|
>The sentiment, that Americans don't NEED fast home internet is probably more accurate than not. No, it's total bullshit, except in the pedantic "well, humans don't NEED electricity/medicine/communications/synthetic shelter/[anything since we wandered the plains of Africa and had a life expectancy in the 40s]" sense. Fast, symmetrical end point Internet is of massive importance and in not just direct but emergent ways that a surprising number of people, even on HN, don't seem to consider. Think for example of many of the concerns that have been raised over the last few years here about how the net has become ever more centralized, or about the difficulties of efforts like Tor. A lot of the core driver for that comes down to lack of end point bandwidth vs demands. If symmetrical gigabit connections were the rule rather then the exception, it'd be possible once again for significantly sized services run directly. Obviously if a service grew large enough they'd eventually need to move towards the core, but for a lot of people it'd completely eliminate the need for many current colo and cloud offerings. Would you be happy going back to thinnet or 10BASE-T on your LAN? Everything that runs there could run over the net with more bandwidth, latency only becomes a significant issue over extremely long distances. More bandwidth also means more can start to be devoted to meta-content issues like privacy. If we consider 25 Mbps to be a target for many services say, then onion routing networks are hard to make use of because they tend to impose significant overhead and be limited by slow nodes (particularly as even someone with a "enough" 25 Mbps connection themselves often would not be willing to allow all of that to be used by the larger network). But if everyone had 1000 Mbps (or more) to play with, then they could easily devote a bunch of that to sharing networks, take even a 90% overhead hit, and still have as much bandwidth as they needed. Just as more power in CPUs/GPUs has allowed us to not merely run things faster, but optimize towards the value of human time vs computer time, more bandwidth "then needed" represents leeway to optimize towards goals beyond the bare necessities. Seriously, giving everyone (or close enough) fast, symmetrical connections should be one of the absolute highest priorities of anyone concerned about market competition, centralized control, privacy, and so on. It'd be one of the best investments America could make, just as national electrification, telephone, and roads were. |
That assertion seems reasonable to me. What does "fast" even mean? Don't know? Then it's a good assertion. You don't need what isn't defined.
This is not the same as "humans don't NEED access to nuclear materials", because the acceptable safety measure is currently 0. So that's also a feasible assertion. The question about tradeoffs and what is practical has to be answered first. Until you say what the minimum is, there's no point arguing about how supposing an upper limit (to what is needed) is wrong.