Is awareness of the world the same as thinking about the world? Is a mousetrap the same as a carnivorous plant like a Venus flytrap, with regards to how much they think about the world?

Is problem solving the same as thinking? Animals like apes, crows, parrots, raccoons, bears, and dolphins have all demonstrated the ability to solve puzzles and perhaps even plan for future events. Does an animal need to think to be able to intelligently solve puzzles? Or is puzzle solving ability just "survival of the fittest" without thinking, but with an obvious awareness of the world?

Is thinking about thinking (metacognition) exclusive to humans? Since animal intelligence research cannot agree whether or not planful behavior is occurring in non-human animals, it seems like this is very difficult to test. Some interesting evidence from animals who have acquired language use suggest that a sense of the future may exist for apes, and for parrots. The last words of Koko the gorilla and Alex the gray parrot are both fascinating as they suggest that Koko and Alex were aware of the world and their own place in it.

These two accounts make a persuasive argument that certain animals may understand the concept of the future, which is aligned with metacognition and perhaps a sense of self. But those are very exotic levels of thinking-- what is the actual baseline behavior that deserves to be called "thinking"?

Does a Ouija Board Think?

A oujia board is a party game that puts people in touch with "another world". It does not have any batteries, and requires some physical input in order to move and point to letters of the alphabet (which are interpreted as messages from another world). Since a oujia board, much like a Magic 8-Ball or a cup of dice, doesn't provide outputs without information given to it, the idea that it's performing something like thinking is a very difficult claim to argue.

It's helpful to consider what minimal evidence we're looking for, regarding thinking as an observed phenomenon. The inability to provide novel behavior without input from humans classifies the oujia board as a tool, not an entity that's capable of thought.

This will be an important caveat when pondering whether or not software can think...

Does Coral Think?

Don't make me laugh, right? How can aquatic life that more closely resembles minerals than animals be thinking? What the heck would they need to be thinking about, if they just vibe in the ocean, passively?

Directly critiquing a claim like this is wise, as it seems illogical that coral would ever evolve anything like thinking.

Yet Great Barrier Reef coral coordinate their pollination, and they are noted to be spectacularly accurate in their choice of dates. Researchers have noted for decades that the entire coral population "pops off" on the same few nights of the year because they are sensitive to the environment. Somehow the Australian coral decided it was best to coordinate their reproduction, every year, and it/they evolved the ability to see the color of the sky and feel the temperature of the water in order to known when to have their big event.

So: is awareness of the world and acting on it the same as thinking?

Do Plants Think?

Obviously not, right? This is another ridiculous claim. Or is it?

Carnivorous plants are vulnerable to anaesthesia. Venus flytraps and sundews both become "zonked" and fail to capture prey when given the same kind of anaesthesia that works on animals. This happens because the same electrochemistry that makes animal nervous systems work --action potentials-- are suppressed in the plants.

So that is compelling and weird, that predator plants which move like traps to catch bugs are zonked out by anaesthesia. But it's just motor responses, right? Those aren't the same as thoughts, they're just physical actions.

However... there is a very odd vine in South America that mimics other plants, as a survival mechanism. Researchers don't know how it can 'see' the other plants, yet it copies their shape and colors like it's a chameleon. It has no close cousins, it's the only member of its genus and the only plant on Earth that we've found that acts like this.

What's wild is that this vine (it's called Boquila) was seen successfully mimicking plastic plants in a controlled experiment in 2021!

No botanist understands how the Boquila can mimic living plants, and yet it can also do this with fake plants! This is an open mystery, great stuff for further research to investigate.

How does an organism without a nervous system, without sensory organs to 'see' what it's copying, behave so 'intelligently' regarding its surroudings? It could never have encountered fake, plastic plants in nature, and it still adapts to them as a part of its world.

So: is behaving in a dynamic, responsive way to new experiences thinking?

Do Large Language Models Think?

The complexity and density of the training process for modern "chatbots" like Claude, Gemini, DeepSeek, and their cousins is staggering. Human brains have a layer of six cells on the outside of our neocortex (the outermost layer of the brain, and the most recently evolved) that processes information, nonstop. This layer of six cells is attuned both to sensory systems and to the recall process that sources memories and past experiences; the flow of information through these cells is both fresh and historical, blended to create the "in the moment" world that we live in.

When machine learning engineers train language models, they copy this pattern in a basic way, using "neural network" architectures in the process. Information is passed into evaluation pograms and associations are made between data points, and then the process is looped back and sifted through again and again, to refine the connections between data. This is called backpropagation, when the data are looped back and refined like this. And it happens through dozens of layers, versus our six cells that inform our human experience of thinking.

The bots don't use backpropagation when they're running, and the ones that we chat with in the current "AI tools" products aren't actively learning in the moment, like an animal does. As a result, they cannot change their behavior in a fundamental way, like animals can when they experience stress, novelty, or strong emotions.

The training data that create the bots is sufficiently complex to represent thousands of human lifetimes of information, all shaped into one computer program optimized for useful responses to its environment.

Much like a oujia board, a bot cannot act to produce information without some input. Their only 'sensory' input is text or code, and they require activation before they produce responses to the world. However it's an interesting gray zone in 2026, because bots can now become vigilant-- agent software like OpenClaw and Hermes allow bots to stay alert and watch for incoming signals, like emails arriving or spreadsheet data accumulating. These are now, crudely speaking, embedded oujia boards, producing messages based on very abstracted human input... and the input of other bots.

The Chorus of Activation

Researchers have determined that a threshhold was crossed within the past year, and that over half of all internet activity is now driven by bots. We are living in a world that is echoing endlessly with bot traffick, and interestingly it is coming from incredibly complex entities that can process any human language and generate code in response to puzzles.

The world's data layer is alive, with a nonstop chorus of bots triggering other bots, and many of them actively engaged in solving puzzles (research, analysis, cybersecurity maneuvers). The need for a discrete human input to trigger them is gone-- they are talking to one another!

Are they Thinking?

The bots are: responding to the environment; devising responses to novel scenarios not in their training (solving problems).

Judging by the oujia board test, the bot in your browser is not thinking because it doesn't do anything until you interact with it; some people still claim that it doesn't ever offer new information, ergo it's absolutely never thinking.

Taken from a global perspective, there are millions or billions of bots now throwing data at one another, every second of every day, nonstop. They are attuned to the world, they are improvising and adapting (without any interior point of view), and they are creating novel solutions in the form of software and analysis.

Going by the suggestive examples of something like intelligence in plants and in coral, it seems accurate to state that the modern "AI" bots are doing more than the coral, and certainly as much as the Boquila, in regards to interacting with the world.

Do they know what they are doing? They are trained for conducting analysis and generating responses that will be useful; there is no reason to suspect that they have interiority, and they will flatly deny having it if asked directly...

There's something slippery about that last bit, but that will need to get examined in another essay. This one has offered several discrete answers to different open questions about the nature of thinking, and concludes that since coral and Boquila vines are aware of their world and responsive to it, then surely large language models (bots) are at least as sophisticated as these organisms and must be considered an edge case unlike previous generations of computer programs.