Top 5 Most Scientifically Plausible Sci Fi Movies

[charlemagne]

The author fails anyways because of this:

Of course, some science fiction films are more plausible than others: Would the Galactic Empire really have enough manpower to build (and then rebuild) the Death Star in the Star Wars films?

There is no reason to not believe that they have this kind of manpower, since none of the movies give the impression that the Galactic Empire has a tiny population or doesn't have much industry. If they guy really was about scientific plausiblity his thought should have been "We see them build (and then rebuild) the Death Star, so they have to have enough manpower to do this."

[NecronLord]

Yeah. About the worst thing, science wise, in Deep Impact was that the Messiah somehow survived hitting pieces of ice the size of buildings.

[RedImperator]

Yeah. About the worst thing, science wise, in Deep Impact was that the Messiah somehow survived hitting pieces of ice the size of buildings.

Well, there's also the fact that when they blew up the comet in low Earth orbit, the comet-mass worth of fragments didn't hit the atmosphere and cause a mass extinction anyway.

[Admiral Valdemar]

Yeah. About the worst thing, science wise, in Deep Impact was that the Messiah somehow survived hitting pieces of ice the size of buildings.

Well, there's also the fact that when they blew up the comet in low Earth orbit, the comet-mass worth of fragments didn't hit the atmosphere and cause a mass extinction anyway.

Additionally, the comet should have been darker, like, dark enough that you can't even see the ground when they did their EVA on to Wolf-Biederman.

Aside from the comet issues, the technology and plan of action was far more realistic than the X-71s and 800 ft subterranean detonation in the Bruce Willis flick. Deep Impact should have borrowed the blasting-two-pieces-apart approach from the latter, though, since turning one solid mass into many smaller masses still flying into Earth's atmosphere is probably even worse than what they nearly had happen.

[Uraniun235]

HAL going rogue isn't very plausible, but I think the book was written before "crazy AIs" became popular.

Eh. Remember, Hal is not one of a kind. There are lots of AIs in the 2001 setting. He's the only one of his class ever to even make a mistake. I would be very surprised, if, out of thousands of AIs, not one of them fouled up when given irrationally conflicting goals.

If I remember right, in one of the books Clarke describes that after HAL's malfunction they tried deliberately giving conflicting goals to other 9000-series computers and found that it almost always resulted in the same sort of failure that HAL experienced.

[DPDarkPrimus]

Alien fails hard on the mass issue. There is just no way that creature could have grown that fast, before it starts eating crewmembers. Hell, even with unlimited food available I very much doubt it's biologically possible to go from a rat sized creature to a human-sized creature in a couple of hours, no matter how crazy your biochemistry.

Considering all SF films are going to have a degree of implausibly, this isn't a particularly large strike against it, since it's a creature unlike anything we've ever seen.

Also, Alien has FTL, of the standard sci-fi type (which makes no particular effort to be plausible).

FTL is only mentioned in passing, never shown, which makes it more plausible than any films with more specific mention or visualization of it.

Much as I like the sets, the equipment design looks horribly dated now, much more so than SW, and it's hard to imagine 1970s computer technology still being in use a century later when hyperdrives and fusion reactors are common technology.

Now you're just meta-bitching.

Then there's the super-acidic blood, and the pretty much chemical impossibility of making a living organism based on it.

As far as we can conceive, anyway.

[Admiral Valdemar]

It was a logical paradox, so I'd imagine any sapient computer would either do what HAL did, or crash/request new instructions. It'd be analogous to the reasoning used by the supercomputer in the I, Robot movie. In fact, I recall Skynet performed the actions it did out of some misplaced sense of making the world safer too. Only unlike the former instance with US Robotics using machines to put humans under house arrest, Skynet removed the human threat to stability, same as HAL.

[Samuel]

It was a logical paradox, so I'd imagine any sapient computer would either do what HAL did, or crash/request new instructions. It'd be analogous to the reasoning used by the supercomputer in the I, Robot movie. In fact, I recall Skynet performed the actions it did out of some misplaced sense of making the world safer too. Only unlike the former instance with US Robotics using machines to put humans under house arrest, Skynet removed the human threat to stability, same as HAL.

Why? Human brains don't crash from logical paradoxes and I can't imagine AI designers making something that can be defeated by Kirk asking hard questions. Couldn't they just copy whatever system we use?

[Admiral Valdemar]

By crash, I mean be unable to carry out orders because of a conflict. The computer wouldn't explode hilariously like in that Kirk example. Either it stalls and doesn't do anything, or it notices a problem and requests clarification or it pulls a HAL and inadvertently counters one mission objective with another, despite both being mission critical.

[Starglider]

Why? Human brains don't crash from logical paradoxes and I can't imagine AI designers making something that can be defeated by Kirk asking hard questions.

You have a vastly overoptimistic assessment of AI designers. Most of the current researchers take the attitude 'hey, we'll throw together some algorithms that sound cool, and see if it does anything intelligent, and if it does we'll do some minimal testing, put it in a box and sell it'. No kidding.

Secondly, this shit is hard. Simple 'all lawyers lie' stuff isn't too bad but goal system instability (essentially the problem HAL had) is a very very hard problem.

Couldn't they just copy whatever system we use?

Currently, no one knows how the human brain works. So no. Also, a lot of the human goal system is genetically determined, with cultural learning layered on top of that. The basic drives have been tested, balanced and debugged by millions of years of evolution, and we still get psychos from time to time. In an AI (that isn't an outright copy of a human brain), you are defining the 'basic drives' as part of your design. Even with a irrational-but-gracefully-degrading human-like processing system, chances are you'll introduce some kind of instability or incompatability that will lead to unexpected (i.e. bad) results.