Comment Re:Bad way to conduct policy (Score 1) 131
The legislators are very much involved. If the FCC makes a decision that the corporations don't like, their pet legislators will reverse it.
The legislators are very much involved. If the FCC makes a decision that the corporations don't like, their pet legislators will reverse it.
True, but 10TB is not the average user. Probably most of their customers have single-drive laptops or desktops.
Newegg doesn't seem to have any laptops with more than 1TB disk, so let's call that a typical disk size. Most users won't have a second disk. Assume a typical home user's desktop/laptop has about a terabyte (my laptop has half that), and assume they've managed to get it 75% full of the kind of stuff Backblaze's app will back up. That's 750GB per user, which at their posted hardware price (for the drive and the box it fits in, but not the rack and datacenter) of $0.051/GB (with their storage pod 4 design) that's about 40 bucks of hardware, or 8 months to break even.
I myself have something like 5TB of space on my fileserver, and it's probably also about 75%, so that'd be 5 times as long to break even on my account. But there's a lot more folks with single drives than with file servers; assuming 10:1 (which I think is conservative) that's 15 typical user's worth of data for 11 accounts, pushing the overall breakeven time to about 11 months.
There's certainly folks with more space than me, but they're comparatively rare; I think it would be reasonable to assume that about 1 in 4 of our posited 5TB users are really 10TB users. That means out of 44 accounts, we have 1 at 10TB, 3 at 5TB, and 40 at 1TB; average per person: 1.4ish TB; keeping our 75% full assumption that brings breakeven up to just under 12 months.
Certainly they have to be counting on long term subscriptions and bulk hardware purchases, and certainly you don't want them to be your only backup for something irreplaceable in case something goes wrong and they go under and you can't get your stuff back from them. But they reportedly are storing 100PB, which (based on our figured average of 1.4TB/user) is some 70k accounts. Now, that estimate is wrong, but it's probably not off by a factor of more than two, so figure it's really only 35k accounts, and they're making 175k/mo. At that scale, if they were losing money, I think they'd have adjusted their price by now...
I remember the Mythbusters looked at the Hindenberg. It's been a while so I may be misremembering, but if I recall, their replication of the skin did burn with an inert gas, but it burned more spectacularly with hydrogen (I suspect that they were able to get a different and more energetic reaction with hydrogen present).
However GGP does have a point; even if the bag had been full of helium, once it caught, the ship was going down.
Curse you. You have just destroyed my afternoon free time
and optimism, don't forget that. Pessimists would just give up instead of trying to find funding.
This requires giving the pilot an opportunity to spend that time, and according to GP a lot of places outside the US make that very difficult.
I agree with your idea, but implementing it seems to hit some roadblocks.
only one case where the seizure was found to be unjustified does not actually mean all the rest were really justified. (It also doesn't mean they weren't, of course. Insufficient data. But it feels unlikely that there were no other incorrect seizures; 65000 instances with no false positives is a better accuracy rate than almost any human activity.)
If I recall, there's on-the-fly dedupe and batch dedupe. Is the 5GB/TB estimate for both, or just on the fly? (I like dedupe as a concept, and I've gotten worthwhile results from it, but I can live with batch to avoid getting enough ram to require a new motherboard
that's pretty cool. In that case, it seems like it would be much harder to fake being a legit device nearby. Thanks for the info
It depends. If the cafeteria has multiple good dishes, I gotta get 'em all! Fortunately this is somewhat rare
that's a good question. How feasible is it to build a widget that can tell the difference between transit time and (remote) processing time?
He may be taking into account the time it takes to mentally get back "into the zone". I personally get interrupted enough during a day that a couple of extra times don't really matter (I already go to get water, hit the head, or get lunch, several times a day; it would only take a couple of additional 5 minute walks to get an average of one per hour).
I want it controlled by a sensor that monitors the oxygenation level of the blood.
I should clarify. The near field, being defined as a multiple of wavelength, is still only that far. But the signal doesn't just stop there; interception of "near field" communications from substantial distances has already been done.
with sufficiently good antennas, "no further" becomes very hard to enforce.
Memory fault - where am I?