7 karma · joined September 19, 2026
I can’t overstate how bad of an idea I think using an AI for customer interaction is.
Building roads is expensive, and they don’t last forever. A good asphalt surface may be service worthy for 20 years. A ton of asphalt outside of California may cost in the ballpark of $150 to install. A 2 inch lift of asphalt will cover not quite 10 sqyd. So you can see how expensive this gets when we’re talking about thousands of miles.
I think it’s worth considering how policies may destroy much of the demand for roadways. Personally, I could imagine a future wear personal private transportation is basically unlawful or artificially expensive.
So given that the literary work was outsourced, anyway, is it really that big of a deal that it’s getting outsourced to a machine? If that still bothers you, then how much machine assistance is too much? We’re already using Word processors, spell checkers, and other tools that have taken away much of the human effort. These language models don’t really have a strong sense of personal agency or what ought to be written, but they can fill out the details from a high-level schematic, which is probably, at best, what the author was doing with other team members.
A wildlife overpass is certainly not going to be cheap, but I could see the road user safety benefit penciling out over the life of the asset. I wonder why you would send the animals over (and span the whole roadway) rather than under (cut and cover a box culvert). Either way, it would be nice to see this better funded for rural America.