Very interesting discussion. Years ago, I did renovation work and often used the same gambit. Often it was a model of "OK, here is the labour price, here are the material prices, but what you're asking for is something we've never done before and it's impossible to estimate a fair price." Surely my poor estimating skills were on display - but that's not what I was being paid for. Could we do it? Yes. Could we price it out to the penny? No. But we still got the job.

Once in a while it was simply, "Yes, we can do that. No, I don't know how much it's worth. Let's do it and figure it out later." And, with people who had confidence in my work, there was never a problem with this approach. I was often surprised at what people would pay. And as the article mentions, I would never use this approach with someone new or a large enterprise. But with familiars, this always worked out OK for me.

Fascinating to see this model applied to untried, almost purely intellectual work (code).

As with renovations, where you're often busting you're ass against someone with a pickup truck looking for beer money for the weekend. Working with code pits you against every chump with a computer. So how can you differentiate yourself?

Being able to say, "Hey, no problem. I can do that, We'll square up on Monday." is one way. Weekend warriors won't do that.