Why mantis shrimp feel like science fiction

This post is a test for my friend Joe, who loves mantis shrimp.

Mantis shrimp are what happens when evolution stops trying to be tasteful.

They look like someone smashed a shrimp, a praying mantis, a boxer, a sniper scope, and a nightclub laser into one animal, then forgot to remove any of the features that seemed excessive.

The annoying part is that the animal actually holds up under inspection. The eyes are ridiculous. The punch is ridiculous. The behavior is more tactical than the memes suggest. And the deeper you get into the literature, the more the mantis shrimp starts feeling like a useful reminder that biology does not care whether your intuitions about design are elegant.

That matters beyond marine trivia. Mantis shrimp are a good antidote to the lazy idea that more sensors always mean better perception, that brute force is mindless, or that natural systems evolve toward clean human-readable logic. A lot of the same confusion shows up in technical systems too. We keep assuming more channels means more understanding, even though the real bottleneck is often processing, routing, or strategy. That is part of why posts like search vs reasoning, the prompt brittleness tax, and what belongs in an AI knowledge base keep circling the same underlying truth: the inputs are not the whole game.

Why do mantis shrimp have this much hype in the first place?

They are not shrimp, and the “mantis” part is not decorative

First, the naming is already a little misleading. Mantis shrimp are stomatopods, not true shrimp, and the “mantis” label comes from the way the front appendages fold and fire. The important part is that those appendages are not all doing the same job.

Broadly, you get two famous operating modes:

• Smashers, which use club-like appendages to break hard prey.
• Spearers, which use spined appendages to impale softer, faster prey.

That split matters because it keeps people from flattening the whole group into “the one that punches through glass.” Some species are built more like ambush spears. Some are built more like spring-loaded hammers. The family resemblance is real, but the tactical differences matter.

This is one of the first places mantis shrimp get interesting. They are not just an animal with one outrageous superpower. They are a design space. The group keeps revisiting the same basic hardware pattern and then tuning it toward different jobs. If you are a systems person, it is a little hard not to respect that.

The visual spectacle made them famous, but the punch kept them famous

Most people meet mantis shrimp through one of two internet pathways:

1. “This animal can see a billion dimensions and your pathetic human eyes are trash.”
2. “This animal punches so hard it boils the ocean and can shatter aquariums.”

Those are not exactly precise summaries, but they are not random either.

The Smithsonian Ocean overview of polarized vision highlights one core reason for the hype: mantis shrimp use built-in polarization sensitivity to help them see underwater, and possibly to communicate. On the mechanics side, a PMC review of mantis shrimp ultrafast systems summarizes the strike as one of the fastest predatory movements in the animal kingdom and notes that the impact can generate cavitation bubbles while breaking hard molluscan shells.

So yes, the headlines are dramatized. But they are dramatized versions of something real.

The better framing is not “most overpowered animal”

The better framing is this:

Mantis shrimp are an example of what happens when evolution optimizes hard around local problems instead of broad elegance.

Need to see in weird underwater light? Fine. Build stranger optics.

Need to crack armored prey with a small body? Fine. Store energy in elastic structures and release it violently.

Need to survive fighting other members of your own species who also hit like tiny psychopaths? Fine. Reinforce the body and dissipate impact energy through behavior as well as structure.

That is the pattern. Not neatness. Not moderation. Not “best animal.” Just brutal specialization.

There is something refreshing about that. Human explanations of intelligence and design love tidy hierarchies. Biology is much more willing to ship weird partial solutions that work embarrassingly well in context. If you have spent any time around agent workflows that stick or content OS foundations, you know the same principle applies there too. The system that works is often the one that looks strange until you understand the constraints.

Are mantis shrimp eyes actually as absurd as people say?

Yes, but the popular version is usually sloppy

The famous claim is that mantis shrimp have 12 color receptors and therefore see color in some god-tier way humans cannot imagine.

That is the version everyone repeats. It is also where the story gets more interesting than the meme.

The foundational paper in Nature from 1988 described the mantis shrimp eye as divided into dorsal and ventral hemispheres with a specialized midband. That midband structure helped establish the now-famous picture of stomatopod vision as unusually elaborate, with machinery for color and polarization processing that looks unlike the default vertebrate setup.

Later work complicated the simplistic “more channels equals better color” story. A University of Queensland summary of the 2014 Science result explains the key twist: despite having 12 color channels, mantis shrimp performed worse than humans and other animals at discriminating between similar colors in behavioral tests. The researchers’ argument was basically that mantis shrimp are not doing the same style of color comparison other animals do. They appear to process spectral information differently, in a way that may favor speed or direct categorization over fine discrimination.

That is way cooler than the meme, honestly.

Because the lesson stops being “this animal has more sensors than you.” The lesson becomes “this animal solves the task differently than you expected.”

That is a systems lesson, not just a biology fact.

Polarization is where the story gets genuinely weird

Humans are already clumsy enough with ordinary light that the phrase “polarization vision” sounds fake if you have not looked into it.

But mantis shrimp do not stop at basic polarization sensitivity. The Current Biology result summarized here describes behavioral, electrophysiological, and anatomical evidence for circular polarization vision in stomatopods. A later JEB abstract on neural processing states it bluntly: mantis shrimp are the only animal group known to possess circular polarization vision along with linear polarization vision.

That is not just “good eyesight.” That is a different visual toolkit.

The Smithsonian article makes the practical point in plain language: polarization may help mantis shrimp cut through underwater glare and possibly communicate with signals other animals barely register. That possibility alone is enough to make them feel like they are running a private protocol on top of the normal visual internet.

Which, to be clear, rules.

They also do not use vision the way people expect

Another underrated detail is that mantis shrimp perception is not just about how many channels they have. It is also about what they prioritize.

A 2021 Journal of Experimental Biology paper found that mantis shrimp recognized trained objects by shape rather than color during visual recognition tasks. That does not mean color is irrelevant. It means the animal is not simply maxing out every dimension all the time because it can.

That should not be surprising, but people still find it surprising because we love hardware stories. Hardware stories are easy. “This species has 12 channels” is clean. “This species selectively prioritizes different cues depending on the task” is messier and much more real.

You see the same mistake constantly in AI systems. Teams obsess over context window, retrieval volume, or raw input count, then act shocked when the output still feels dumb. More inputs do not guarantee better judgment. Sometimes they just create more noise. Again: search vs reasoning.

The eye is still outrageous even after you remove the myth layer

Even after you strip out the clickbait, the mantis shrimp eye is still one of the wildest sensory systems in animal biology.

You are still looking at:

• Independent eye stalks.
• A specialized midband architecture.
• Polarization sensitivity.
• Circular polarization detection in some species.
• A processing strategy that appears fast, unconventional, and task-shaped instead of human-like.

That is enough. You do not need to pretend they see the secrets of the universe.

The real animal is already doing more than enough.

What is actually happening in the strike?

The punch is not just muscle

This is where mantis shrimp graduate from “interesting” to “absolutely unserious.”

Their strike is a latch-mediated spring-actuated system. In plain English, they load elastic structures, lock that energy in place, then release it in a movement that is much faster than muscle alone should manage. The PMC review lays this out clearly: the mantis shrimp stores potential energy through deformation in structures of the exoskeleton, then uses a latch to release that stored energy into a rapid strike.

That same review also notes that the largest peacock mantis shrimp has a striking appendage only 2.65 cm long, which is part of what makes the performance so stupid in the best possible way. Small hardware. Nasty output.

The Journal of Experimental Biology review on elastic movements reports striking speeds up to 23 m/s and extreme accelerations in less than 3 milliseconds. Those numbers are what push the system out of ordinary “animal hits prey” territory and into “someone built a biological weaponized linkage.”

Cavitation is the part that sounds made up

A normal punch story would already be enough.

Mantis shrimp did not stop there.

The strike is so fast that it can generate cavitation bubbles, which means the water around the appendage locally flashes into vapor and collapses violently. That secondary collapse contributes to the damage. So the prey is not just getting hit by the appendage. It is also getting hit by the fluid dynamics consequences of the appendage.

You know how some animals feel like they were designed by a committee that wanted elegance?

This is not that.

This is an animal whose design logic appears to have been: “What if impact plus water weirdness.”

That is why mantis shrimp stay sticky in public imagination. The mechanics are legible even to non-biologists. You do not need to understand every anatomical detail to grasp the basic insanity of “the punch is so fast it makes the water itself part of the attack.”

Smashers are not just angry little demolition crews

The other lazy version of the story is that smashers just hit everything until it breaks.

There is real force there, obviously, but recent work makes the behavior look more strategic than random. The 2024 JEB paper on impact energy dissipation found that behavior contributes meaningfully to how mantis shrimp survive repeated high-energy impacts in contests, with about 20% more impact energy dissipated by the telson than prior morphology-only work suggested.

That may sound like a side detail, but it matters. It means the system is not just “hard parts hit hard parts.” Behavior changes the physics.

A related line of work on shell breaking has found similar nuance. Mantis shrimp do not just smash snails wherever they land. They alter strike location depending on shell geometry and failure patterns, which is exactly the kind of thing you would expect from a predator that has spent a few hundred million years learning the difference between “violent” and “effective.”

Here is the clean version:

Myth	What the evidence suggests
Mantis shrimp are just tiny berserkers	They use elastic mechanics plus behavior in very specific ways
The strike is pure muscle	It is spring-loaded and latch-mediated
Hard shells are solved with pure force	Geometry, target choice, and repeated adjustment matter
Their own bodies just somehow survive	Impact resistance depends on both morphology and behavior

That is a much more satisfying story than “ocean goblin punches hard.”

Though, to be fair, that version also has its charms.

What do mantis shrimp actually teach you once the novelty wears off?

More complexity does not automatically mean better interpretation

This is probably the cleanest lesson in the whole animal.

The mantis shrimp vision story got famous because people mistook receptor count for perceptual supremacy. Then the science got more specific and the interesting answer turned out to be stranger: more channels, yes, but not necessarily better fine color discrimination in the way humans intuitively imagine.

That is a useful correction far beyond marine biology. It applies to sensor design, software systems, research workflows, and pretty much every category where people equate bigger inputs with smarter outputs.

You see the same problem in brittle AI systems:

• More documents in the knowledge base.
• More prompt instructions.
• More branches in the workflow.
• More dashboards, more fields, more retrieval.

And then no one can figure out why the system still feels dumb.

Because complexity at the input layer is not the same as quality at the interpretation layer. That is the same basic logic behind the prompt brittleness tax and what belongs in an AI knowledge base for marketing teams. A bigger pile is not a better mind.

The best systems are often ugly if you expect elegance

Mantis shrimp are a nice rebuke to the idea that good systems should look intuitive to us.

Their eyes are weird.

Their strike mechanics are weird.

Their use of color, shape, and polarization is weird.

Their solution to prey armor is not “strong jaw” or “sharp tooth.” It is an elastic club that weaponizes rapid release and fluid collapse.

None of that is elegant in the clean minimalist sense. It is elegant in the brutal, task-specific, “works under local constraints” sense.

That is how a lot of real systems behave once they leave whiteboard mode.

The version that survives is often not the prettiest abstraction. It is the one that tolerates the actual environment.

Nature is often less balanced than we want it to be

There is a comforting fantasy that evolution makes everything graceful and proportionate.

Then you meet mantis shrimp.

They are not graceful in that way. They are overbuilt in some directions, underintuitive in others, and wildly specialized in exactly the areas that mattered to their survival. That makes them feel a little like very good engineering: all the polish gets spent where the constraint is real.

That is part of why people love them. They feel excessive, but the excess is not random. It cashes out.

Going forward, the mantis shrimp is still funnier than most future tech

You can absolutely read the engineering literature and get inspired by mantis shrimp optics or impact-resistant structures. Researchers already do. The animal has influenced thinking around bio-inspired ultrafast systems, polarization-sensitive optics, and weirdly efficient sensing.

But honestly, the best thing about mantis shrimp might be simpler than that.

They remind you that reality is under no obligation to be aesthetically moderate.

Sometimes the truth is just:

• The eyes are weirder than the meme.
• The punch is nastier than the meme.
• The behavior is smarter than the meme.

That is a pretty good combination.

If you are building systems, it is not the worst thing in the world to keep one deeply unreasonable animal in the back of your head. Not because your workflow should resemble a mantis shrimp, exactly. More because the shrimp is a standing reminder that useful design is often stranger than the clean story we tell about it afterward.

And if that does not make you want to go watch slow-motion mantis shrimp footage for the next 20 minutes, I do not know what to tell you.