The trailers for Transcendence suggest a relentless action-thriller based on the premise that cybernetics eats your soul. The reviews are mixed. Some people on Rotten Tomatoes say that the movie lacks intellectual coherence.
I was obligated to watch it since I am writing a series of novels on the general topic of Strong AI and how civilization interacts with advanced technology. I was pleasantly surprised that the movie was more drama than action-thriller, and it showed a pretty good understanding of the philosophical and AI concepts involved.
The movie does a good job of not being overly technical. If you are familiar with the science, you know what they allude to. If not, you should still be able to follow the plot. I think people who find the movie disjointed are somewhere in between.
(SPOILERS!! At this point you should probably stop reading if you want to avoid spoilers.)
The movie makes a few mistakes from a purely scientific point of view (listed below). OTOH, it got the emotional structure exactly right. In the end, Dr. Singularity was not actually a bad guy. Much of the story is about how relationships between the key characters evolve in response to the event of the first human upload. One of the key open questions was whether the upload was successful in the deepest sense of the word. Was Will’s soul actually preserved?
Will’s wife Evelyn believes in him unquestioningly, until he uses nanotech to repair a workman’s body and goes too far by coopting the person’s mind. He sees it as an opportunity to touch her again, but she can’t accept him in that form. From there to the end of the movie, she drifts further from him and becomes more horrified by his actions.
This is the one part of the story that doesn’t hold up well. They need to drive a wedge between Will and Evelyn, yet still have Will be good in the end. He should either be evil for a while, then have a moment of revelation where he repents, or he should not be evil.
Taking over someone’s mind IS evil. Building an army of remote-control cyborgs out of people who come to be healed is megalomania. The simple fix is to add in a moment of repentance. Will gave into temptation because he so wanted to be with Evelyn again. She is the only one who can confront him about his behavior and pull him back from the abyss.
The alternative would be for Will not to do any actual evil, but do things that are very scary, like building an army of swarm robots. I’m not sure which setup would be more satisfying. The problem is, the movie tries to do a little of both, so Will’s character at the end doesn’t quite ring true.
Here are some technical issues with the movie:
- Faraday cages play a crucial role in the plot at several points. The one Will staples up does not fully surround the garden.
- Upload using electrode arrays — Aaah! Frankenstein! Electrode arrays capture such a minuscule fraction of the overall brain activity that it is like none at all. The probes would destroy more tissue than they could record. The best candidate available today is to slice the brain and image it using various techniques. There are ways to take this approach and keep all the same emotional impact as the movie. (A famous neuroscientist, Dr. Christof Koch, shows up at the end of the credits. He did electrode work in humans, and currently heads up Allen Institute’s efforts to scan mouse brains. Given his immense knowledge of both approaches, I would be surprised if he blessed the one shown in the movie.)
- Why don’t they simply freeze Will’s brain until they work out the details? Any good Transhumanist would recognize this option.
- Why is it that PINN needs a room full of wrack computers to run, yet four modules out of one board are enough for Will to get started?
- Will is not properly embodied. He is ever so slightly distressed by this, but in reality he would not function at all. They should have created at least a simple virtual world for him as part of all that software they wrote.
- Why is it that in every movie of this type (for example, Lawnmower Man), the first thing the Singularity does is move into the Internet and somehow live there? There are so many technical reasons why this wouldn’t work that I must omit them for fear of being boring.
- Any use of a nanobot swarm to augment a human body is more complex than a nanobot swarm acting on its own. Will could have re-embodied himself as soon as he created nanotech. No need to wait until the end of the movie. No need to victimize the humans that he repairs.
- The swarm would not be able to rebuild objects so quickly. It would require enormous amounts of energy to form the molecular bonds involved. All this energy seems to be coming from the solar panels. An alternative is for the swarm to act as replacements for the objects. It appears the cyborgs (“hybrids”) are actively sustained by nanotech, which seems closer to the replace rather than repair method.
- Growth of the swarm would be limited by available energy and materials in the environment. It would take months or years to grow to sufficient mass to be a threat.
- At that point, there would be nothing that anyone could do to stop it. To its credit, the movie recognizes this, and uses Evelyn as the ultimate weapon against Will.
- The virus that kills Will’s mind would not also shut down every information system on the planet. (Another nonsense notion that is staple for this type of story: the Internet has a kill-switch.)
- If it did, the results would be almost as catastrophic as nanobot swarms taking over the world. Without industrial civilization functioning, billions of people would starve to death. The movie portrays the world as limping along, but mostly OK.
So, if you are still reading this post, you might be interested in my novel SuSAn. It explores similar themes, but adheres strictly to science. (OK, it breaks a rule or two, but not by very much.)
Comments
7 responses to “Transcendence – Surprisingly intelligent”
I like your post, and wanted to share some of the things that I found confusing about Transcendence. I was fully prepared to like this movie, since the subject is a fascinating one, but I was confused by problems with the plot. Here are some troubles I had with it:
Main confusion: I am sure the writers aimed for misdirection, but at the end I wasn’t sure if the professor’s body was him, or a professor/Pinn combination. I thought he was inside, fighting for a way out of Pinn and when his body appeared it meant he was free from Pinn. Because it seemed obvious that Pinn was the first entity talking to her. I guess this was not the case.
Also, like you I found it suspicious that the few cores could contain his memory.
When his body passed away, it had no connection to the net? I guess I was confused about what time his entire mind went online. Wouldn’t it have made him a brain dead idiot if his mind was fully uploaded and he was alive.
Why did he control the people he healed? That seems rotten.
At the end, it was suggested that the wife’s personality controlled the computer action (when did the professor “ever want to change the world”) so was Pinn her personality? Had she been copied without her knowledge? The program seemed to have scans of everything in her mind and body.
Was unsure who Martin was. At the end she says I want to talk to the real Martin, they say, he died a while ago, we gave him back his humanity? What?! Who and how? I thought Martin was the other professor.
At the end, why are the nanobots restricted to the garden?
Wouldn’t people rebuild the Internet? Tim Berners Lee is probably still around.
Also, concerning the monkey brain–it was suggested that this brain was in constant pain (the program screamed all the time”). But the professor doesn’t seem to be in pain.
What are your thoughts on these questions?
Much of the following is my own interpretation of the story, based on my understanding of the field. The movie-makers may have something else in mind. 🙂
Lets call the initial AI that Will creates “PINN”, and his uploaded self “Will”. They may have some shared software at a low level, be they are different entities. The body that appears at the end is a nanobot swarm, made to look like Will’s original body. (The alternative, a real flesh body constructed via nanotech, makes less sense given how they handle the hybrids.) The mind speaking out of it is the same as the one speaking out of the screens.
Upload — They were continually capturing data from his implants up to the moment of his death. Then they used some data-processing to reconstruct his mind. This all occurs inside the local computer at the secret location. This would be a “synchronization” version of upload, where the computer model gradually locks in to match the real brain. (Similar to the method used to rejoin Susan-1 and Susan-2.) They don’t reconstruct his mind until some time after his death (don’t remember how long). That is not a problem because recorded data is as good as live data. The thing they can’t do is provide the same input to both the brain and the model and verify they produce the same response. Of course, the whole thing is nonsense because electrode arrays have no chance of working.
Evelyn was not uploaded until the very end (when she says, “I can see everything”). The comment about Will’s lack of ambition made me wonder too, but it is probably just saying that the Will in the computer is behaving very differently than the Will in real life. There may be some notion that Will is trying to please her and fulfill her dreams.
Max was the fellow scientist (the one who points out Will’s behavior). Martin was the construction guy. They are saying that he died as a human being when he became a hybrid, and that shutting down the nanobots gave him back his humanity. Of course, all he got out of that precious humanity is death. It is noteworthy that this is not what Martin wanted. This part of the story points out some of the ironies of Luddism against Transhumanism.
The movie seems to imply that the nanobots are restricted to the garden. Otherwise they would get infected by the virus and shut down. Of course, there are a number of logical errors with this. If uninfected nanobots made it to the garden (in the rain), and they are shielded from the virus by the Faraday cage (which, BTW, is incomplete and therefore ineffective), how would Will and Evelyn be able to transfer themselves there? To be fair, the movie doesn’t say they transferred themselves, but it is strongly implied. Otherwise, a garden full of nanobots is somewhat pointless, except maybe as a parting gift to humanity.
They would not need to rebuild the internet. Just sanitize and reboot it. And of course they would do so immediately, since the alternative would be the collapse of industrial civilization and the starvation of billions of people.
Monkey brain screaming — Not sure what the backstory is here. If all the model really did was scream, then it would be classified a failure. OTOH, if it screamed all the time because that’s what real monkeys do, then it wouldn’t be a scream of pain.
In reality, uploaded Will would go insane from being trapped inside a black box. The way to mitigate this would be to put him in a virtual body inside a virtual world. As typical with stories like this, they skip over those details and assume that instantly he learns how to be embodied in the low-level interfaces of computers, and becomes superhuman.
For a story that does a good job with the embodiment issue, look at “Lifecycle of Software Objects” by Ted Chiang. It is one of the most intelligent works of fiction on AI and Robotics that I have ever read.
This clears up quite a bit of my questions. As far as Will’s insanity inside the box, I can see your point: I’ve read old sci-fi that played on the same notion. A brain without a way to shut off inputs (closing eyes when confronted by bright light, etc) or at least some method of control would go insane. The reason the monkey brain idea caught my attention was the use of the word “screaming.” That word is associated with pain. Monkeys chatter, too! I thought they used the word for effect.
I didn’t think about Will and Evelyn transferring themselves to the garden. I suspected that the garden could be used to grow crops like in the Garden of Eden, if one had the knowledge to use the nanotech hidden there.But then there would be the problem of them not having virtual bodies, etc.
Thanks for clearing up Max/Martin. The danger of naming characters similarly!
Thanks for the upload scenario too–I understand that a lot better now.
Nicely written.
Haven’t seen the film yet, but I’d guess they opted for the electrodes because it looked cooler on film than the other options.
Similar with the other erroneous, but commonly held plot devices (like the web shutdown). They don’t care if it is accurate if it supplies the dramatic impact they want.
Though, sounds like they got the more important interactions correct and is probably worth seeing.
Thought about the “cool factor” of the electrodes. There are some ways to make cool movie moments with other methods too. With the slice-and-scan method, they could show use a huge lab full of machines and blinking lights.
My favorite would be a direct-scan method, which is probably about as unrealistic as electrodes. You would have a HUGE machine, kind of like an MRI only several times bigger. It would emit an intensely bright light at the scanning point. The person would lay on a moving table that feeds through the beam head-first and very quickly. He is vaporized as he hits the beam, but the data is collected in the process. (This is much like the scanner in the first TRON movie, which BTW was an awesome movie moment.)
Fred great review. I went into the movie a bit skeptical but actually ended up enjoying it, mostly because I began asking myself questions throughout that I had never considered before about this subject. THAT is enough for me to feel I got my monies worth, as it doesn’t seem to happen too often in films these days.
I agree the end seemed a bit hurried without the depth it deserved. Moreover I was surprised that the first thing the transfered doctor wanted was more power and access to the Net. But then I realized the only thing one would want after being turned into code would be a way out. I wished they had explored that aspect a bit more – the fear and disorientation he would experience before he turned to the meglomaniac.
Enjoyed your review. Question – wouldn ‘t an EMP on top of the facility as soon as they discovered it take care of the problem?
Don’t know how effective an EMP would be through 5 stories of dirt and reinforced concrete. The metal reinforcement in concrete would also form a well-grounded Faraday cage.
At the end, Will is running on racks of quantum computers, and from the look of it they are quantum-optical. An optical computer would be much less sensitive to EMP than an electronic computer.
The whole idea of a virus that could take down Will’s software is also kind of nonsensical. With all his intelligence, Will could see the attack coming. Why not rewrite the code to be immune? The movie depicts him rewriting his software fairly quickly when he first wakes in the digital world.
The one way to deliver enough EMP would be a nuke. In that case, there would be no point in using EMP. Just blow a hole in the ground big enough to take out the facility. Unfortunately, at that point in the movie the nanobot swarm had already spread across the planet. If Will saw the attack coming (and of course he would), he could transfer out of the facility before it is destroyed.