Proto-Platypus Species Generative Modeling: Search for the Ancestors of Nature’s Most Diverse Creature

Dr. Brandon Staggs

1 Department of Mutant Homunculi Design and Analysis, Cranberry-Lemon University, Pittsburgh, PA


Across all the creatures on planet earth, few have more perplexing characteristics than the platypus. They have bills like ducks and nest with eggs like birds. They are venomous in the same way vipers are. Finally, they swim and lactate as a mammal would. This paper answers the question, what did the platypus evolve from. There is no known creature or even species which would explain their evolution. Using AI modelling of known animals, proto-platypus species are generated to create a model for the physiology and reproductive characteristics of an animal that could have feasibly evolved into a platypus. Then the proposed proto-platypus species will be cross-analyzed to model a more refined species that could have come soon before the platypus in a reverse phylogenetic tree. The AI analysis of the data determined that it isn’t probable for anything to evolve into a platypus and that God created the creature. 

Keywords: Platypus, Hairy Lizards, Busty Ducks, Beaver Eggs, Creationism, Artificial Intelligence, Evolutionary Models

1. Introduction

After years of blaming the weirdness of the platypus on the fact that it lives in Australia [1], the scientific community has lost the patience, credibility, and means to form a convincing explanation of their origin. By grouping the platypus into the order Monotremes, scientists have created the smallest possible order  of two species, because they could only fit in echidnas, the other egg laying mammal. These creatures still aren’t similar as echidnas who look like a browner version of sonic the hedgehog with a long snout and platypuses are duck beavers with genetics that doesn’t support that claim.

2. Background

The Aboriginal explanation for the Platypus involves a duck having a sordid love affair with a water rat. This oral tradition unfortunately doesn’t explain half of the Platypuses biology and has still yet to be accepted by the academic community. It’s a centuries old oral story that is still under review by Nature. They can’t accept it, but they also don’t have the heart to reject it.

The current evolutionary history claims that the platypus evolved from the Amniotes, who then evolved into the Synapsids who then evolved into the Monotremes instead of Marsupials and Placental mammals. The only issue is that the monotremes have as much in common with the birds and lizards of the Sauropsids than they do with the Synapsids. Their male-female genetics are more similar to birds, and their physiology is closer to lizards than any mammals. Yet they ooze milk like some Cronenberg monster. 

When Europeans discovered the Platypus, the scientific community thought it was a joke until someone captured and transported some back to London. After decades of studying them, the mystery only grew as to how convergent evolution could possibly produce such a unique creature. Just as AI has been used to generate art and extrapolate art outwards as a recent extrapolation of Starry Night [2] has been done, the same sort of misunderstood AI [3] will be further misapplied into the scientific realm to answer the question. What did the Platypus come from? 

Figure 1: Unguided AI generation of a cross between a Duck, a Beaver, and a Lizard
Figure 1: Unguided AI generation of a cross between a Duck, a Beaver, and a Lizard

 3. Proposed Proto-Platypus Species

As shown in figure 1, if we ask an AI to determine what a cross between a duck, a beaver, and a lizard would look like, it will only generate pure nightmare fuel. The AI must be guided using careful elements and then resynthesized as shown in [4] in which homunculi derived from alchemic equations were shown to have no souls and are extremely creepy. To guide this process, we set out to create generative models of ducks with mammary glands, egg nesting beavers with bills, and finally furry lizard/crocodile-Platypus hybrids.  

3.1 Duck Hybrid Proto species 

Two strategies were used to generate a feasible generative AI model of duck based proto platypuses. Due to the massive training dataset of busty humanoid women and ducks, the AI was asked to regenerate images of species crossed with women in bikinis crossed with ducks until it made something that wouldn’t scare away our reviewers. The least freaky trials were refined until we came upon the model which generated figure 2. Unfortunately, according to the beauty standards of the humanoids, the duck-humanoid homunculi is in project runway shape and not very close to the cozily plump and round platypus modeled in [5]. This model shape phenomenon in AI generation is further addressed in [6].

Figure 2: Duck species crossed with a woman humanoid
Figure 2: Duck species crossed with a woman humanoid

To combat this, we attempted to guide the AI into generating a female version of Daffy duck who was not close to model shape but sufficiently close to human shaped. Due to the plethora of images of Daffy and women, the AI had plenty of training data to work with. Unfortunately, there were not any female ducks portrayed for Daffy to court which would have further aided our AI generation. Much of the literature around this issue among the Saturday Morning cartoons has suggested this is in large part due to Daffy being too sassy [7], not knowing how to communicate [8], and frequently having its bill blown off by Elmer Fudd after mistaking rabbit season for duck season. Not having a bill is a sign of duck infertility [9].

Figure 3: Woman version of daffy duck
Figure 3: Woman version of daffy duck

As shown in image 3, the woman Daffy duck woman model is every bit as nonsensical of a comparison to the platypus as the humanoid woman version of a platypus but in a way we think it’ll average out to be a platypus. 

3.2 Beaver Hybrid Proto Species

The platypus appears to be very similar to the common beaver by being water dwelling and having a similar tail. The primary difference is that instead of wood cutting teeth, Platypuses have bills to find and eat shrimp with using tens of thousands of shark-like electro receptors, and the fact that they nest with eggs. To seed our model with an explainable creature, the AI was instructed to imagine a beaver like creature nesting with eggs as shown in figure 4, and what a beaver would look like with a bill as shown in figure 5. 

Figure 4: Beaver nesting some eggs
Figure 4: Beaver nesting some eggs

Thankfully, these images were much more wholesome than the duck generated proto-platypuses. I started getting far less concerned email traffic in between my advisor and my parents about what my home life was like. As generated from the AI, it is feasibly possible for a beaver to care about an egg! 

Figure 5: Beaver with a bill
Figure 5: Beaver with a bill

As shown in figure 5, beavers’ anatomical structures won’t generate anything terribly monstrous once imagined with a bill. While there are concerns about why a beaver would evolve a bill and give up the whole lumberjack thing for something to sense and eat tiny shrimp with, it’s good enough to seed our evolutionary model with. 

3.3 Lizard Hybrid Proto Species

There are no known images of furry lizards or billed lizards and the AI was very confused when we asked it to draw a furry Gecko in a toupee. The AI had a lot of trouble imagining any anatomically feasible lizards with watertight hair or bills. Using a bout of created thinking, the AI was prompted to cross a lizard with a weasel because they were similar shaped. 

Figure 6: Lizard crossed with a Weasel
Figure 6: Lizard crossed with a Weasel

While the furry gecko like thing shown in figure 6 isn’t exactly cute enough to sell car insurance, it was the best the AI could generate to create some lizard like proto-platypus. The weasel-gecko homunculi will provide our AI generative model a species that can scurry along the land with its arms and legs to the sides while potentially having a venomous spur on its back leg to stab sexual competitors with.  

Figure 7: Crocodile crossed with a platypus
Figure 7: Crocodile crossed with a platypus

Finally, we instructed the AI to cross a platypus with a crocodile in the off chance that some Australian crocodiles got cold a few million years ago and decided to evolve some fur because the climate refused to turn the thermostat up as proposed in [10]. Our shaky argument for how a Platypus could be related to a crocodile is that their snouts both face up in the water so they can breath while they’re floating around. A sample of the generative model of the platypus can be seen in figure 7 in which you will see why we didn’t weight it very high when generating the final proto-platypus. 

4. Evolutionary Model Development

In a move that will most certainly get this study rejected by any serious publication due to our extreme level of genius, we proposed the question, what if a phylogenetic tree didn’t just branch but could merge back together. Despite all of our species’ ingredients not being sexually attracted to each other in any way, we created a backwards phylogenetic tree seeding all the potential proto-platypus species into one massive genetic homunculi orgy the likes of which traditional biologists would never dream of or dare propose. 

Figure 8: Platypus Reverse Phylogenetic Tree
Figure 8: Platypus Reverse Phylogenetic Tree

Because we are bold, creative, and courageous innovators in the field of science, figure 8 is our proposed evolutionary model for the platypus. Just as the English language merged Nordic, Romantic, and Germanic speech into a seemingly unexplainable international phenomenon propagated by a monarchical lust for more tea and spices the Platypus could have come from several different species. According to [11], we are being idiots, but according to [12], we may be on to something. If our model is so insane, where’s the fossil record between Amniotes and Monotremes? 

5. Results and Discussion

After seeding the AI with all the proposed proto-platypus inter-species homunculi, our AI driven reverse phylogenetic model dared to go where no evolutionary model has gone before and has come to produce the revolutionary and controversial figure 9. 

Figure 9: AI sampling of God making the Platypus. 
Figure 9: AI sampling of God making the Platypus. 

In an upset which will certainly force my university to fail my dissertation while simultaneously and instantly securing my tenure at the creation museum which includes a life size Noah’s Ark [13], our AI has determined that God created the platypus. We didn’t even tell the AI about God. This scientific result nearly raises more questions than we were already dealing with when attempting to explain the platypus. Questions such as; if God is all powerful, can He create a rock that He cannot lift? Or the question, if God is all powerful, and all good, why did he let someone invent mumble rap? Or the most controversial question generated by this conclusion, which God? All these questions will have to be addressed in further research and probably some miracles, a messiah, or the fulfillment of several prophesies. 

6. Conclusion

According to the new revolutionarily reverse phylogenetic evolutionary model, and a creative use of artificial intelligence, God is the only possible explanation for the platypus. Maybe we shouldn’t have allowed for our AI to be so unsupervised or have unlabeled constants, or used an AI model that we don’t understand, but the science has spoken. According to the blindly trusted mathematics behind the mass of neural networks creating our generative models, we live in a much more spiritual world than the evidence previously suggested [14]. Some may call it faith; our research team calls it black box machine learning. 

7. Conflicts of Interest

I just got a Mid-Journey subscription and needed an excuse to play with it. It’s not my fault it became religious in the process. 


  1. Bruce ‘The Man Down Under’ McDonalds 2015 Australia’s strange animals and how our diversity makes us more special than You  :: Journal of the Land Down Under
  4. Edward Elric and Alphonse Elric Dec 2005 Alchemic Methods to Bring Humans back to Life: The Search for the Philosopher Stone Continues :: Journal of State Alchemy
  5. Cynthia Ronalds Aug 2019 A Dad Bod inspired Anatomical Model of the Duckbilled Platypus :: Deviant Art Account
  6. Susan Donaldsondurson Feb 2021 How Unrealistic Beauty Standards will Reinforce Barbee Bodies in AI Image Generation :: Journal of Machine Learning Feminist Theory
  7. Elliot Herbs Dec 2006 Where are all of the Daffy Duck Lady’s? Is he too Sassy? ::
  8. Elliot Herbs Jun 2008 Where are all of the Daffy Duck Lady’s? A Failure of Communication ::
  9. Elliot Herbs Dec 2006 Where are all of the Daffy Duck Lady’s? The Duck Bill Infertility Theory ::
  10. Bill ‘The Tacoma Kid’ Weathers Nov 2018 Like What if Crocodiles got Cold in the Ice Age? :: Journal of Hypothetical Biology
  11. Unimportant Academic probably a long time ago A Scientific Criticism Reverse Phylogenetic Tree Model :: Journal of Biological Has beens
  12. Esteemed Scientist Aug 2022 The Reverse Phylogenetic Tree Model is the Answer to Life’s Mysteries :: Every Journal worth Reading 
  14. Richard Dawkins The God Delusion

If you enjoyed this article in which we misused AI image generation machines to justify Creationism please like, share, and subscribe with your email, our twitter handle (@JABDE6), our facebook group hereor the Journal of Immaterial Science Subreddit for weekly content.

If you REEEEALY love the content, for the equivalent price of a Chipotle Burrito, chips and Queso, you could buy our new book Et Al with over 20 hand picked Jabde articles for your reading pleasure, it’s the perfect Christmas/Birthday gift for confusing your unsuspecting family members! Order on amazon here: Please rate and review so that you can brag to your friends about having opinions or showcase your excellent taste in reading material!

Published by B McGraw

B McGraw has lived a long and successful professional life as a software developer and researcher. After completing his BS in spaghetti coding at the department of the dark arts at Cranberry Lemon in 2005 he wasted no time in getting a masters in debugging by print statement in 2008 and obtaining his PhD with research in screwing up repos on Github in 2014. That's when he could finally get paid. In 2018 B McGraw finally made the big step of defaulting on his student loans and began advancing his career by adding his name on other people's research papers after finding one grammatical mistake in the Peer Review process.

Leave a Reply

%d bloggers like this: