Posts by pdh 52 posts • joined 18 Mar 2014
An old quote from Steve Grand: "[W]e used to have intelligent cars; they were called horses. And they used to know stuff that our cars don't know. They used to know where they lived and how to get home and how not to knock people over. Even how to refuel themselves. The amazing thing was that they could even make new copies of themselves. The intelligent car would be like a horse. Something that really enjoyed a good drive and prided itself on not knocking people down."
> I'm not sure how changing CEO is going to brush this under the carpet. It's not as if they can lay the blame on the outgoing CEO for the MCAS palava that has cost hundreds of lives!
Perhaps it's not a question of: was he directly responsible for the original problem; but rather: how did he handle the situation after the problem occurred?
> In banking, for example, you can accept a few glitches but when it comes to human life you cannot have that, of course
I'm surprised that anyone can say that with a straight face, given that hundreds of thousands of patients die from medical errors every year.
http://www.uscourts.gov/about-federal-courts/court-role-and-structure
"There are 13 appellate courts that sit below the U.S. Supreme Court, and they are called the U.S. Courts of Appeals." There's a map on the page that shows the boundaries of the different appellate courts.
> as those who said "no" would see it, they're probably right
Yes. If you're big enough to afford a proper IT staff, and the necessary software licenses, and redundant hardware, then maybe you can do it better in-house. But for smaller outfits the cloud often makes more sense.
Sure there's going to be occasional downtime, no matter how many 9's your cloud provider claims, and of course you still need to keep backups somewhere other than in that same cloud, but unless you can afford multiple competent IT people (more than one, since they tend to take vacations sometimes) and redundant hardware (elsewise you have a single point of failure in your one crucial server, and how long will it take to re-build the server when it eventually fails) you can probably get better uptime in the long run if you use cloud solutions.
It's like mains power: a few sites really do need a large onsite generator and etc so they can ride out multi-hour power outages, but that's very expensive. So most people just have enough UPS so they can run for a few minutes and then safely shut things down if the power doesn't come back quickly. Long power outages are rare enough that for most businesses it doesn't make economic sense to maintain your own fully-capable power-generation equipment. So it is becoming with IT.
> If it works, it will slot seamlessly into existing systems, and the users need never even know it's there.
Exactly. Blockchain is just a key-value database with certain special properties. If banks and etc find it useful, it will take its place in their IT environments alongside other databases that have other specific properties. The concerns about liability and protection of certificates and so on, they apply to the banks' entire IT estates and presumably come under the same regulatory requirements and legal structures. Blockchain is just one particular tool among many that may be useful to certain organizations, if their cost-benefit analysis justifies it.
Figure roughly 1000 miles each way for low-earth orbit, depending on the exact orbit height... speed of light is about 186,000 miles per second, so 2000 miles takes about 11 milliseconds -- that is, it would take that much longer for a one-way data packet from ground station 1 to satellite to ground station 2 as compared to a perfect zero-latency link. If the internal processing in the satellites can keep up without introducing more delay, that's not too bad.
> Who actually needs their transaction records to be distributed?
Maybe think of it as "shared" rather than "distributed." Here's an example of why that's useful...
A few months ago my cellphone provider threatened to cut off my service because I hadn't paid the previous month's bill. I logged into my bank's website and I was able to see that my check had cleared several weeks ago, and the phone company had cashed it, so this was clearly an error on the part of the phone company.
However, because the three of us have three different databases (me with my checkbook, the bank with my bank account info, and the cellphone company with their information about my account) and since one of the three databases disagreed with the others, it took two weeks and several phone calls and a bunch of emails to straighten things out.
A trustworthy shared / distributed database can help prevent this kind of thing. We got everything corrected in the end, but the dispute resolution process could have been streamlined pretty substantially if my bank and the cell provider had a shared database that they both trusted.
> But then blockchain becomes simply a distributed, crypto-verifiable database, right? And that already exists, right?
Mostly right. It's a tamper-proof, permissioned, append-only distributed database. (Tamper-proof by virtue of the crypto.) If you've already got one of those then you probably don't need blockchain.
Lots of questions about what blockchain is good for in a business context...
Blockchain gives you an immutable (write-once) database which is sharable and permissioned. The use case is for transactions amongst business partners who today must each record every transaction in their own private databases. The basic benefit for blockchain in a business context is that participants can all share one copy of the transaction log, so everyone sees the same consistent data.
The features that blockchain adds, beyond what you'd get by sharing a traditional database, are these: You have cryptographic proof that the database isn't tampered with. The permission system means that participants can see only those transactions that they themselves are involved with. Smart contracts allow transactions to be proposed automatically, subject to confirmation by the parties involved in the transaction. And contrary to the way cryptocurrencies like Bitcoin work, the confirmation process does not have to be resource-intensive. (That resource-intensive confirmation process is central to how Bitcoin works, but it's not a fundamental requirement for blockchain in general.)
So the benefit in a business context is improved efficiency. A group of business partners who regularly interact with each other can share a single, permissioned, tamper-proof ledger, and routine transactions can be automated but are still subject to approval by the participants in the transaction.
How many companies actually need blockchain is another matter, but that's basically what it gives you.
> Poor weather is a barrier to cycle adoption, true, but it can be engineered around.
You've never lived in a cold, snowy city, have you?
I live in the northeastern U.S. in a city of 100,000. We get a lot of snow -- there's about a foot of it on the ground right now, thanks to last week's thaw which reduced the snow cover. The current temperature is 13 degrees F -- it was colder yesterday. Most of our roads have packed snow on them and there are many icy spots. If it warms up a bit, some of the snow and ice will turn to slush, thanks to the road salt that the city applies. Have you ever ridden a bicycle when temperatures are below freezing and the streets are a mix of snow and ice and salty slush?
Nevertheless, there are still a few maniac winter bicyclists out there, including a few of my friends (I'm a warm-weather cyclist myself). But they ride fat-tire bikes with studded tires, and they generally spend hundreds of dollars on specialized winter clothing. There's no way that any significant percentage of people would ride a bicycle in the winter in a place like this, even if it was motorized with a fairing and heated bars. (We call those "motorcycles;" they're common here in the summer, but not in the winter.) And there are lots of places like this.
Two things: one is that there are reports from people who have checked the math, and they claim that the bitcoin energy consumption is nowhere close to what's been reported. As far as I know there's just one source for the sky-high energy estimates (Digiconomist) and they admit that there's no way to verify their numbers. See for example:
https://www.cnbc.com/2017/12/21/no-bitcoin-is-likely-not-going-to-consume-all-the-worlds-energy-in-2020.html
Second, bitcoin and blockchain aren't the same thing, although of course bitcoin uses blockchain technology. Bitcoin mining is CPU-intensive on purpose -- that's a design goal that's central to how it works -- but blockchain systems in general don't have to work the same way.
His fault; Musk himself blurs the distinction between robots and AI. From a speech of his this past summer:
“I keep sounding the alarm bell, but until people see robots going down the street killing people, they don’t know how to react, because it seems so ethereal.”
> It's a Shriek (the exclamation mark in that equation)
When I was in school (decades ago) I also heard it called "admiration" -- that is, some people pronounced 3! as "three admiration." I was told that was a British thing (I went to school in the U.S.)... is there any truth to that?
> it would be useful for funding niche/hobbyist sites and would be OK with it provided the user is kept informed
If managed properly (maybe via "official" browser or protocol support) maybe this could be a way to do truly unobtrusive micropayments? I.e. I let you use 10% (or whatever) of my CPU cycles for the duration of the time that I'm visiting your website, and in exchange you show me no ads and you collect no data about me.
Re: doing an unscheduled test with no advance notice: eventually you *will* face this sort of test, whether it's your hand on the breaker, or the hand of Chaos. Given that this is the second time that SF has failed the test, it's hard to blame anyone other than SF themselves.
Many of those who tried to build the first airplanes failed because they made aircraft with flapping wings, like birds have. But it turns out that even though birds fly very well, imitating them is not the best way to build a machine that can fly. Same thing with the horseless carriage -- our mechanical horses don't have legs like real horses do. And submarines don't swim like fish do. Based on this history, I suspect that if we ever succeed in building a machine that thinks, it will not do so using the same techniques that humans use.
> Following a firmware update sent late last week, the Trådfri smart lighting system will work with competitor Hue, Philips' wireless lighting system. Thanks to a more open approach taken by big tech companies, it also works with Amazon's Alexa/Echo digital assistant and the Google Home.
Does that mean that a security hole in any one of those systems other can now be leveraged to compromise Tradfri as well?
It's a lock, fer the cryin out loud. I wonder why a door lock would need a software upgrade in the first place -- how complicated can the software be?
It would be interesting to see the list of bug fixes that the firmware upgrade was intended to address. Maybe the CPU in the lock is mining bitcoins for the company in its spare time, and they had to introduce new logic to deal with the recent bitcoin forking?
> I know commenters here are rightly suspicious of big-business but to say "you can bet they're doing it" is tinfoil hat zone
I agree. If they were doing that, then a number of Amazon employees would necessarily know about it. If just one of those employees ever became disgruntled, that person could pretty much destroy AWS by spilling the beans. Way too risky.
@John: yes, sadly this kind of thing happens all the time. I lean toward the right myself, but I remember back in the fall of 2001, after 9/11, telling my like-minded friends that we would surely come to regret the Patriot Act and all its trimmings, as soon as some left-wing liberal became president and took control of the machinery. And sure enough...
"Will this still sound like a good idea when the roles are reversed and the other side takes power?" I wish more people considered that when making legislative proposals...
> Nice one, Mike! Keep going!
Bah. He's as bad as Trump. He took a good idea and chose to frame it as a partisan taunt, wording it in a way that guarantees it can never pass a Republican Congress. That's just a stunt, it's not leadership.
If he was serious, he could and should have written the proposal using neutral language so it actually had a chance to be considered and maybe passed into law. But it so much easier to just bluster and taunt, isn't it? And people apparently love that, as evidenced by all the up-votes on the "nice one" post.
Putting partisanship aside (something that Congressman Quigley should have done if he really wanted this proposal to be taken seriously), it's a good idea. I can imagine the knee-jerk responses -- Republicans trying to shoot this down because it targets Trump, Democrats trying to keep it going in order to annoy Trump -- but if the shoe were on the other foot -- if a Democratic president had taken to Twitter like Trump has -- you can imagine that the roles would be exactly reversed.
Many people do take Trump's tweets seriously, they're unquestionably a crucial element of the current political discourse in the U.S. Whether you agree with them or not, officially archiving them is the right thing to do.
It would be interesting if accusations like these also included a line saying: "And in that time, we ourselves have mounted or sponsored X number of attacks against Russia and China, using the same definition of 'attack'."
Even if the number X is zero, it would be interesting to hear the government state that out loud, with a straight face. (And then to hear them explain why they believe zero is the most appropriate value of X.)
According to the U.S. Securities and Exchange Commission website: "Illegal insider trading refers generally to buying or selling a security, in breach of a fiduciary duty or other relationship of trust and confidence, while in possession of material, nonpublic information about the security." So you'd have to argue that MedSec was in a "relationship of trust and confidence" with St Jude, which seems unlikely. They're slimeballs, but this isn't insider trading.
> If the fridge was able to read this info and then you planned your meals for the week, it could tell you what you needed
I've seen comments like this in quite a few places... but does anyone actually pre-plan a week's worth of meals in any kind of detail?
And even the warnings about expiry dates don't seem all that useful. In my fridge, leftovers (rather than packaged products straight from the store) account for most of the stuff that's in danger of going bad. So the AC said, I'd have to tag them myself if I wanted warnings, which most people would probably be unlikely or unable to do. (How long does half of a tuna casserole last in the fridge? It depends...)
> The fascination with the Red Planet is driven by tantalising evidence that life may have once existed
This has always bothered me. I'd still be very interested in, and very impressed by, these remote exploratory missions even if we were absolutely 100% certain that life had never existed on other planets. I wonder sometimes whether we're missing out on other significant discoveries because of this over-focusing (in my opinion) on the Search for Extra-Terrestrial Life. Was there ever life on other planets in our solar system? That's a very interesting question, but it's not the only very interesting question.
Exactly! Many people don't realize that small-scale solar (a few hundred watts or less, as in the car-roof example) costs very substantially more than grid power. Or they don't realize that "it costs a lot more" means "it requires substantially more resources to build and maintain" -- because to a first approximation that's what cost is, a measure of the resources that are required to create and deliver something.
So choosing to put solar panels on the roof of a car actually wastes resources -- you're not helping the earth by doing that, you're just increasing your own resource footprint with a bit of conspicuous consumption. Telsa may well decide to offer the option anyway, but if they do it will be a marketing / business decision, not a sustainability decision.
I got 1 kWh by assuming that the roof of a car is about one square meter. Wikipedia says "Ignoring clouds, the daily average insolation for the Earth is approximately 6 kWh per square meter." Today's solar panels are about 15 - 18% efficient -- call it one-sixth, so on average you can expect to harvest about 1 kWh per day per car roof -- but only if the car is sitting under an unobstructed cloudless sky all day long.
There are are losses involved in taking the output from the solar panels to the batteries -- you have to adjust the voltage and depending on where the batteries are in their charge cycle you may have to control the current as well -- so the batteries won't actually get all of the energy that the solar panels capture, but I'm ignoring those losses for the purposes of this discussion. Like I said, the really big loss in a practical system comes from shade -- clouds and trees and such.
How did you arrive at the 65 kWh number?
When you think about it, 12 cents per kWh is just absolutely freaking amazingly low, especially when you consider the geographic reach of the system and the reliability that it manages to sustain, but that's the approximate average cost for grid power in the US today. I agree that it's not sustainable over the centuries if you have to get the energy from fossil fuels.
Thing is, you're probably not going to get half the average in the UK -- you'd probably be lucky to get one-fourth the average. Juan in Malaga will do better, but in a climate with average or worse-than-average cloudiness, you're not going to get as much power as you might expect. Have you ever tried it?
I have a panel similar to the one you link to, connected to a small charge controller to harvest the electricity and properly feed a battery. I've been using this setup to charge a couple of lead-acid batteries that run the lights in my one-room office (two 3-watt DC LED light bulbs) for over a year now, just as an experiment. I'm in the northeastern US, which is probably similar in cloudiness to you in the UK, and I'd be thrilled to get half of the theoretical maximum rated daily power from my panel. As it is, on a sunny day in the summer I can get about 80% of the rated output, for about 4 to 6 hours. Outside of those hours I get almost nothing, partly because of the angle of the sun and partly due to shade from surrounding trees and buildings. And on a sunny day in the winter I don't even get 4 hours of usable sun, since the it's so much lower in the sky.
Shade just kills the output -- any shade at all. Clouds, or trees, or buildings, or any other source of shade. One surprising aspect of solar panels is that if you shade even a small part of a panel, the output drops dramatically -- way more than you'd expect. A panel that's rated at 50 watts may output 40+ watts in full sun here where I live, but if you put half of your hand in front of the panel, shading maybe 5% of of the panel area, your output drops to 5 or 10 watts. This is why I don't think most people in most places would get even half of the maximum rated power -- is the entire roof of your car really in full sun half of the time between sun-up and sundown, on average?
South-facing solar panels on the angled roof of a house in a very sunny and treeless area -- those can work really well. Tilted to capture maximum solar energy, no clouds, no shade from trees or other buildings -- that's just ideal. But the roof of a car in most places, not so much. Based on my own experience, I just don't think you'd harvest enough energy to be worth the expense.
You get that 657 USD only if the sun shines down from a cloudless sky for 15 years, and there's never any shadow on the car, and it's never parked in a garage. How realistic is that? I bet you'd be lucky to get $200 worth of electricity in real life, which would just barely cover the cost of the panels.
Now if he really meant the roof of your house rather than the roof of your car, that's more sensible -- but with today's technology it's still not economically viable in many areas unless there's a subsidy involved.
I admire Musk's achievements to date, but some of this seems less than realistic. Solar panels on the roof of a car? If you do the math, the best you can expect from a car that's out in the sun all day under a cloudless sky would be about 1 kilowatt-hour per day -- about 12 cents' worth of electricity. Under real-world conditions you can expect to get considerably less than that. Hardly seems worth the effort.
I think they actually *can* pack heat in most places, but only if they're properly certified. Here in Pennsylvania, for example, you have to have "Act 235" training and certification to be an armed security guard. A mall could choose to hire Act 235 guards, but they'd have to pay those guards substantially more than minimum wage.
Indeed, but I personally would still be interested in a poll asking how people would vote if they could do it all over again. Here in the US, the media is full of stories about how most of the crazy Leave voters are already regretting their votes; I'm really curious to know if that's true or not.
Maybe instead of bi- or tri- or quad-modal it makes more sense to think about a continuum. At one end, when a project or idea is new, speed and ease of experimentation are paramount. So move fast and sure, go ahead and break things from time to time when you're at that stage. But if the project succeeds, people start depending on it, and stability becomes more important. If you're doing it right, there's a natural progression rather than a fixed dichotomy.
This is true even for new, disruptive businesses. In the early days of, say, Uber or Twitter or Facebook, an occasional burst of unexpected downtime was probably quite acceptable if that was the cost of rolling out new functionality at a rapid pace. But now that those companies are successful and established, they're probably much less willing to risk unplanned downtime -- service disruptions now trigger immediate snarkery from world + dog + vulture. Or think about AWS -- an occasional glitch in some new experimental / beta service is to be expected, but we expect EC2 and S3 to remain stable and reliable all the time.
So you evolve -- you use agile when it's appropriate, and you use stricter change control when that's appropriate. Different processes and standards for different stages of the project's lifecycle -- a continuum as the risks and costs and opportunities change over time.
You can't really trust those test -z's... if $1 and $2 are space characters, or slashes, or periods, (and I bet there are more), you still lose. And you probably can't afford to perform that careful testing procedure for every new set of $1 and $2 values.
Writing "rm -rf $variable" does require a certain hubris.
To put it another way: who's going to be interacting with that single pane of glass? Is there any one person in your enterprise who has full permissions on all of those different cloud resources? Is there even a person who has read-only permissions on all of those resources? Should there be?
The Register - Independent news and views for the tech community. Part of Situation Publishing
Biting the hand that feeds IT © 1998–2021
