For example, it will carry substantially larger payloads than the SLS, at less than 100th the cost per launch ($15M vs $3B), and launch almost daily, vs twice a year. The Shuttle flew every couple months, at a payload cost of over $30,000/lb, vs under $100/lb for the BFR.
Now I don’t think the first versions of the BFR will achieve those goals, but the crazy part is they are achievable. Reusability drives costs rapidly down towards to the cost of fuel, which is under $1M per launch.
The problem Elon faces is refurbishment costs. It seems premature to think that the BFR and Starship can fly ten flights with only minor refurbishment, and fly one hundred times before replacement. But the benefits of reuse are so huge that even if they fly only ten times each launch costs can be less than a Falcon 9.
It's all conjecture obviously. The point is that we probably don't need to carry all our heavy tools and infrastructure up there- we just need to find new (hopefully better) ways of creating the stuff we need.
I'm convinced you cannot bootstrap a colony in an unhospitable place in the Solar System without ships 3-4 orders of magnitude above what we have today.