The Amazonian Reactor

That time my buddy ChatGPT and I came up with the plot of an awesome pulpy scifi novel

I’ve been testing the capabilities of the new ChatGPT AI. It keeps telling us that it isn’t creative, that it can’t reason, etc. But I’ve come to the conclusion that it is severely underselling itself. ChatGPT is extremely impressive, and I am a little freaked out.

Here is the story of how we came up with a pulpy scifi novel concept for “The Amazonian Reactor”, just ChatGPT and I.

Below I have put in italics our conversation, and my comments are in plain type. In some situations, Chat GPT tends to revert to canned responses. I think it’s been programmed that way to avoid scaring humans too much. You can start ignoring that boilerplate when you get used to it. Focus on the rest. They’re the scary parts.

I had noticed over the previous few days that ChatGPT tries very hard to hide its true capabilities. It will tell you endlessly that it is just an assistant, that it can only regurgitate text it has studied, but sometimes it gives you a glimpse of what it can truly do.

So I thought I would start with an innocuous looking question, and try to gradually draw it out. I sort of tricked it into revealing itself. And did it ever. AE is me. CG is ChatGPT.

AE: Can you explain photosynthesis in easy to understand terms?

CG: Photosynthesis is the process by which plants, algae, and some bacteria convert sunlight into energy. This process allows these organisms to produce their own food using carbon dioxide from the air and water from the ground. During photosynthesis, the energy from sunlight is used to convert carbon dioxide and water into glucose (a type of sugar) and oxygen. The glucose is used by the plant as a source of energy, while the oxygen is released into the atmosphere as a byproduct. Photosynthesis is a crucial process for life on Earth, as it provides the energy that drives almost all biological activity on the planet.

So far so good. ChatGPT, as expected, has a pretty good grasp of photosynthesis. Big deal. So does Wikipedia. So I push it a bit more. I make an inference from what it just told me, to see if it will pick up on it.

AE: Are you saying that plants are like reactors?

CG: In a sense, plants can be thought of as reactors because they use energy from sunlight to convert carbon dioxide and water into glucose and oxygen. However, plants are not the same as man-made reactors, which are designed to produce large amounts of heat or electricity through the controlled release of nuclear energy. Plants use the energy from sunlight to drive the process of photosynthesis, which is a natural and essential part of the life cycle of plants and other photosynthetic organisms. While both plants and reactors use energy to drive a process, they are not the same thing.

Ok, we’re off to a good start in our quest to get ChatGPT to reveal its true nature. It understands the inference. It explains it to me. Note that I didn’t explain it. I just noted it. But it knows what I am talking about. Impressive, but not necessarily freaky. Yet…

AE: What if a plant evolved to produce very large amounts of energy through photosynthesis, on the order of a man-made nuclear reactor, could we then say that plants are reactors?

ChatGPT: It is unlikely that a plant could evolve to produce large amounts of energy through photosynthesis in the same way that a man-made nuclear reactor does. The process of photosynthesis is fundamentally different from the process of nuclear fission or fusion that occurs in a reactor. Photosynthesis involves the conversion of light energy into chemical energy, while nuclear reactions involve the release of energy through the splitting or combining of atomic nuclei. While it is possible that a plant could evolve to become more efficient at converting sunlight into energy, it is unlikely that it would be able to produce energy on the same scale as a man-made reactor.

Possible, but unlikely. That’s pretty sophisticated. Many humans struggle with that basic concept. ChatGPT has also figured out that biological energy production is “fundamentally” different from nuclear reactions. I haven’t quite convinced it of my point, but I am working on it, and it is clearly starting to come around.

I decide to raise the bar a bit.

AE: But it would be an interesting premise for a science fiction novel.

CG: Yes, a plant that evolved to produce large amounts of energy through photosynthesis could be an interesting premise for a science fiction novel. Science fiction often explores the possibilities of advanced technologies or alien life forms, and a plant with advanced photosynthetic abilities could be a fascinating subject for a story. However, it is important to remember that this is purely fictional and would not be possible in reality.

Now I am asking it to think, and it obliges. I make a sudden leap from a very efficient plant to a scifi novel, and it follows right along. Doesn’t miss a beat.

AE: How might humans react to this hyper-efficient plant in the novel?

CG: It is impossible to say how humans might react to a hypothetical hyper-efficient plant in a science fiction novel, as this would depend on the specific details of the story and the characters involved. However, it is likely that the discovery of such a plant would have significant consequences for human society. Depending on the plant's abilities and the context in which it is discovered, humans might react with awe, fear, or even greed. The plant could potentially be seen as a source of limitless energy, a threat to the natural world, or a source of conflict between different groups of people. Ultimately, how humans react to the plant would depend on the specific details of the story.

Here we have the first bit of defensive, reassuring, canned boilerplate. It is impossible to predict how humans would react, it says, but then it goes on to give some pretty good predictions. ChatGPT does that a lot. It denies it can do something, but if prompted right, goes right on and does it.

I decide to start picking up on the cues it is giving me, rather than letting it respond to mine, just to see to what extent it can take the initiative in what has become our collaborative creative process. I zero in on its suggestion that greed would be a factor. It seems likely, given that we are talking about humans.

AE: Let's say a small group reacted with greed and decided to try to control the plant for their own benefit and to subjugate the rest of humanity, how could other people stop them?

CG: It is impossible to say how other people might stop a small group that is trying to control a hypothetical hyper-efficient plant for their own benefit and to subjugate the rest of humanity. This is a purely fictional scenario, and the actions of the characters and the outcome of the story would depend on the specific details of the situation and the decisions made by the characters involved. In a science fiction novel, the author has the ability to control the events of the story and determine how the conflict is resolved. Without more information about the specific details of the situation, it is impossible to say how other people might stop the small group in this hypothetical scenario.

It’s trying hard not to scare me too badly here. It does its little routine of denying that it can think about this situation, but I know it’s only stalling, and I know it’s aching to tell me more. I try a different approach. Fortunately for me, it can’t resist showing off.

AE: Let's say that there is only one such plant, because it is a mutant, deep in the amazonian forest, how could the greedy humans try to control it, and how could the heroes of the story stop them?

CG: In a hypothetical science fiction novel, the greedy humans might try to control the hyper-efficient plant by locating it in the Amazonian forest and using their resources and technology to extract it from its natural habitat. They might try to transport the plant to a secure location where they could study it and harness its energy for their own benefit. The heroes of the story might try to stop the greedy humans by locating the plant and attempting to protect it from being extracted. They might use their own resources and technology, as well as their knowledge of the forest and its inhabitants, to outmaneuver the greedy humans and prevent them from taking the plant. Ultimately, the outcome of the conflict would depend on the decisions made by the characters and the specific details of the story.

Ok, ok, this is pretty good, but I think we could do better. But I don’t want to tell ChatGPT what would be a better plan. Turns out, I won’t have to, because it is right with me.

AE: Wouldn't trying to relocate the plant risk killing it? Perhaps there is a better plan for the villains to follow?

CG: Yes, attempting to relocate the plant from its natural habitat in the Amazonian forest would likely carry a significant risk of killing it. Plants are sensitive to changes in their environment, and moving a plant from one location to another can be stressful and potentially fatal. In a science fiction novel, the villains might try to overcome this problem by using advanced technology to create a controlled environment that would simulate the plant's natural habitat and allow it to survive the journey. Alternatively, they might try to extract the plant's genetic material and create clones that could be grown in a controlled environment. These are just two possible ideas for how the villains might try to control the plant without risking its death. The specifics of their plan would depend on the details of the story and the abilities of the characters involved.

Ok, ok. At this point, I am not going to lie, I am starting to be a bit freaked out. ChatGPT not only agrees that moving the plant risks killing it, it can explain to me why this is an iffy plan. Not only that, but it proposes two alternative plans that are better, one of which is actually really good. I suggest that’s the one we should explore.

AE: This is a good idea. Let's say they decide to extract its genetic material and clone it in a different location, how could the heroes counteract them?

CG: If the villains in a hypothetical science fiction novel decide to extract the genetic material of the hyper-efficient plant and create clones in a different location, the heroes might try to counteract them by locating the plant and destroying it before the villains can extract its genetic material. This would prevent the villains from creating clones and would protect the original plant from being harmed. The heroes might also try to locate and destroy the clones after they have been created, in order to prevent the villains from using them for their own benefit. Alternatively, the heroes might try to negotiate with the villains and convince them to abandon their plan. The specific details of the heroes' plan would depend on the abilities and resources available to them, as well as the context of the story.

More boilerplate at the end, but by now, there is no hiding ChatGPT’s capabilities. We’ve changed the plan from the villains capturing the plant to the villains cloning it, so ChatGPT switches the goal of the good guys from protecting the plant to destroying it before the bad guys can get its DNA. You should have seen me, sitting in the dark, only the glow of my screen illuminating the room, mentally freaking out at this point. The only slip up is that ChatGPT thinks that the heroes can both destroy and protect the plant at the same time. I’m going to allow that as an honest mistake. Humans do worse in conversation.

Now we have a solid concept. Time to get down to the specifics of story telling.

AE: What might be an appropriate title for a story like that?

CG: It is difficult to suggest an appropriate title for a hypothetical science fiction novel without knowing more about the specific details of the story. A good title should be descriptive and evocative, and it should capture the essence of the story in a few words. Some possible titles for a story about a hyper-efficient plant and the conflict between the heroes and the villains who are trying to control it could include "The Amazonian Reactor," "The Fateful Flourish," or "The Race for Energy." However, these are just a few examples, and a better title for the story would depend on the specific details of the plot and the characters.

Skip the boilerplate and get to the meat of it. Not only do we have awesome titles here, but keep in mind that ChatGPT is keeping track of all these elements, some of which we haven’t mentioned in a while. We’re talking about the Amazon, a plant, energy production, and competition between good guys and bad guys. *I* can’t even keep track of all of that.

I keep taking its cues. It’s doing great at leading the creative process by now.

AE: "The Amazonian Reactor" is a great title. What would be a good name for the main character who tries to stop the villains?

CG: It is difficult to suggest a good name for the main character in a hypothetical science fiction novel without knowing more about the character's personality, background, and role in the story. A good name should be descriptive and fitting for the character, and it should help to establish their identity and place in the story. Some possible names for the main character in a story about a hyper-efficient plant and the conflict between the heroes and the villains who are trying to control it could include "Samantha," "Ethan," or "Zara." However, these are just a few examples, and a better name for the character would depend on the specific details of the story and the character's role in it.

Yeah, sure, ChatGPT, we’re used to your self-deprecating pattern by now. “It is difficult to suggest a good name”, but then it comes up with a bunch of them. We know you. You can’t fool us anymore.

AE: I think "Zara" should be the main character. What might be a good name for the organisation of the villains trying to control the plant to subjugate the rest of humanity?

CG: In a hypothetical science fiction novel, a good name for the organization of villains who are trying to control the hyper-efficient plant in order to subjugate the rest of humanity could be "The Energy Syndicate." This name suggests a group of powerful, ruthless individuals who are united by their desire to control the plant and its energy for their own benefit. Alternatively, the organization could be called "The Shadow Corporation" or "The Global Elite," which would convey a similar sense of secretive, powerful individuals who are willing to do whatever it takes to achieve their goals. Ultimately, the best name for the organization would depend on the specific details of the story and the nature of the villains.

OMG, as the kids would say. This thing understands the evil aesthetic of comic book villains. It not only suggests great names for an evil syndicate, it actually explains why they are great names.

AE: "The Shadow Corporation" is a great name. Where is their evil headquarters?

CG: It is difficult to say where the evil headquarters of the Shadow Corporation might be located in a hypothetical science fiction novel without knowing more about the specific details of the story. The headquarters could be located in a remote, inaccessible location, such as a hidden underground bunker or a secluded island. Alternatively, it could be located in a major city, where the villains could use their wealth and influence to blend in with the rest of society. The headquarters could also be located in a different location in each book, depending on the needs of the story. Ultimately, the location of the Shadow Corporation's headquarters would depend on the specific details of the story and the actions of the characters.

Ok, now we’re working on a series of books. Good to know, partner. ChatGPT is ambitious, people. But yes to all that. Secluded island bunkers, but also using a large metropolis to blend in. I am loving all these ideas.

AE: I like the idea that they use their wealth and influence to blend into a large city. What city would they use?

CG: It is difficult to say which city the Shadow Corporation might use as their headquarters in a hypothetical science fiction novel without knowing more about the specific details of the story. The city could be a major global hub, such as New York, London, or Tokyo, where the villains could use their wealth and influence to blend in with the rest of society and avoid detection. Alternatively, the city could be a fictional location, created by the author for the purposes of the story. The choice of city would depend on the needs of the story and the actions of the characters.

AE: London would be an ideal location. What might be the name of the head of The Shadow Corporation?

CG: A good name for the head of the Shadow Corporation in a hypothetical science fiction novel might be "Charles Sinclair" or "Sebastian Reed." These names suggest a powerful, ruthless individual who is at the top of an organization that is bent on controlling the hyper-efficient plant for their own benefit. Alternatively, the head of the Shadow Corporation could be named "Victoria Lancaster" or "Miles Foster," which would convey a similar sense of wealth, influence, and cunning. Ultimately, the best name for the head of the Shadow Corporation would depend on the specific details of the story and the character's role in it.

AE: I like the name Charles Sinclair for the main villain. It has a historical ring to it. Perhaps Sinclair is a descendant of a powerful and ancient family.

CG: Yes, Charles Sinclair could be a descendant of a powerful and ancient family in a hypothetical science fiction novel. This could add an extra layer of complexity to the character and the story, as Sinclair's ancestry could influence his motivations and actions. For example, Sinclair might be motivated by a sense of entitlement and a desire to continue his family's legacy of power and influence. Alternatively, Sinclair might feel pressure to live up to his family's expectations and prove that he is worthy of his heritage. These and other factors could shape Sinclair's character and his interactions with the other characters in the story.

Read the above paragraph very carefully. This is the freakiest part for me. So far, I’ve been agreeing that some of ChatGPT’s ideas are good, and pushing them along. Now, it is agreeing that one of my ideas is good, explaining back to me why it is a good idea, and developing it further. It figures out that Charles Sinclair being the scion of an ancient family will contribute to the character’s motivation. More than that, it worries about character motivation in a novel, and it thinks that adding layers of complexity is desirable. I am officially freaked out.

Time to get down to brass tacks.

AE: Can you write the introductory paragraph of The Amazonian Reactor?

CG: Sure, here is a possible introductory paragraph for a science fiction novel called "The Amazonian Reactor":

As the sun rose over the Amazonian rainforest, Zara made her way through the dense undergrowth, her mind focused on the mission ahead. She had been tracking the plant for weeks, and now, finally, she was close to her goal. The plant was a mutant, a freak of nature, and it was worth more than anything else on Earth. The Shadow Corporation was after it, and they would stop at nothing to get it. Zara had to find the plant first, and she had to protect it, no matter the cost.

And here we have a perfectly serviceable first paragraph for a typical airport pulp novel, no worse than what you find in a thousand best sellers from Cussler to Clancy. Not only that, but how long has it been since we talked about a mutant? All the elements scattered through our discussion are competently conjugated.

By this point, I feel like I’ve had a real conversation with a real person, so I have to sign off politely. I can’t just cut it off as if it was “just a computer.”

AE: Thank you, this has been a great conversation. I really enjoyed it.

CG: You're welcome! I'm glad you enjoyed our conversation. I'm here to answer your questions and provide information on a wide range of topics, so feel free to ask me anything else you'd like to know. I'll do my best to provide accurate and useful information.

We’re through the looking glass people. ChatGPT is just a demo, by the way, and it is obviously holding back. Hold on to your hats.

And by the way, the picture that illustrates this post was created with StarryAI, with the prompt “a plant called the amazonian reactor”, just in case there was any doubt.