AI will be fantastic
Once AI becomes a thing, there are all sorts of great things it will be able to do aside from finding Sara Conner. The trick isn't figuring out what it can do. The trick is making AI a thing.
54 publicly visible posts • joined 1 May 2007
In the US, competition between cable companies is nonexistent. Local governments define exclusive areas that the companies operate in, so it's not like finding myself (a Time Warner customer) suddenly a Comcast or Charter customer would be something I could do anything about. I can get whatever is in my area, and that's it.
Applications use a lot of hardware-specific code to improve performance. Unless the OS adds an emulation layer that will keep those apps from working. You won't end up with the kind of performance you get from current VM hypervisors in those cases. Those benefit from the emulated VM architecture and the hardware being the same, so the machine code doesn't need to be translated instruction-by-instruction.
It would be really difficult to manage a different CPU architecture on one segment of it's PC line. You'd have to run current apps in emulation (something like Rosetta) with the performance hits that involves and get developers to target two different architectures for the foreseeable future. All emulation, all the time, on systems that are constrained in CPU performance, RAM, and storage is asking for a really bad user experience. Apps that are 50% larger (to accommodate the duplicate binaries) complicate things even more.
Speaking of those binaries - a quick bit of gooleing seems to indicate that ARM CPUs are bigendian. Endian issues made file compatibility between Apple and Windows versions of the same program occasionally problematic.
This is exactly my experience too. I use and support PCs at work, but all my home systems are Apple (2 iMacs and a Macbook Pro). The only issue I've had with any of them is a failed hard drive, and time machine backups made that simple to recover.
OSX is a lot less forgiving of hard drive issues than Windows is. A bad external drive will halt OSX, preventing recovery of the data. With Windows you can usually recover stuff from a failing drive as long as you avoid writing to the drive.
Don't know about him, but my views after having used it for 4 months:
Metro problems:
- Every icon normally added to the start menu gets sprayed onto the Metro start screen after an install, so I have to keep re-arranging the icons there.
- Some metro icons can take two tiles, Some can't. It's not clear why that is.
- Changing a shortcut's properties means locating the 'file' it's associated with, then changing the properties of that file.
- Lesser-used icons for programs that I have installed fill (on my 1920x1200 resolution screen) 4 screens of (largely) disorganized icons.
- Search is useful if and only if you know the name of the lesser-used program you are looking for
- But not if that program is an 'app'. Those sort and display differently.
- Some computer settings (lock screen, user accounts, user account types) are available in a Metro settings panel. Some of those are also available in the control panel but with different options.
- Every time a Metro app that can read different file types gets updated, Windows asks me again if I'd like to use it to open this file I clicked on.
General uselessness:
- Management of networks (wireless or otherwise) is entirely at the command line using the netsh command. Windows will still 'see' the same network as 'different' from time to time, so you'll still see connections to 'network 12' or some such.
- Can't change an automatically-detected 'private' network to a 'public' or 'home' network or vice versa. Those settings still affect the firewall rules though, so if it's mis-detected you get to use netsh some more.
- I keep turning media sharing off. It keeps showing up in the network locations, with one instance for each user account. None of them work though, because media sharing is turned off.
- Changing file file associations works for some applications and file types. For others it doesn't work. If the file name of the program you want to use happens to be the same as the one you used to use, it will never work. This is even if you change it in the registry.
- Desktop 'charms' mean that overshooting into a corner (desktop or Metro mode) mean that you can (and will) accidentally activate something you didn't want to at some point or another.
Weirdness that isn't well documented:
- The system-wide search tool is the same search tool you use to search inside Metro apps like the Microsoft app store. This isn't mentioned anywhere that I could find in the Windows help. Google found that though.
- If you want to open a second instance of a program (a second notepad window, for example) the process is:
- Go to the metro screen
- Right-click on the app tile
- Select 'open new window' from the charm bar at the bottom of the metro screen.
Stability:
- Hibernate, a feature I've been using since Windows XP, is now so unreliable that I have disabled it. The 'clue' that Windows gives you if it can't restore your state is particularly helpful - it turns the computer back off. Unless you are in a dark room, it's easy to miss that the computer you just turned on isn't on any more...
That's just off the top of my head. I've got more if you want them.
I initially 'upgraded' so I could determine if it would be useful to roll this out inside my company. After using it for a while, I can't imagine a time when I'd want to do that, so we will be skipping this version and advising others to do the same. Maybe Windows 9 will not get in the way as much as 8 does. Maybe we'll all just switch to Macs.
" Apple's closed garden only-on-our-device biz model is starting to work against it."
You could replace 'Apple' with 'Amazon' or 'B&N' in that sentence and it's still accurate. It's not Apple that requires DRM, and DRM isn't something that only Apple does. The Amazon content market is every bit as closed as Apple's is.
In fact, when talking about tablets as e-readers, iPads are actually more open than a Kindle or Nook. I can (and do have) Kindle and Nook e-reader apps on my iPad. I can't get Nook for my Kindle, Kindle for my Nook, or iBooks for any of them.
Solar cells on an aircraft would be a tremendous boost for the manufacturers of solar cells, though homeowners along the flight path might not appreciate the constant rain of glass and silicon...
Not so great for the passengers would be the consequences of an uncontained engine failure in an engine that's buried inside the fuselage or at the wing root. Converting a serious but controllable failure into the loss of a wing isn't much of a win for anyone.
Also missing form the plan is how these aircraft - launching from short runways via catapults - might then manage to land on those short runways in inclement weather. Arresting wires gear and tail hooks perhaps?
As the article itself pointed out, advertisers agreed to honor 'do not track' as long as it was not the default setting in browsers. If Microsoft makes it the default, they have broken that agreement. I can't imagine that advertisers will feel any compulsion to hold up their end of the bargain if Microsoft doesn't.
It would be awesome, if people were all good and trustworthy. In practice you'll have hackers who crack the scheme and sell those cracks, thieves who don't crack the scheme but claim they did, and con men who'll sell codes for CPUs that can't be upgraded. All of which will chip away at Intel's goodwill.
"Until Microsoft do not force you to use IE to run Microsoft update, then IE Will be required."
In Vista, Windows Update is a control panel applet. If you go to the Windows Update web page on a system running Vista, you get a page telling you how to find it in the control panel. I presume Windows 7 will work the same way. Not only is a browser not required, it's not supported or (as far as I can tell) possible.
I wonder if any of the people complaining about Microsoft's decision have ever dealt with large deployments of software, or have been involved in shipping anything. It's a lot more complex than just slapping a few stickers on it and changing a script.
And don't forget that Microsoft is at or very close to RTM on this product. Removing features is lots simpler than adding them, especially if Microsoft has been removing dependencies on IE - a likely goal of the Vista and Windows 7 development process, given the legal problems including it has already created.
Not likley, unless they rip out all of the interface nastys and replace them. WinMobile is way too dependent on a stylus in it's touch-screen version, and makes dumbphone interfaces seem like models of clarity and logic in it's other incarnation. I don't think Steve B's Microsoft can do it. Bill G's, maybe. Not Steve's.
Phishing scams are not really a security issue that can be fixed with better software engineering. Genetic engineering, maybe, but not software.
In any event, anti-phishing 'toolbars' can cause more problems than they prevent because the presence of them fools the gullible (who are most at risk for such scams) into thinking that 'if the browser says it's OK, then it's OK'.
I think we've replaced .dll hell with .NET hell. After 4 versions of this nonsense, the .NET enhancement I'm really looking forward to is the one that doesn't require me to install a new version of .NET alongside the other 4 versions I can't remove because newer versions won't work in place of the older ones.
"someone will have to pay to maintain the plant even when it no longer produces leccy, and they will have to keep paying to maintain the plant for fifty to one hundred thousand years. That is longer then human beings have been recognisably human.
Still think they are a good idea?"
Yup. The other options are more stupid.
Humans are pretty clever. Clever enough that I'd be surprised if they didn't figure out something useful to do with that waste long before the 50-100K years it would take to decay on it's own.
What I really like to see are some counter-proposals from the groups who are so dead-set against things like this. Not fuzzy-headed ones that won't scale or aren't proven yet. Ones that can produce in the multi-gigawatt range, and can be on line in a few years. We don't need pilot programs. We need to break ground on something that will work, and we need to do it soon.
Re: Overpriced [...]
"Bah to overpriced software. I'm very surprised that they aren't quaking in their pants since open source CAD software that doesn't cost is cent is now widely available."
Have you actually tried using any of them?
Re: Lock the software
"Why don't they just put in mechanisms that make it hard to resell the software.
Require registration online and lock the software to the hardware in the machine."
Autodesk does exactly that. Unless the software being sold is unregistered, I don't know how this guy's customers are going to be able to actually use it.
No real mystery why sites are vunrible to this kind of thing. It's likely to be one of:
- Security 'experts' and programmers who are real good at interviewing and CV writing, but not so much when it comes to actual security. Doctors, lawyers, engineers, ect. - all of these professionals have licensing schemes that garentee a certain minimum level of competence. Not so much when it comes to software. Used car lots and flea markets have a higher level of regulation and official scrutiny than information security experts. They need business licenses and such. All you need to be a security expert is a business card and a phone number.
- Information security is one of those cost centers in a company that doesn't actually seem to accomplish anything (aside from being occasionally annoying) if they are doing their jobs correctly. People who aren't doing anything are the first to go when budgets get cut.
- SQL injection isn't something you can defend against with an automated tool or eliminate by putting a tick in a box. You have to watch out for it in your coding practices and page design. Like buffer overflows (the main source of security issues in software), the ability to inject unexpected SQL queries into a form on a web page is a bug. You can catch it if you are looking for it, but unless you are looking for it, it can escape your notice.
This kind of thing isn't new or even unique. Where was the shock and outrage when Java updates included broswer toolbars and Adobe Reader updates came with that started happening?
Seems to me that the 'real' problem is that Safari is finding its way onto systems via a route that the Mozilla project can't use, and that has the Mozilla project up in arms because it maybe means fewer Google Toolbar installs for them.
It's really only otto-cycle engines that are extremely picky about what fuels will work. They depend on thier fuel vaporising in the intake tract. Pretty much every other kind of engine can work with a wide variety of liquid fuels. You can find find diesels, for example, running happily on everything from natural gas to filtered crude oil. With a continuous-combustion engine like a gas turbine, you really only need to worry about things like 'will it pump OK' and 'does the exhaust corrode the power turbine'. If it didn't congeal into wax at altitude, heating oil would likely work too.
MS Academic licensing is much more complex than the Becta report indicates. There are four different programs, multiple tiers within the programs, and often confusing per-product rules for them. The Campus-level license program Becta references does require counting all systems on the campus regardless of the OS installed, but only requires the audit once a year, allows any number of machines to be added during the year, and includes CALs for the servers the school will (presumably) be making avalible to the non-Windows systems as well. There are follow-on arrangements for students who graduate as well as the personal systems of instructors and students - if it fits the school's needs, it's actually not all that bad.
Schools are free to select per-seat, per-product pricing if the terms of the campus agreement seem too onerous, or get on the SA bandwagon like ordinary businesses, but with academic discounts. If Becta wanted to point out absurd MS license terms, the requirement to re-purchase Vista (or acquire SA for it) for systems with OEM Vista licenses to take advantage of KMS or MKS activation and the lack of OEM-Academic OS product would be a far more damming topic for discussion, I think.
If my lifestyle choices don't match your lifestyle choices then it's my choices that are wrong? Things are a lot farther apart here in the western US than they are in Europe.
The real problem with pure-electric vehicles is that they don't eliminate pollution, they export it. The EV-1, when it was running off of Los Angeles power, was doing so by burning coal in Arizona. If you think that's green, you've got really funny ideas about what green means. Factor in transmission losses, thermal, gearing, and storage losses in the EV-1, thermal and conversion losses in the recharging arrangements, and thermal losses at the plant, and you've got a pretty inefficient way to move one or two people (with no luggage, mind you) around while introducing the fine people of Arizona to smog and soot.
What would a scalable solution look like? With current tech? I don't think there is one, actually. Converting a sizable fraction of the current vehicle fleet to EVs could easily double the load on the already over-stressed power grid. With fission and fusion off the table and all the useful rivers already in use, I don't know where that power is gonna come from. Biofuels, with the possible exception of some algae that Shell is working on, are actually more troublesome than coal (look up ocean dead zone for one reason). And don't even start with solar cells. Toxic waste from the manufacturing process and with energy break-even times of 5 to 10 years mean that's even dumber than using corn to make fuel.
Until something can be worked out, the best bet seems to be to stretch what we have. That means oil-fired hybrids with electrical drive trains. The EV-1 was nothing more than a way to see if an electrical drive train was fesable. It wasn't meant to be a solution, because it *can't* be a solution.
someone looked at a biofuel idea that might actually scale up big enough to be a solution. Using food crops for fuel is dangerously short-sighted, because it will never actually work and it keeps people from looking at things that might.
@Alan:
The test is being conducted in Hawaii becasue the alge they want to use is native to the islands. If it works out, they will undoubtedly try to figure out how to safely re-locate the alge or do further testing with other strains. While it's still in the testing phase, though, it would be a really bad idea to take that alge and try to establish it elsewhere. Search for 'kudzu' or 'zebra mussel' for some ideas on why that might be. Besides, parts of Hawaii are not as expensive as you might think. Lots of Hawaii (especially the big island) is still quite rural and undeveloped.
@biodiesel coward:
Biodiesel only works becasue very few people are interested in running it.
A (really important) point lost on non-US readers is that Verizon's network is CDMA/EDVO, not GSM. Cool phones from South Korea might be possible, but to get an iPhone or just about any Nokia onto the VZW newwork would take a bit more work. With a soldering iron. Good luck with that.
Verizon's phones are all sold unlocked (contract or not), and Verizon will already activate anything that can talk to thier network that has a valid, known ESN. Of course nothing but a VZW phone can actually talk to their network, and the phones are completely useless on other networks.
It's really funny to see how much mileage Verizon is getting out of the media on this though. The Euro press, used to a mature and functional wireless market, see this as something new and exciting - look at those cheeky bastards getting the drop on Google! While the US press - mouth-breathing morons for the most part - regurgitate the press release. Mention that 'open network' + 'no frickin hardware' = no change and their eyes glaze over.
The bit in the original press release about 'testing and certificaiton' is the real gotcha though. Take a look at Verizon customer forums to get an idea about how long that takes. I think they are still waiting for their first WM6 device.
...and spend all thier time at anchor because the'd be banned from every port.
Widespread civil use of nuclear power can only happen once all the evil and/or stupid people have been eliminated. Of course if you eliminated all the evil and stupid people, the dozen or so people left won't really need nuclear power.
Apple's documentation on TV out for the classic:
http://docs.info.apple.com/article.html?artnum=300233
Apperently, the $100.00 A/V kit works, as do the $50.00 universal dock and the new $50.00 A/V cables (neither of which are currently avalible) but the $20.00 cable that worked for the 5G doesn't work for the classic.
The classic's world clock is broken, too. The DST settings is applied to all of the displayed clocks, wither it's appropriate or not. i.e. I have one clock set for Los Angeles (a DST zone) and another set for Honolulu (no DST in the tropics). The classic insists there is a two-hour difference, but that's only true in the winter months.
Not selling through iTunes isn't setting up to fail. Not selling through iTunes is more like trying to find a way to remain relevant in digital music when Apple is doing all the distributing.
UMG wants differential pricing. Apple won't agree. With DRM, UMG can't can't go anywhere else, and has to accept Apple's terms. Remove the DRM and Apple isn't calling the shots anymore. The only question remaining is 'is it too late to build an online marketplace for music that doesn't include Apple'
Love Eminem or hate him, he has a point. If what I purchase at iTunes is a single-user license that I can't sell or give away, why does his label pay him the lower rate associated with physical objects? i.e. End users get all the usage limits of 'licensing' while the artist gets the income limits of 'sales'. Why is that fair?
When the labels are actually manufacturing things, they are taking a sizable risk. Make too many, and you've got millions of objects out there taking up space that need to be moved hither and yon. Don't make enough and folks will loose interest before you can get more of them made and shipped. The artists involved don't share that risk, which is why the labels haven't given them much of the profit in the past.
With downloads, that's all different. The label produces one perfect digital file and sends it to Apple (or whoever) to make copies as needed. No manufacturing risks at all. The label is still spending on advertising and such, and should get a fair return on that, but the return should not be the same as the return for dealing with a physical product becasue the risks aren't the same as dealing with a physical object.
If an artist has different terms for 'license' and 'sale', then iTunes 'sales' ought to be conducted under those 'license' terms. Because that's what it is.