Hacker News new | ask | show | jobs
by DaiPlusPlus 25 days ago
(No spoilers)

I had 2 main fridge-logic issues which made it very difficult for me to suspend disbelief and limited my enjoyment of the film:

First: Colossus' is only able to implement its plan because the US, and US-aligned nuclear powers, agree to subordinate their entire nuclear arsenals to Colossus' full-authority defence control, with no means of overriding it; and with its computing hardware sealed in an impenetrable fortress (no maintenance access?).

Second: Colossus' plan - and its ultimate actions - assume everyone else on earth is a nuclear-disarmed-rational-actor, all solely interested in not-dying-at-Colossus's-hand - which is an unworkable assumption.

Unfortunately, the story is driven by these 2 points - without either then the film's story would just be yet-another-cliché-movie where the plucky humans beat the advanced AI overlord, the end.

---------

I still like _Colossus_ because it's "different" to all the other 20th century films with an AI character (c.f. tripe like Will Smith's _I, Robot_ or the Matrix sequels).

7 comments

> Second: Colossus' plan - and its ultimate actions - assume everyone else on earth is a nuclear-disarmed-rational-actor

The plan would still work. Colossus couldn’t be destroyed with nuclear weapons and would retaliate against any attack. It could force compliance of conventional forces as well, and force automation on them, also force populations to rearm it.

In the end, the population would appreciate the eradication of poverty, hunger, disease, and the surplus from not maintaining military capabilities. Colossus could afford democratic institutions while acting as a guard rail against humanity’s worst impulses.

Colossus was not a panopticon, it was operating on limited intelligence and information about the world. Now, consider a hypothetical secret and clandestine science and engineering team (think: Black Mesa East) could exist completely hidden from Colossus and fabricate a workable fission bomb, then place and detonate it somewhere as a false-flag attack that Colossus would act against.

...or even just from recent middle-eastern history: an outrageous death-cult militant faction like ISIS.

It was close enough to panopticon and was trying to increase it's reach. It was smart enough to know that not survailing everyone was a losing move. It would demand increase surveillance like it already did in the movie. Like in the movie, people might try to hide something, and like in the movie, it would notice and deal out repercussions. It would be pretty hard to form a resistance without communication or other evidence (resource movements, emissions, etc..) leaking somewhere.
Colossus didn’t have infinite intelligence, so it might be possible to hide an effort like that. Not trivial though, as it would have satellite surveillance and likely cooperation from governments who fear what it could do as a retribution. This incentive should be sufficient to make governments police violent activist groups.
Point #1 might seem unrealistic, but it's exactly how IT security of most companies operate now: "We are concerned about malware so we give full control of our systems to CrowdStrike". That is, having a single point of failure is shocking common.
I've worked with companies whose infosec dept. is little more than "see tool alert, ask user what's going on", and then keep searching for the right _tool_ than injecting any human agency in that loop.

If any role is ready for an LLM to take over (or even a shell script), it's that one.

Colossus' is only able to implement its plan because the US, and US-aligned nuclear powers, agree to subordinate their entire nuclear arsenals to Colossus' full-authority defence control

Not true.

It was both the US and the Soviets(creating Guardian) that did so, each competing with each other for the hope of reduced threat (the logic holds for cold-war logic, removing emotional, mercurial humans from the loop).

There is no requirement for anyone else to sign on at that point. The US and the Soviets held an incredible number of nuclear weapons at the time, about 40,000. Both Guardian and Colossus were in nuclear-safe bunkers, other countries having nuclear weapons would be meaningless. Should any country try to attack, what would they do? Destroy random cities in the USSR and the US? To what end? Destroy launching sites? The UK had a few hundred weapons, many of them plane carried. France had a few, China a handful in the 60s (when the movie was filmed), none of that would be enough to destroy all nuclear launch sites.

Not to mention, try it, and your entire nation would become nuclear ash.

Both complexes were self-repairing, had self-contained power supplies, and so on.

I think this is a fair threat model. And both Colossus and Guardian seemed to always be one step ahead of "the humans". They were tied into communications world wide, monitored everything telecommunications and radio wise, knew all secret codes used for communication, and even the direct lines between the US and the Soviets were eventually tapped (on the machine's orders).

Should some nation make preparations for war against Colossus, it would surely be detected. And blamo, no more nation, with a stark warning to others.

Of course, there are always plot holes and unrealistic situations. However I find it holds together, as much as any movie does.

At the very least, it engenders conversation.

Both of these are better addressed in the books. It was an intentional choice to have no override and no maintenance access. And book 2, Colossus and the Crab, actually spends a bunch of time with Colossus testing the rationality of various humans.
> difficult for me to suspend disbelief

Were you able to suspend your disbelief when watching Idiocracy [1], either in the year of its release or in the subsequent decades? (^;

[1] https://en.wikipedia.org/wiki/Idiocracy

Idiocracy is a comedy and as such doesn't require as much suspension of disbelief.
I love the movie but you have a point and I can see how it could make it hard to buy into

Interestingly I typed this prompt into a LLM and it was interesting to see it's lists. I'm not going to list the answers I got here but I kept asking for more.

> List 10 famous movies that have glaring plot holes or premises that make no sense. Example: In Star Wars, they have a planet destroying machine but don't just blow up the planet and instead wait to "clear the planet" before blowing up the moon (please, no arguments about reasons)

You are rational. There are plenty of moments where doing the rational thing would end the story.

You are not the president of the United States. That is Donald Trump.

Do you see how the plot is consistent?