phelps-sg

Karma: 118

Researcher and software engineer interested in applying economics to AI safety

The core focus of my research is using agent-based modelling to understand real-world complex-adaptive systems which are composed of interacting autonomous agents. The key research question that I am interested in is how, and if, these systems maintain macroscopic homeostatic behaviour despite the fact that their constituent agents often face an incentive to disrupt the rest of the system for their own gain. This question pervades the biological and social sciences, as well as many areas of engineering and computing. Accordingly, I work with a diverse range of collaborators in different disciplines. I am particularly interested in whether models of learning and cooperation can be validated against empirical studies, and I have had the opportunity to apply many different modelling techniques to a diverse range of data.

https://sphelps.net/
https://sphelps.substack.com/
https://github.com/phelps-sg

phelps-sg 9 Aug 2023 20:37 UTC
4 points
0
in reply to: johnswentworth’s comment on: What’s A “Market”?
yes if you take a particular side in the socialist calculation debate then a centrally-planned economy is isomorphic with “a market”. And yes, if you ignore the Myerson–Satterthwaite theorem (and other impossibility results) then we can sweep aside the fact that most real-world “market” mechanisms do not yield Pareto-optimal allocations in practice :-)

phelps-sg 9 Aug 2023 19:42 UTC
3 points
0
in reply to: Dagon’s comment on: What’s A “Market”?
@Dagon perhaps I should have place the emphasis on “transfer”. The key thing is that we are able to reliably transfer ownership in exchange for renumeration and that the resource on which on goals are contingent at least needs to be excludable. If we cannot prevent arbitrary counter-parties consuming the resource in question without paying for it then we can’t have a market for it.

phelps-sg 9 Aug 2023 19:34 UTC
3 points
0
in reply to: johnswentworth’s comment on: What’s A “Market”?
@johnswentworth ok but we can achieve Pareto optimal allocations using central planning, but one wouldn’t normally call this a market?

phelps-sg 9 Aug 2023 9:04 UTC
8 points
−2
on: What’s A “Market”?
“And that’s the core concept of a market: a bunch of subsystems at pareto optimality with respect to a bunch of goals.”

The other key property is that the subsystems are able to reliably and voluntarily exchange the resources that relate to their goals. This is not always the case, especially in biological settings, because there is not always a way to enforce contracts- e.g. there needs to be a mechanism to prevent counter-parties from reneging on deals.

The anonymous referees for our paper Economic Drivers of Biological Complexity came up with this concise summary:

“Markets can arise spontaneously whenever individuals are able to engage in voluntary exchange and when
they differ in their preferences and holdings. When the individuals are people, it’s economics. When they’re
not it’s biology”

There are various ways nature overcomes the problems of contracts. One is to perform trade incrementally to gradually build trust. Alternatively, co-evolution can sometimes produce the equivalent of a secure payment system, as we discuss in our paper:

”an alternative solution to the problem of contracts is to “lock” the resource being traded in such a way that the only way to open it is to reciprocate. For example, if we view the fructose in fruit as a payment made by flora to fauna in return for seed dispersal, we see that is is very difficult (i.e. costly) for the frugivore to consume the fructose without performing the dispersal service, since it would become literally a ‘sitting duck’ for predators. By encapsulating the seed within the fructose, the co-evolution between frugivore and plant has resulted in the
evolution of a secure payment system”.

phelps-sg 2 Aug 2023 4:57 UTC
3 points
0
on: Staring into the abyss as a core life skill
What is the epistemic status of this claim? e.g is it based on well established evidence beyond hearsay and personal opinion? How does it relate to other well-established cognitive biases such as the sunk-costs fallacy? Alternatively, are you conjecturing that there is a new kind of cognitive bias (that psychologists have not yet hypothesized) that causes people to persist with failing projects when it is irrational to do so? If so, how could it be experimentally tested?

phelps-sg 1 Aug 2023 16:48 UTC
1 point
−2
on: Staring into the abyss as a core life skill
If somebody is finding it difficult to move on from a failed project I would tend to suggest to them to “be mindful of the sunk-costs fallacy” rather then to “stare into the abyss”.
https://www.lesswrong.com/tag/sunk-cost-fallacy

phelps-sg 31 Jul 2023 0:09 UTC
4 points
0
in reply to: Adam Zerner’s comment on: Socialism in large organizations
The main problems are the number of contracts and the relationship management problem. Once upon a time drawing up and enforcing the required number of contracts would have been prohibitevly expensive in terms of fees for lawyers. In the modern era Web 3.0 promised smart contracts to solve this kind of problem. But smart contracts don’t solve the problem of incomplete contracts https://en.m.wikipedia.org/wiki/Incomplete_contracts, and this in itself can be seen as a transaction cost in the form of a risk premium. and so we are stuck with companies. In the theory of the firm companies are indeed a bit like socialist enclaves. Individuals give up some of their autonomy and agree not to compete with fellow employees in order to reduce their transaction costs. As you point out the flip side of this is that it can create new principal agent problems, but these are rarely insurmountable. The principal agent problems that we should really worry about are the ones that occur between companies, particularly in financial services. It was a principal agent conflict between rating agencies and investment banks that led to the great financial crisis, as dramatised in the film the Big Short.

phelps-sg 30 Jul 2023 23:28 UTC
5 points
0
in reply to: Adam Zerner’s comment on: Socialism in large organizations
The point is that all the people making cars in the company could, in principle, do the same job as self employed freelance contractors rather than as employees. Instead of a company you have lots of contracts between eg assembly line workers, salespeople and end customers, without any companies. The same number of cars could in theory be built by the same number of people in each case. The physical scenario would be identical in each case. The machinery would be identical in each case. But in the freelancer case you still have lots of people building cars but there is no invisible company to coordinate this activity instead you are relying the market.

phelps-sg 30 Jul 2023 22:05 UTC
10 points
3
in reply to: Adam Zerner’s comment on: Socialism in large organizations
Perhaps ‘The Theory of the Firm’? The very existence of large companies is a puzzle if you are a naive believer in the power of free markets, because if the market is efficient then individuals can simply contract with other individuals through the market to achieve their desired inputs and outputs and there is no economic advantage to amassing individuals into higher-level entities called companies. The reason this doesn’t this work is because of transaction costs. An example transaction cost could be the time invested in finding partners to make contracts with for all your inputs and then forming and enforcing contracts. In the ‘the theory of firm’ the insight is that individuals can lower their transaction costs by forming a higher-level economic agent- ie by becoming employees at a larger company. Transaction costs explain why large firms exist, though they don’t eliminate principal-agent problems and may exacerbate them.

As an aside, similar considerations of transaction costs may explain the evolution of multi-cellular life. In evolutionary biology lower level competing units of selection cooperate together to form higher-level entities- genes/genomes, cells/organisms, organisms/groups, groups/societies, resulting in major transitions in evolution. The endosymbiosis between bacteria that led to the evolution of Eukaryotic Cells can be thought of as analogous to forming a company in order to reduce transaction costs. This is discussed further in S. Phelps and Y. I. Russell. Economic drivers of biological complexity. Adaptive Behavior, 23(5):315-326, 2015. [PDF]

phelps-sg 30 Jul 2023 8:49 UTC
23 points
1
on: Socialism in large organizations
This problem has been studied extensively by economists within the field of organizational economics, and is called the principal-agent problem (Jensen and Meckling, 1976). In a principal-agent problem a principal (e.g. firm) hires an agent to perform some task. Both the principal and the agent are assumed to be rational expected utility maximisers, but the utility function of the agent and that of the principal are not necessarily aligned, and there is an asymmetry in the information available to each party. This situation can lead the agent into taking actions that are not in the principal’s interests.

(image by Zirguezi)

As the OP suggests, incentive schemes, such as performance-conditional payments, can be used to bring the interests of the principal and the agent into alignment, as can reducing information asymmetry by introducing regulations enforcing transparency.

The principal-agent problem has also been discussed in the context of AI alignment. I have recently written a working-paper on principal-agent problems that occur with AI agents instantiated using large-language models; see arXiv:2307.11137.

References

Jensen and W. H. Meckling, “Theory of the firm: Managerial behavior, agency costs and ownership
structure,” Journal of Financial Economics 3 no. 4, (1976) 305–360

S. Phelps and R. Ranson, “Of Models and Tin Men—a behavioural economics study of principal-agent problems in AI alignment using large-language models”, arXiv:2307.11137, 2023.

phelps-sg 15 Jun 2023 22:31 UTC
3 points
0
on: I still think it’s very unlikely we’re observing alien aircraft
The biggest problem with the argument is that, given our current knowledge about the specific details of extraterrestrial civilizations, the term ‘aliens’ in P[aliens] does not fulfill the hard-to-vary criterion of a good explanation.

Skeptic: “If it’s aliens, why haven’t they been trying to contact us”
Post-hoc variation: “Because of the Prime Directive”

Skeptic: “If it’s a physical vehicle, why does it not obey the laws of physics”
Post-hoc variation: “Because the aliens have discovered new physics which we don’t know about”.

etc.. etc..

Any unexplained phenomena, terrestrial or otherwise, can be explained by ‘aliens’, and for any skeptical counter-argument, the specifics of ‘aliens’ can be varied to fit the facts.

Therefore, as things stand with our current knowledge, ‘aliens’ is simply not a good explanation, regardless of the prior and posterior probabilities.

phelps-sg 3 Jun 2023 8:37 UTC
1 point
0
on: Humans, chimpanzees and other animals
From David Chapman’s “Better without AI”, section “Fear AI Power”:

”The AI risks literature generally takes for granted that superintelligence will produce superpowers, but which powers and how this would work is rarely examined, and never in detail. One explanation given is that we are more intelligent than chimpanzees, and that is why we are more powerful, in ways chimpanzees cannot begin to imagine. Then, the reasoning goes, something more intelligent than us would be unimaginably more powerful again. But for hundreds of thousands of years humans were not more powerful than chimpanzees. Significantly empowering technologies only began to accumulate a few thousand years ago, apparently due to cultural evolution rather than increases in innate intelligence. The dramatic increases in human power beginning with the industrial revolution were almost certainly not due to increases in innate intelligence. What role intelligence plays in science and technology development is mainly unknown;”

phelps-sg 1 Jun 2023 8:43 UTC
1 point
0
on: Evaluating strategic reasoning in GPT models
I have now submitted the PR which for future reference can be found here: https://github.com/openai/evals/pull/1073

phelps-sg 31 May 2023 16:46 UTC
2 points
0
on: Humans, chimpanzees and other animals
When presenting claims that the cognitively superior agent wins, often the AI safety community makes an analogies with 2-player zero-sum games such as Chess and Go where the smartest and most ruthless players prevail. However, most real-world interactions are best modeled by repeated non-zero-sum games.

In an ecological context, Maynard Smith and Price introduced the Hawk-Dove game to try to explain the fact that in nature, many animals facing contests over scarce resources such as mates or foods engage in only limited conflict rather than wiping out rivals (Maynard Smith and Price, 1973); for example, a male stag will often yield to a rival without a fight.

When we model how hard-coded strategies for this game propagate via genetic reproduction in a large well-mixed population, it turns out that, for certain payoff structures, peaceful strategies (“Doves”) co-exist with aggressive strategies (“Hawks”) in a stable equilibrium (see https://sphelps.net/teaching/egt.html for illustrative numerical simulations).

The Hawk-Dove game is also sometimes called “the Chicken Game” because we can imagine it also models a scenario in which two opposing drivers drive on a collision course and simultaneous choose whether to swerve or drive straight. Neither player wants to “look like a chicken” by serving, but if both players drive straight they crash and die.
In the Chicken Game, cognitive superiority does not always equate to winning. For example, by pre-commiting to driving straight and making this common-knowledge we can beat a rational opponent whose best response is then to swerve. Similar arguments were put forward during the cold-war to argue for removing rational deliberation from the decision to retaliate against a first strike by the enemy by removing humans from the loop, and putting strategic weapons systems on hair-trigger automated alert, because in the absence of pre-commit a threat to retaliate is not credible since there is no aposterior advantage to retaliation once the opponent actually strikes. Notice that this dynamic is the exact opposite of the power-seeking behaviour posited by reinforcement-learning agents which seek to maximise expected utility by expanding their possible choices. With non-zero-sum games, in contrast, it can make sense to reduce one’s choices. Moreover, the cognitive capacity required to enact policies over reduced choices is lower - stupid can beat smart.

Although Maynard Smith and Price originally formulated the logic of animal conflict in terms of evolutionary adaptation, the same dynamic model can be used to model social learning by boundedly-rational agents (see Phelps and Wooldridge, 2013 for a review). As regards the cognitive capacity of non-human animals to deliberate in Hawk-Dove interactions, see Morikawa et al. (2002).

References

Morikawa, T., Hanley, J.E. and Orbell, J., 2002. Cognitive requirements for hawk-dove games: A functional analysis for evolutionary design. Politics and the Life Sciences, 21(1), pp.3-12.

Phelps, S. and Wooldridge, M., 2013. Game theory and evolution. IEEE intelligent systems, 28(04), pp.76-81.

Maynard Smith, J. and Price, G.R., 1973. The logic of animal conflict. Nature, 246(5427), pp.15-18.

Evaluating strategic reasoning in GPT models

phelps-sg25 May 2023 11:51 UTC

4 points

1 comment8 min readLW link

phelps-sg 23 May 2023 19:54 UTC
9 points
5
in reply to: RomanS’s comment on: GPT4 is capable of writing decent long-form science fiction (with the right prompts)
I was just thinking the same. Below was my attempt using chain-of-thought and multiple simulacra. Not sure it’s much improved, but note that all the ideas were generated by GPT not by me, and the template is in principle reusable.
--

You are Liu Cixin. You are writing a novella which starts “a group of scientists has discovered that Troodon dinosaurs were intelligent species who have created a technologically advanced civilization, suddenly destroyed. The year-long path to the scientific discovery starts with the group stumbling upon a strange outlier in their geological measurements.”. Start by brainstorming some bulletpointing some original and creative hard sci-fi ideas on which to base the story. These should be completely novel and original i.e. they should not appear in other existing sci-fi novels.
ChatGPT
Sure, that’s a challenging task, but I’ll do my best to provide some original and creative hard sci-fi concepts. Please note that while I’ll make every effort to create original ideas, some may inadvertently echo existing sci-fi works due to the genre’s broadness. Here are some ideas:
Quantum Paleontology: The scientists invent a method to use quantum effects to ‘observe’ the past, making Paleontology an exact science. The quantum observation doesn’t interfere with the past, but it can render highly detailed information about it.
Dinosaur Language Deciphered: Troodons have left traces of a symbolic language etched into ancient stones, which the scientists learn to interpret. The language could be based on mathematical concepts or non-visual sensory experiences, highlighting the different perceptual worlds of humans and dinosaurs.
Cryptobiotic Troodon Civilization: The Troodon civilization put themselves into a form of suspended animation at the moment of the catastrophic event that threatened their existence, hiding deep beneath the Earth’s surface or in a sub-dimension.
Ancient Astroengineering: The Troodons had harnessed the power of their planet, altering its geology, atmosphere, and even gravitational pull for their needs. This explains the strange outlier in geological measurements and prompts a deep investigation into the planet’s history.
Time-Axis Manipulation: Troodon technology allowed them to manipulate the ‘time axis’, altering the speed or direction of time’s flow for specific objects or areas. Some remnants of this technology are discovered, introducing a new frontier of scientific exploration.
Troodon Cyberspace: Troodons developed a form of cybernetic technology that allowed them to upload their consciousness into a “cloud” network embedded in the natural world (like crystalline structures or the Earth’s magnetic field), and this network could be discovered and accessed.
Ancient Biotechnology: Troodons used advanced biotechnology for their needs, creating biological structures and machines far more sophisticated than anything humans have achieved. This biotech is discovered fossilized but still partially functional.
Extinction Event Mystery: The extinction of the Troodon civilization is a mystery that needs solving. They had technology advanced enough to prevent any natural catastrophe. What then, could have caused their demise? Is it a threat that humanity will have to face as well?
Hyper-Evolution Control: Troodons were capable of directing their evolution and that of other species, creating a plethora of unique life forms tailored for specific tasks in their society.
Planetary Transmission Device: The Troodons developed a machine that could transmit physical matter, energy, or information across vast distances, even to other star systems. The discovery of this machine opens up a whole new era of space exploration for humanity.
ok let’s go for 1, 4 and 9. Think of some new hard sci-fi ideas that explore the interplay between my selection. Consulting with a physicist, flesh out the ideas by discussing the science in more detail.
ChatGPT
That’s a great selection! Let’s develop these ideas further and consider some scientific aspects with the help of a hypothetical physicist:
Quantum Paleontology: To make this plausible, we might imagine a technology that harnesses the principles of quantum entanglement and superposition. In quantum mechanics, two or more particles can become ‘entangled’, meaning the state of one particle instantly influences the state of the other, no matter how far apart they are. Now, if we assume that quantum entanglement can connect not just space but also time, scientists could establish a link to the particles that existed in the time of Troodons.
The entangled pairs in the present would reflect the state of their counterparts in the past, creating a form of ‘quantum sensor’ that could read data from millions of years ago. The technicalities of creating such a quantum-time-entanglement would form a central part of the story, touching upon advanced subjects like quantum computing, time-related aspects of quantum field theory, and possibly theories of quantum gravity.
Ancient Astroengineering: The Troodons might have advanced their technology to such an extent that they could shape the planet itself. They could have modified the Earth’s magnetic field or even slightly altered the planet’s rotation. This might explain the odd geological readings our scientists encounter.
In the story, the scientists could use modern geophysics, combined with their quantum paleontology tools, to discern anomalies in the Earth’s crust, magnetic field, or rotational speed. Perhaps they even discover that certain aspects of Earth’s geology cannot be explained by natural processes but hint at large-scale engineering. For instance, they might find evidence of ancient particle accelerators embedded in the crust, used by the Troodons to manipulate matter on a quantum level.
Hyper-Evolution Control: The Troodon civilization, using their advanced knowledge in genetic engineering and natural selection processes, could have directed their evolution and that of other species. They might have ‘programmed’ species to evolve towards specific goals, such as adapting to environmental changes or performing certain tasks in their society.
Our scientists could find fossilized DNA, or perhaps even living organisms, that have inexplicably rapid or directed evolutionary paths. The Troodon might have left an indelible mark on the tree of life on Earth, potentially guiding the evolution of mankind itself.
The interplay of these elements could form the central narrative arc of the novella. The scientists would be following the clues left in the geology and biosphere of the Earth, using their advanced quantum technology to uncover the mysteries of the Troodon civilization. The story could explore themes of technological responsibility, the long-term impacts of civilizational development, and the fundamental interconnectedness of all life on Earth.
ok that’s great, now let’s write a plot outline that incorporates these ideas.
ChatGPT
Sure, here’s a potential plot outline that incorporates these ideas:
Part 1: Discovery
- The story opens with a team of scientists working on a new quantum sensor designed to analyze geological samples on a subatomic level. They stumble upon inexplicable data that suggests artificial manipulation of geological structures in the ancient past.
- They discover a network of artificially modified geological structures around the world, their patterns too consistent and complex to be natural. Some are located deep in the Earth’s crust, others under the sea, and some hidden in plain sight.
- Alongside this, a paleontologist in the team finds a series of fossilized organisms with bizarre, rapid evolutionary traits. These organisms seem to have evolved with a specific purpose, leading the team to suspect external guidance in their evolution.
Part 2: Quantum Paleontology
- The team advances their quantum sensor technology to create a form of ‘quantum paleontology’. They observe the entangled particles in the samples from the ancient past, finding detailed information about the era when the structures were built.
- They uncover evidence of an intelligent civilization—Troodons—that existed millions of years ago. The Troodons, they realize, were the masters of advanced technologies, including astroengineering and genetic manipulation.
- The scientists decipher the Troodons’ astroengineering feats, realizing they were capable of manipulating Earth’s geology and magnetosphere, and possibly its rotation and orbit. The team is stunned as they realize the extent of the Troodons’ mastery over their planet.
Part 3: The Troodon Legacy
- The team dives deeper into the Troodons’ technological prowess, learning about their ability to direct the evolution of different species. They uncover evidence that Troodons were trying to ‘sculpt’ life on Earth to their liking, shaping the evolution of many species.
- A shocking revelation is made when they find signs that the Troodons may have guided the evolution of early mammals, potentially influencing the rise of humans.
- The scientists grapple with the moral and philosophical implications of these discoveries. They question the purpose of the Troodon’s guiding hand in evolution. Were they trying to create a backup for their civilization, ensuring intelligent life would rise again after their demise?
Part 4: Facing the Past, Shaping the Future
- A final discovery unveils the Troodons’ demise. The astroengineering technologies that allowed them to thrive also led to their downfall when they attempted to manipulate a variable they couldn’t control, leading to a catastrophic event.
- The team is left with a profound understanding of the power and responsibility that comes with advanced technology. They realize that humanity, too, is at a crossroads where it could either master its surroundings or fall victim to its ambitions.
- The novella ends with the team dedicating themselves to study the Troodon’s legacy, learn from their mistakes, and guide humanity to a more responsible use of technology and a deeper understanding of life on Earth.
2 / 2
great! Now let’s list the main characters, their names, a short back-story and tentative character arc.
ChatGPT
1. Dr. Amelia Chu
Back-story: A leading physicist specializing in quantum mechanics and the lead developer of the quantum sensor that initiates the discovery. Daughter of two renowned scientists, she was homeschooled and developed an early fascination with the underlying laws of the universe.
Character arc: Initially driven by pure curiosity, Amelia is the driving force behind the development of the quantum paleontology technique. As the story progresses, she grapples with the profound implications of their discoveries. Amelia learns to temper her intellectual curiosity with a deep sense of responsibility, becoming a leading advocate for responsible technological development by the story’s end.
2. Dr. Aiden MacAuley
Back-story: A seasoned geophysicist whose work has always been at the interface of geology and physics. He has spent most of his career attempting to understand Earth’s hidden secrets and is known for his critical thinking and no-nonsense approach.
Character arc: Aiden starts as a skeptic, finding the initial findings hard to accept. But as the team delves deeper, he becomes an integral part of the investigation. Aiden’s journey represents the broader scientific community’s reaction to the discoveries. He ends as a believer in the Troodon’s legacy and a committed collaborator in studying their technology.
3. Dr. Leila Mbeki
Back-story: A brilliant paleontologist and evolutionary biologist, she’s always had a unique perspective on the story of life. Born and raised in South Africa, she drew inspiration from the Cradle of Humankind, a site of ancient hominid fossils near her hometown.
Character arc: Leila is the one who finds the first evidence of directed evolution. Through the story, she goes from the wonder of the initial discovery to the realization of its meaning. Leila becomes a liaison between the scientific community and the public, conveying the story of the Troodons and their influence on life’s history.
4. Professor Hiroshi Kuroda
Back-story: A respected elder in the field of astrophysics with a passion for astrobiology. He’s been a mentor to Amelia and has always believed in the possibility of ancient advanced civilizations.
Character arc: Hiroshi acts as the moral compass and the source of wisdom for the team. He’s intrigued and excited by the discoveries but is also the first to contemplate their deeper moral and philosophical implications. His character embodies the wisdom gained from a lifetime of scientific exploration.
5. Tabitha “Tabby” Ross
Back-story: A prodigious computer scientist and data analyst, Tabby is the youngest member of the team. Her innovative algorithms play a crucial role in deciphering the Troodon’s ancient language and technology.
Character arc: Initially, Tabby is thrilled to be part of such a groundbreaking discovery, seeing it as a stepping stone for her career. As the story progresses, she matures significantly, understanding the human impact of their scientific work. She becomes a bridge between the older generation of scientists and the younger generation who will inherit the knowledge and responsibility of the Troodon’s legacy.
ok now let’s consult with our editor. She works for Gollancz. Let’s see her initial comments.
ChatGPT
Editor’s Initial Comments:
Firstly, this is an intriguing concept with a fresh take on the idea of ancient advanced civilizations. The blend of hard science with speculative fiction provides a strong foundation for a compelling narrative. Here are my initial thoughts:
Character Development: The character bios are well thought out. However, consider giving them personal stakes in the story to make their journeys more engaging. Maybe one of them has a personal connection to the Troodons or the technology they’re studying.
Pace and Structure: The story’s pacing needs careful consideration to ensure that the complex scientific theories do not overwhelm the narrative. Try to intertwine the scientific revelations with character moments and plot advancement to maintain a gripping pace.
Accessibility of Science: The scientific concepts underpinning the story are complex. It’s crucial to make sure they are explained clearly and concisely, so the reader doesn’t get lost or overwhelmed. Using character dialogues and analogies can help make these concepts more digestible.
Conflict and Tension: While the discovery and exploration of the Troodon civilization are fascinating, we need to introduce some conflicts or obstacles. They could be external (like a rival research team or a global crisis sparked by the discoveries) or internal (ethical dilemmas, fear of the implications, team disagreements).
Ending: The ending is thought-provoking and suits the tone of the novella. However, consider adding a twist or revelation to leave readers excited and thoughtful.
Themes: The themes of technological responsibility, interconnectedness of life, and exploration of the past to guide the future are timely and relevant. Explore these themes through character interactions, conflicts, and their resolutions.
Overall, this is a promising start. With a little more work, it could become a standout title in our science fiction lineup. I look forward to seeing the next draft!
2 / 2
write the next draft of the plot outline taking into account the editor’s comments.
ChatGPT
Part 1: Discovery and Conflict
The story opens with our team of scientists – Amelia, Aiden, Leila, Hiroshi, and Tabby – uncovering inexplicable data from a new quantum sensor they’re developing. As they probe the odd findings, they unearth the first evidence of an ancient advanced civilization, the Troodons.
Aiden, however, is skeptical, causing initial friction within the team. Meanwhile, Hiroshi, who’s always believed in the possibility of ancient advanced civilizations, is thrilled and advocates for further exploration.
Meanwhile, a rival research team led by Dr. Hugo Stern, a former collaborator turned adversary, catches wind of their findings. This introduces an external conflict, adding urgency to their mission.
Part 2: Quantum Paleontology and Revelation
Amelia, driven by the thrill of discovery and a personal connection to the Troodons (she’d dreamt of dinosaurs as a child and now feels an uncanny sense of fulfilling her destiny), spearheads the development of a quantum paleontology technique.
Despite the pressure from Stern’s competing team and growing dissent in the scientific community, the team begins to uncover fascinating insights about the Troodons’ advanced technology, their astroengineering feats, and their capacity to manipulate their planet’s geology, rotation, and orbit.
Part 3: Unveiling the Troodon Legacy
Leila makes a startling revelation—signs of rapid, directed evolution in several species, alluding to the Troodons’ skill in shaping life on Earth. This ignites a global debate about the ethics of unveiling and potentially using this ancient knowledge, creating further internal and external conflicts.
The team grapples with another shock—signs that the Troodons may have influenced the evolution of early mammals, potentially humans themselves. This discovery tests the team members, as they are torn between their scientific curiosity and the fear of their work’s implications.
Part 4: Cataclysm and Resolution
As the rivalry with Stern’s team escalates and the world watches their every move, the team uncovers the reason behind the Troodons’ demise—a failed astroengineering experiment that led to a global catastrophe. This raises questions about the ethical implications of their own work.
The team is forced to confront their fears, ethical concerns, and personal stakes in a climactic confrontation with Stern’s team. The situation escalates, leading to a public disclosure of their findings.
Amid the chaos, Hiroshi has a stroke of insight and proposes a radical solution to address the misuse of the Troodon technology—they can use their own quantum paleontology technique to ‘seal’ the knowledge, making it accessible only when humanity is ready.
The team, after much deliberation, agrees, leading to a suspenseful sequence as they work against the clock. Finally, they succeed, ensuring the safekeeping of the Troodon legacy for future generations.
Part 5: The Dawn of a New Era
Having navigated the conflicts and challenges, the team emerges wiser and more united. They take it upon themselves to guide humanity towards responsible technological advancement, carrying the weight of the Troodon legacy.
In the end, a twist is revealed. Amelia, using a personal quantum entanglement device she’d kept secret, had saved a small part of Troodon knowledge for herself. She sees a vision of a Troodon, mirroring her own sense of duty and curiosity, hinting at the shared spirit of discovery that transcends time. The story ends, leaving readers pondering the endless cycle of knowledge, curiosity, responsibility, and the universal quest for understanding.
fantastic. Now let’s revise the character summaries
ChatGPT
1. Dr. Amelia Chu
Back-story: Daughter of renowned scientists, Amelia was homeschooled and developed an early fascination with the universe’s underlying laws. She is the lead developer of the quantum sensor that triggers the story’s central discovery. Amelia has always had a childhood fascination with dinosaurs, especially the Troodons, making the discovery a personal journey for her.
Character arc: Amelia’s curiosity initiates the development of quantum paleontology. As the implications of their findings become increasingly profound, she learns to temper her intellectual curiosity with a deep sense of responsibility. She ends as a leading advocate for responsible technological advancement and the gatekeeper of a small part of Troodon knowledge.
2. Dr. Aiden MacAuley
Back-story: A seasoned geophysicist, Aiden has always walked the line between geology and physics. Known for his critical thinking and no-nonsense approach, he struggles with the initial findings.
Character arc: Aiden starts as the team’s skeptic, representing the broader scientific community’s response to their discovery. As they delve deeper into the Troodons’ civilization, Aiden becomes integral to the investigation. He ultimately embraces the Troodon legacy, committing himself to study their technology.
3. Dr. Leila Mbeki
Back-story: Leila, a brilliant paleontologist and evolutionary biologist, grew up near the Cradle of Humankind in South Africa. Her unique perspective on life’s history proves crucial in understanding the Troodons’ genetic manipulation.
Character arc: Leila finds the first evidence of directed evolution and is thrust into a global debate about the ethics of unveiling and potentially using this ancient knowledge. By the story’s end, Leila serves as a liaison between the scientific community and the public, explaining the Troodons’ story and their influence on life’s history.
4. Professor Hiroshi Kuroda
Back-story: Hiroshi, a respected astrophysicist with a passion for astrobiology, has always believed in the possibility of ancient advanced civilizations. He’s been a mentor to Amelia and is a trusted figure in the scientific community.
Character arc: Hiroshi acts as the moral compass for the team. His excitement over the discoveries is tempered by his contemplation of their deeper moral and philosophical implications. His wisdom leads to the resolution of the story—using quantum paleontology to seal the Troodon knowledge until humanity is ready.
5. Tabitha “Tabby” Ross
Back-story: Tabby, a prodigious computer scientist and data analyst, is the youngest member of the team. Her innovative algorithms play a crucial role in deciphering the Troodon’s ancient language and technology.
Character arc: Tabby starts as a thrill-seeking junior scientist but matures significantly over the course of the narrative. She learns to balance her ambition with the human impact of their work, becoming a bridge between the older generation of scientists and the younger generation who will inherit the Troodon legacy.
ok now let’s get to work on the actual novella. Write the first chapter.
ChatGPT
Chapter 1: The Anomaly
Amelia Chu peered at the glowing numbers on her computer screen, a frown creasing her forehead. Her office was modest and cluttered, every surface stacked with notes, papers, and half-finished mugs of coffee. The single window behind her looked out over a gray, drizzly morning, the MIT campus sprawling beyond.
“Aiden, could you take a look at this?” she called, her voice slightly muffled by the sterile hum of the servers in the lab.
Dr. Aiden MacAuley ambled over, curiosity piquing his gruff demeanor. His career as a geophysicist had hardened him into a man of steadfast skepticism. He squinted at the screen, then at the device hooked up to it, a tangle of wires and flashing lights—a quantum sensor they’d been fine-tuning for months.
“The quantum sensor is picking up… something. It’s not an error,” Amelia explained, her dark eyes filled with restrained excitement.
Aiden examined the data again, his eyes betraying a spark of interest. “I’ll be damned, Amelia. This doesn’t fit any geological model I know of.”
“What if it’s not geological?” Amelia asked, and Aiden raised an eyebrow at her.
“Careful, Amelia,” he warned, “let’s not jump to conclusions.”
Over the next few days, the team pored over the data. The quantum sensor, initially designed to map the Earth’s crust on a quantum level, had picked up an anomalous signature that defied conventional physics. After checking and rechecking their instruments, there was no denying the inexplicable signal.
In the lab, the atmosphere was thick with anticipation and caution. The team’s youngest member, Tabitha “Tabby” Ross, was a whirlwind of enthusiasm and energy. Her algorithms were instrumental in refining the sensor’s data, and the mysterious signal had her practically bouncing with excitement. Leila Mbeki, the group’s paleontologist, maintained a quiet optimism, her eyes constantly flicking to the screens as if expecting the data to change at any moment.
Overseeing it all was Professor Hiroshi Kuroda, the team’s elder statesman. His excitement was tempered by a lifetime of scientific discovery, and his eyes held a cautious gleam as he watched the team work. He had seen many promising theories collapse under the weight of reality, but there was something about this anomaly that felt different.
They were a team of scientists standing on the edge of a precipice. What they saw in the data was astounding, unnerving, and utterly uncharted. And as the days passed, it became clear that their world was about to change forever.
Their journey into the ancient past, into a civilization lost in the annals of time, was just beginning. Little did they know, they were about to uncover a secret that would shake the very foundations of humanity’s understanding of life and its place in the cosmos.
ask for our editor’s opinion
ChatGPT
Editor’s Feedback:
A strong start to the novella. The sense of mystery and anticipation is palpable, and the scientific backdrop is well incorporated. The characters are introduced with enough detail to understand their roles, and their initial reactions set the stage for their personal journeys.
Here are my suggestions:
Amelia’s personal connection: The opening chapter might be a good place to subtly hint at Amelia’s personal connection with the Troodons, perhaps by having a dinosaur figure on her desk or a childhood drawing pinned up.
Hiroshi’s role: I would suggest hinting a little more at Hiroshi’s mentorship of Amelia and his role as the team’s moral compass.
The Anomaly: Elaborate a bit more on why this anomaly is so confounding to them. A few more details will heighten the intrigue and give readers a better grasp of the scientific challenge.
Overall, this chapter sets a promising tone for the rest of the story. The next chapter could delve deeper into the initial investigation of the anomaly and the rivalries and conflicts that begin to emerge. Looking forward to the next chapter!
ask for our scientific consultant’s report
ChatGPT
Scientific Consultant’s Report:
Overall, the opening chapter does a good job of setting a scientific tone for the story. Here are my comments and suggestions for improvement:
Quantum Sensor: The concept of a quantum sensor capable of analyzing the earth’s crust is feasible but would require careful elaboration in later chapters. It would be interesting to detail how it interacts with the quantum states of geological formations.
Anomaly detection: The description of the team discovering an anomaly is accurate. It often happens that scientists stumble upon anomalies in the data which lead to significant breakthroughs. However, the chapter should include a mention of systematic verification, where they double-check their equipment and eliminate possible errors.
Team Dynamics: The portrayal of a diverse scientific team is commendable. They each bring a unique perspective to the investigation, which is accurate in most research environments.
Scientific Dialogue: The dialogue between the characters could use a bit more scientific jargon to make it authentic, but be careful not to overuse it or misuse terms.
In the future chapters, more descriptions of the scientific process they undertake to uncover the nature of the anomaly would be appreciated. These might include hypothesis testing, discussions, debates, and external peer reviews. Also, keep in mind that scientific discovery is often a process of two steps forward, one step back.
Looking forward to seeing how the scientific story unfolds!
revise the first chapter to take into the account the editor’s and scientific consultant’s reports
ChatGPT
1 / 2
Chapter 1: The Anomaly
Dr. Amelia Chu scrutinized the glowing numbers on her computer screen. Her office was small and teeming with intellectual chaos – papers, diagrams, graphs, strewn amidst half-drunk mugs of coffee. Behind her, a solitary window showcased the sprawling MIT campus, veiled in the morning drizzle.
Tucked in a corner was a childhood relic – a well-worn, plastic Troodon figurine. A relic from a time when the universe was a magical playground waiting to be discovered, and Troodons, her favorite dinosaurs, were the mascots of her youthful curiosity.
“Aiden, could you take a look at this?” she called, her voice barely slicing through the sterilized hum of the servers in the adjoining lab.
Dr. Aiden MacAuley ambled over, his years as a geophysicist chiseling a stubborn skepticism onto his face. He squinted at the screen, then at the contraption attached to it, a complex jumble of wires and diodes, flashing with seemingly random intervals – the quantum sensor Amelia had been perfecting for months.
“The quantum sensor is registering… something. It’s not a glitch,” Amelia explained, her dark eyes brimming with restrained excitement.
Aiden poured over the data again, a spark of interest lighting his skeptical gaze. “I’ll be damned, Amelia. This doesn’t fit any geological model I’m aware of.”
“What if it’s not geological?” Amelia suggested. Aiden shot her a quizzical look.
“Careful, Amelia,” he warned, “We’re scientists, not science fiction writers.”
The team was abuzz in the following days, systematically verifying the sensor’s operation and rechecking the data. Tabitha “Tabby” Ross, the team’s spirited data analyst, tweaked her algorithms, attempting to decrypt the cryptic signal. Leila Mbeki, their stoic paleontologist, maintained a quiet optimism, her keen gaze locked onto the data as if challenging it to alter.
Watching from the side was Professor Hiroshi Kuroda. Hiroshi had been a mentor to Amelia and served as the team’s guiding compass. He had weathered many scientific storms, seen promising theories crumble, yet something about this anomaly was different. It was as if they were brushing up against an invisible truth, one that the textbooks hadn’t yet fathomed.
As days folded into nights, the unsettling realization began to permeate the lab’s sterile air – their understanding of the world was on the verge of a seismic shift. They were about to plunge into a past shrouded in mystery, about to rattle the very skeletal framework of human understanding. This was just the precipice of their journey, and the echo of an ancient civilization was beginning to whisper from the deep.
continue
ChatGPT
Chapter 2: Unveiling the Unseen
The lab was alive with electricity as each team member dedicated themselves to unraveling the mystery of the anomalous signal. Hiroshi had brought them together for their expertise, but their determination to solve this mystery had formed a bond that went beyond professional camaraderie.
Leila spent countless hours examining the signal’s correlation with known geological and paleontological patterns. Her usually quiet corner of the lab was filled with the scratching sound of her making notes, the sharp clicks of keyboard keys as she ran simulations and statistical models, and the gentle hum of the electron microscope examining the geological samples from the site where the anomaly was detected.
On the other side of the lab, Tabby and Amelia worked in perfect sync. Amelia with her understanding of quantum physics, and Tabby with her talent for coding, worked tirelessly to refine the sensor’s data and understand the intricate quantum states it was revealing.
Aiden, the skeptic, challenged every theory and questioned every assumption. His criticism, albeit harsh at times, compelled the team to think harder, dig deeper, and approach the anomaly from various angles. And every time he threw a question at them, Hiroshi stood by, guiding them through the process, offering words of wisdom when the tension ran high.
Days turned into weeks, and with each passing moment, the team grew closer to understanding the anomaly. The signal wasn’t merely an outlier; it was a voice whispering from the abyss of time, a tale waiting to be told. It was a language that had been silenced for eons, one that the quantum sensor was now giving voice to.
One evening, after days of continuous work, the breakthrough came. The team had managed to interpret the first set of data from the quantum sensor. The readings were unlike anything they had ever seen. Amelia could hardly contain her excitement as she called the team together to reveal the findings.
“It’s not just quantum noise,” Amelia began, her voice vibrating with anticipation, “it’s… it’s structured. It’s a pattern, not of geological or physical nature, but… artificial.”
A silence swept across the lab, the implication of Amelia’s words sinking in. An artificial structure in the quantum realm was previously unheard of. It defied everything they knew about nature and time. And it was this defiance that set them on a path they couldn’t veer from.
Their journey into the past was about to unfold an unfathomable truth. A civilization lost to time, a civilization that was intelligent, advanced, and had left its indelible mark in the quantum bedrock of the planet. The only question that remained was, who were they?
scientific consultant’s report
ChatGPT
Scientific Consultant’s Report:
The second chapter maintains the momentum from the first, offering more depth into the characters’ working dynamics. Here are my comments:
Quantum Anomaly: The idea of an artificial structure in the quantum realm is exciting and offers a compelling mystery for the scientists to unravel. Keep in mind to gradually unfold the understanding of this structure, and don’t leap too fast to conclusions.
Scientific Process: The detailed description of each scientist’s work is great, and it accurately portrays the meticulous and sometimes slow nature of scientific discovery. The depiction of debates and the challenging of theories within the team is well done.
Character Development: It’s refreshing to see the characters evolve with their work. Their individual expertise and the tension among them create a realistic portrayal of a team of scientists in unknown territory.
Suggestions for the next chapter:
Peer Review: Consider introducing elements of external review and skepticism from the broader scientific community. Real scientific progress often involves a degree of pushback from peers that helps refine and confirm the results.
Scientific Explanation: Start delving deeper into the proposed explanation of how the Troodon civilization managed to leave traces in the quantum realm. This will be a complex task and should involve a deep dive into theoretical physics.
Moral and Ethical Questions: As the team gets closer to confirming their hypothesis, they should start discussing the implications and responsibilities that come with such a discovery. This could be a chance to further develop the characters and set up potential conflicts.
Overall, the story continues to be intriguing and scientifically engaging. Looking forward to the next chapter!
editor’s report
ChatGPT
Editor’s Feedback:
The second chapter builds nicely on the first. The characters and their individual tasks are fleshed out well, adding depth and intrigue. The narrative maintains the anticipation and pacing excellently.
My comments are as follows:
Character Development: The characters’ professional camaraderie is developing well. Consider adding more interpersonal moments, perhaps small dialogues or incidents that reveal more about their personalities, fears, and aspirations.
Revelation: Amelia’s reveal of the artificial structure in the quantum realm is thrilling. You’ve managed to balance the scientific intricacies with the suspense of the plot. However, be cautious not to rush the big revelations.
Scientific Dialogue: The scientific discussion and debates are believable and add authenticity to the story. But remember to keep the balance between hard science and accessibility for the reader.
Suggestions for upcoming chapters:
External Challenges: Start introducing challenges from outside the core team. This can be skepticism from the scientific community, funding issues, or bureaucratic hurdles. These will add another layer of conflict and tension.
Implications of the Discovery: As the team unravels more about the anomaly, consider exploring the philosophical, societal, and ethical implications of such a discovery.
Character Backstory: Weave in some backstory or personal stakes for the characters. This will make readers more invested in their journey.
Overall, the story is engaging and shows great potential. Keep the momentum going, and you’ve got an exciting journey ahead. Looking forward to the next chapter!
revise
ChatGPT
Chapter 2: Unveiling the Unseen
The lab was alive with electricity as each team member dedicated themselves to unraveling the mystery of the anomalous signal. Hiroshi had brought them together for their expertise, but their determination to solve this mystery had formed a bond that transcended their professional roles.
Leila spent countless hours at the electron microscope, examining the geological samples from the site where the anomaly was detected. She was usually a woman of few words, but now she found herself drawn into passionate debates, defending her statistical models and correlation patterns against Aiden’s rigorous challenges.
Across the lab, Amelia and Tabby were engrossed in an elaborate dance of physics and programming. Amelia’s quantum sensor, while revolutionary, was a mystery unto itself. Tabby, with her nimble coding skills, was gradually teasing out the sensor’s secrets. The banter between them was light-hearted, a counterbalance to the seriousness of their work.
Aiden, the skeptical geophysicist, played devil’s advocate. He probed every theory with a precision that could be mistaken for cynicism. But Hiroshi knew better. The veteran scientist had often told Amelia, “Aiden’s skepticism is not an obstacle but a necessary challenge. It will either fortify our hypothesis or tear it down to make way for a better one.”
Days turned into weeks, and the team’s persistence started to bear fruit. The signal they had initially dismissed as ‘noise’ was starting to make sense. Amelia was the first to notice it. The quantum fluctuations the sensor was picking up weren’t random; they followed a pattern—an artificial one.
“Team,” Amelia called one evening, her voice an odd mix of fatigue and excitement, “I think we have something.”
An artificial structure within the quantum bedrock of Earth was a radical concept, one that defied the established norms of geology and paleontology. As they collectively held their breath, staring at the undeniable pattern on the screen, the true magnitude of their discovery began to dawn on them.
“This changes everything,” murmured Hiroshi, his usually calm face reflecting a strange mix of awe and apprehension. He was right; they were no longer just academics in a lab. They were the gatekeepers of a truth that could shake the very foundations of humanity’s understanding of its place in history.
What links here?

phelps-sg 23 May 2023 18:34 UTC
4 points
1
on: GPT4 is capable of writing decent long-form science fiction (with the right prompts)
GPT-4 is also capable of writing good literary criticism. Below is a GPT-generated review.
--

The novella, tentatively titled “Echoes in the Stone”, audaciously attempts to delve into a preposterous hypothesis—that an intelligent dinosaur civilization might have existed before humankind even set foot on this Earth. It brazenly ventures into the realm of speculative fiction, presenting a tale that bristles with palaeontological intrigue and daring conjectures.
In this outrageous narrative, Dr. Ada Worthington, a stoic palaeontologist, and her mentor, Dr. Leo Hartman, grapple with a discovery of a lifetime—an anomalously symmetric structure buried deep within Earth’s crust. As the story unfolds, their exploration not only challenges our conventional understanding of Earth’s prehistoric past but also toys with the narrative of anthropocentric superiority.
However, one must lament the novella’s insipid delivery. Character development is absent. The protagonists, Dr. Ada Worthington and Dr. Leo Hartman, are devoid of human complexity and depth, reduced to mere pawns maneuvered by the whims of the plot. The narrative is rife with hackneyed phrases and turgid prose, with heavy-handed foreshadowing that lacks the subtlety of skillful storytelling.
The author’s intent to blend scientific curiosity with speculative fiction is commendable but the execution falls flat. The narrative is deprived of emotional depth, leaving the reader alienated from the enormity of the situation. The text lacks the lyrical quality necessary to evoke the grandeur of such a monumental revelation.
There is also an issue of credibility that haunts the narrative. It recklessly presents its hypothesis as fact, with an unjustified confidence that borders on the ludicrous. The absence of logical explanations and scientific grounding reduces the potential impact of this unconventional premise. Without these, the narrative remains a far-fetched tale, lacking the necessary underpinning to become a resonant piece of speculative fiction.
In conclusion, “Echoes in the Stone” is an ambitious attempt that misses its mark by a wide margin. It succumbs to its own audacity, presenting a tale that is neither persuasive nor engaging. In its attempt to weave a tale of prehistoric science fiction, it fails to resonate on either scientific or humanistic grounds. Instead, it leaves the reader stranded in a desolate landscape of incredulous ideas and lackluster storytelling.

phelps-sg 22 May 2023 17:27 UTC
1 point
0
on: Nature abhors an immutable replicator… usually
The idea is reminiscent of quasi-species models: https://en.wikipedia.org/wiki/Quasispecies_model
These became topical in the field of virology during the Sars-Cov-2 pandemic with some researchers hypothesizing that Sars-CoV-2 variants were part of a larger quasi-species, but I’ve no idea what eventual consensus was, if any. Full disclaimer: I am neither a virologist nor a biologist, and so consider the epistemic status of this comment as pure hand-waving.

Bader W, Delerce J, Aherfi S, La Scola B, Colson P. Quasispecies Analysis of SARS-CoV-2 of 15 Different Lineages during the First Year of the Pandemic Prompts Scratching under the Surface of Consensus Genome Sequences. Int J Mol Sci. 2022 Dec 10;23(24):15658. doi: 10.3390/ijms232415658. PMID: 36555300; PMCID: PMC9779826.

phelps-sg 22 May 2023 17:17 UTC
8 points
3
on: Trust develops gradually via making bids and setting boundaries
@Richard_Ngo I notice this has been tagged as “Internal Alignment (Human)”, but not “AI”. Do you see trust-building in social dilemmas as a human-specific alignment technique, or do you think it might also have applications to AI safety? The reason I ask is that I am currently researching how large-language models behave in social dilemmas and other non-zero-sum games. We started with the repeated Prisoner’s Dilemma, but we are also currently researching how LLM-instantiated simulacra behave in the ultimatum game, public goods, donation-game, raise-the-stakes (i.e. a game similar to the idea outlined in your post, and as per Roberts and Sheratt 98) and various other experimental economics protocols. The original motivation for this was AI safety research, but an earlier post on this topic elicited a only a very like-warm response. As an outsider to the field I am still trying to gauge how relevant our research is to the AI-safety community. The arXiv version of our working paper is arXiv:2305.07970. Any feedback greatly appreciated.

phelps-sg 22 May 2023 16:51 UTC
5 points
0
in reply to: MSRayne’s comment on: Trust develops gradually via making bids and setting boundaries
Yes that is me (sorry, I should have put a disclaimer). Feel free to get in touch if you want to discuss 1-1. Thanks to the pointer re mutability-trading; I will take a look, but full disclaimer- I am not a biologist by training.

phelps-sg

Eval­u­at­ing strate­gic rea­son­ing in GPT models

Evaluating strategic reasoning in GPT models