Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?

Comment What a moron. (Score 2) 298

even if we were able eventually to create a superintelligence, there is no reason to believe it would be bent on world domination, unless this were for some reason programmed into the system.

Yeah, see, nobody, to a first approximation, is worried about a superintelligence having "world domination" as its intrinsic value. They're worried about a superintelligence adopting world domination as an instrumental value to achieve the end actually programmed into it. If whatever goal actually implemented by programmers and trainers in the superintelligence's code (bugs in implementation and all) is most easily achieved after eliminating the ability of humans to thwart it, then a sufficiently-smart AI carrying out that programmed goal will try to eliminate the ability of humans to thwart it.

The worry is not that AI will be evil, or even directed to do evil by its creators. It's that programmers are notoriously bad at writing complex code that has no unanticipated behaviors, and superintelligent AI will inherently be complex code.

And unless superintelligent AI turns out to be intrinsically impossible, the only question is when, not if, we have to deal with the problem of writing safe superintelligent AI.

Comment Tried, wasted hundreds of billions, and failed (Score 1) 315

Seriously, we already tried federal funding of broadband expansion. All it did was fill the pockets of telecoms; the problem still exists. Why would you expect another attempt to do particularly better? Because Trump's people will do it right?

If you're going to do anything, don't even consider the supply side at all. Set up a program on the demand side where sufficiently-rural addresses can apply for subsidies toward Internet access. That'll make fundraising for the OneWeb and SpaceX constellations easier while letting the individuals get on with HughesNet and Exede right now.

Comment Re:Not a protest (Score 4, Insightful) 465

Ah, yes, who can forget about how all those angry Tea Party protesters, which of course included people with anger issues, resulting in all sorts of rioting.

Oh, wait. Apparently it's actually possible to have protests that don't descend into riots. It's not natural or inevitable, it's just an effect of whether the "protesters" are decent people or scum.

If rioting breaks out at a protest, it's because the "protestors" are choosing to aid and abet violence. There are no innocents at a riot, just co-conspirators. Lock up all the scum.

Comment Re:So Make Hydrogen (Score 1, Insightful) 324

Because, you know, it's a totally worthwhile capital investment to make massive desalination capacity that you run a tiny percentage of the time with no relationship to the demand for water. Tell me, are we also going to pay the workers to stand idle, or will we just expect there to be a bunch of trained unemployed people living nearby that we can hire to staff it when the power's available?

Comment Re:Crap (Score 1) 381

Why is google all of a sudden making absurd, terrible design decisions in its news division?

I assume that they got a new design team in, people who neither actually read news nor remember what a mess last redesign (2010-2011) was. The inability to tell the difference between clutter and information density proves the first (news is not a program UI, it's a presentation of data), and stripping all the features that had to be laboriously re-implemented last time indicates the second.

The real question to me is whether the previous redesign team was the more stupid & arrogant (they tested their new version a while, discovered everybody opted out of the test for the old one, and then decided to impose the new one without opt-out because they were sure it was just getting people used to it rather than major deficiencies), or if this redesign (imposing the new one without testing that would have told them people didn't like the feature-stripping) is more stupid & arrogant.

I guess that question will be decided by if and how quickly the new team restores article snippets, whitespace-sacrificing higher information density layout, real two-column view of news, turning off the sidebars, allowing standard Google search from the input field, and otherwise bringing forward all the first-implementation Google News features the second-iteration design idiots discovered too late that they had to add back.

Comment Re:Thanks for this article. (Score 1) 381

Will they listen? I don't know. Last time they screwed up News this badly (2011), they eventually listened to the angry people (including me) and added enough features back to make it as useful as the previous version. And the reaction on their product support forums has been next-to uniformly negative, just like last time.

On the other hand, the fact that they did this without noticing they were making the same mistake as last time, without an extended period of a/b testing, makes me wonder if they're too arrogant or stupid to listen to feedback this time. Do they have no institutional memory whatsoever, or did they go ahead in spite of institutional memory?

Comment Re:Because Microsoft has legacy business customers (Score 1) 367


You are partially correct in that you can't run real mode code in 64-bit mode, either directly or in a V8086. But real mode and 16-bit are not, in fact synonyms; the all-16-bit 286 had a protected mode, code for which which later processors were perfectly capable of executing in 32-bit protected mode.

That includes in the 32-bit mode on the x64 architecture. You can simultaneously run 64-bit and 16-bit code just fine, if the 16-bit code is protected mode code and the OS doesn't do anything stupid. And any Windows program that can run on Windows 3.1 is able to run in 16-bit protected mode.

The fact that you can't run a Windows 3.1 program on x64 Windows is a very specifically Microsoft fuckup, having to do with how the Windows-on-Windows software was done for Itanium, and then how the x64 version of Windows was ported from Itanium.

But you don't need a virtual machine if you want to run 16-bit Windows 3.1 programs under a 64-bit OS; Linux with WINE will run them.

Comment Re:For the Young... Some Background. (Score 2) 145

Um, no. The Joint Development Agreement didn't have anything to do with antitrust. (You're confusing that with why IBM didn't lock Microsoft into exclusivity in the DOS contract five years earlier, which in part was motivated because of the antitrust settlements on IBM mainframes that required IBM to make its mainframe OSes available.)

Microsoft's original plan for its successor to the limited DOS was a migration path to Xenix, but, when the 1984 AT&T antitrust resolution came down, AT&T got permission to sell Unix as a product. Microsoft decided it would be folly to try to compete with AT&T selling AT&T's OS, and switched over migration plans to a product called "ADOS" or "DOS 4" or various other names in the press. ADOS would then slip under Windows, also in development, which would be the GUI.

At the same time, IBM had been trying to develop its own improved extensions and GUI to DOS to exploit 286 hardware -- Top View.

After a fairly short period of the press speculating about the coming war between Microsoft and IBM over the future of the PC, and the initial failure of Top View to get as many sales as expected, IBM and Microsoft signed a joint development agreement for what the press would, during development, still call ADOS/DOS 4, and which was internally codenamed CP/DOS. The PC would have a single, obvious software future.

And when this OS was released as OS/2 in 1987, it worked just fine on non-PS/2s, which was only to be expected, because A) IBM was already committed to customers that it would work on ATs, which is why they wouldn't let Microsoft make it a 386-only OS; and B) Microsoft actually finished development of it (and released initial outside developer machines with it) on Compaq 386s.

And it was, in fact, that 286 compatibility that hampered it the most, because the 286 had no v8086 mode to hide DOS programs in. Thus the tendency to call the DOS box the "penalty box" By the time OS/2 1.1 shipped with the GUI in October 1988, Windows/386 had already shipped and, because it used virtual 8086 mode to multitask DOS, had better support for DOS apps than OS/2. Added to the ability to drop out of Windows just to pure native DOS if necessary, the installed base of DOS apps then won the day for 16-bit Windows.

NT didn't even release until July 1993, long after 16-bit Windows dominated desktops. And NT wasn't enough to stop OS/2 Warp from making a play, it was 32-bit-extended-16-bit-Windows 95 that shut OS/2's last charge down.

Comment Want to stop this efficiently? (Score 1) 95

It can be done pretty easily -- let people kill any wolves that come into populated areas, like they used to be allowed to do. The wolves that survive will be those that fear people and stick to the wilds.

And it's not like it would actually endanger the wolves -- the IUCN listing for them is Least Concern. The "Endangered Species Act" listing of them as "endangered" merely indicated they were rare in the lower 48 states; Canada, Alaska, Russia, and China have plenty.

Slashdot Top Deals

Just go with the flow control, roll with the crunches, and, when you get a prompt, type like hell.