machine, but a way to reveal assumptions and problems that one person might miss. Final responsibility remains human. Personal anecdotes and deliberately absurd scenarios illustrate context rather than providing individual medical or financial recommendations. Examples include Spiegamelo Facile, his own startup, and an invitation to subscribe to its free newsletter.
Key points
- AI can broaden reasoning without taking the final decision away from the person.
- Start with a second opinion, then compare models and analyse risks and alternatives.
- Complete context makes the discussion more relevant.
- Requesting explicit objections helps counter agreeable answers.
- With agents, decision procedures and different roles can become part of the workflow.
