Most “AI” projects fail because they try to be big. I prefer small tools that:

• reduce a recurring cost (time, money, attention)
• are easy to maintain
• are measurable

That’s the same lens I use for FIRE and investing: build systems that compound.