Sure, you think that now, because you're in the group that interfaces are designed for. But once you're the guy at the farmer's market, you'll think differently. Power only matters if one can get something done with it. A kilogram of plutonium is enormously powerful, but that's no reason to carry it around in your pocket.
I'll also note that we have been sacrificing the power of our devices in exchange for improved ease of use since the 1980s and it's great. Manually configuring X-Windows in those days was enormously powerful, and enormously user-hostile. We sacrifice vast amount of computing power on the altar of usability, and that's exactly what we should be doing. Contrary to my early computing experience, I have never once had to rebuild the kernel on my phone, and I am very happy with that.
And third, you mostly create a false dichotomy here. If we're talking about a typewriter or a steam loom or something, yes, power and user complexity go together. But the true power of software is the ability to hide most of the complexity most of the time. Most of the work of software development is pushing complexity down, wrapping it in abstractions that provide the next level up with a clean interface.
A perfect example here is Google Search. It's a magic box into which I can type (and now, just say) anything, and it will apparently intuit what I want. The "powerful" version of that interface was 1980s search, where you carefully specified the various fields you wanted to search (because all data had to be carefully fed into the system's structure) and a painfully constructed, manually stemmed, boolean query. For maybe 1 search in 500, I still want that kind of power. But for the rest of the time, I'm very grateful that thousands of nice people at Google have made it so that their search's power has increased continuously, with users needing to learn less and less to get good results.
I'll also note that we have been sacrificing the power of our devices in exchange for improved ease of use since the 1980s and it's great. Manually configuring X-Windows in those days was enormously powerful, and enormously user-hostile. We sacrifice vast amount of computing power on the altar of usability, and that's exactly what we should be doing. Contrary to my early computing experience, I have never once had to rebuild the kernel on my phone, and I am very happy with that.
And third, you mostly create a false dichotomy here. If we're talking about a typewriter or a steam loom or something, yes, power and user complexity go together. But the true power of software is the ability to hide most of the complexity most of the time. Most of the work of software development is pushing complexity down, wrapping it in abstractions that provide the next level up with a clean interface.
A perfect example here is Google Search. It's a magic box into which I can type (and now, just say) anything, and it will apparently intuit what I want. The "powerful" version of that interface was 1980s search, where you carefully specified the various fields you wanted to search (because all data had to be carefully fed into the system's structure) and a painfully constructed, manually stemmed, boolean query. For maybe 1 search in 500, I still want that kind of power. But for the rest of the time, I'm very grateful that thousands of nice people at Google have made it so that their search's power has increased continuously, with users needing to learn less and less to get good results.