Brainfuck was originally developed as a joke, the fact that it found a niche use in biological computer science is amazing.

So it turns out life doesn't emerge in spite of entropy...it emerges BECAUSE of it. It implies that not only should life be common, it should be EVERYWHERE. Heck, ANY system left to its own devices should form something resembling life according to this. A universe made of life. Never before has a coding program been more appropriately named.

Brainfuck is just a way to describe a Turing Machine in a programming language format, so even being a joke it was a solid concept far before it was created.

The fact this title can exist in this world on a non click bait channel is wild.

Some thoughts on this: 1. Limited space IS a fitness criterion in itself. 2. Longer code means less probability for a working program to emerge. 3. If space is the only limit, fast replication is the optimal fitness. So nothing more complex has an advantage and will always be suppressed by fast replication.

If ANYTHING is occurring randomly, it will either accomplish nothing, destroy itself, or self-replicate. Over time, the first two do nothing. Only the 3rd causes larger change, and so it will.

A "Game of life" pattern can actually self-replicate. The program "Golly" provides one by default, called "linear-propagator-p237228340". It even has a double line of "gliders" encoding the information on how to build the end stations, like DNA.

Thanks Anton, watching Anton is scientifically proven to improve your day ( even if it's not been great ) and it definitely improves your mind. You even inspired me to make my own YT channel 💛!

Chaos mimicking life in simulations sounds like some 40k lore

Absolutely no evolutionary pressure is a stretch. They would need infinite space in order to avoid the constraints of a closed system to produce an environment with no pressures and then the code itself would be a pressure since it has the ability to alter the local environment. It’s still cool and I’m looking forward to what they discover with it but exactness is not objective.

The way self-replication arises from just random changes - and with no evolutionary pressures - is fascinating!

If I had one vote to cast for "truly wonderful person" it would be you Anton. Your content is so pure and your so jolly about it

They should isolate samples of those self-replicators and give them fresh environments. Then merging different environments at a later state.

This is how we get replicators, have you not seen Stargate??

We need to make a place called "science island" for experiments like this. We don't want self replicating robots getting out in the wild.

“You start with a random clump of atoms, and if you shine light on it for long enough, it should not be so surprising that you get a plant”

Seems like the paper authors might have incorrectly assumed that randomly generated programs in a given language will produce random behavior. They will not. The underlying rules and structure of any programming language will inevitably generate strong biases in the way their respective programs behave.

If I can quote Monty Python- "So remember, when you're feeling very small and insecure How amazingly unlikely is your birth And pray that there's intelligent life somewhere out in space 'Cause it's bugger all down here on Earth"

Thank you very much Anton, for this once again super exciting video! It's great to be kept so well informed by you about exciting new discoveries and opportunities to create new knowledge. And of course for the warm smile at the end, which I always like to return. All the best for you, your family and your channel!

I honestly don't know why it was a surprise. If you have programs set to change their properties randomly and they are put together in a place with limited space, the fact is that the ones that are better able to survive, do survive, and the ones that aren't as good will die out. There doesn't have to be any inherent mechanism. It's simply limited space and random changes. Eventually, a random change makes something that self-replicates and is thus able to survive much better than all other programs, and so it spreads and snuffs out everything else until something better takes over. Seems obvious to me.

I don't understand why this is unexpected? If you give what is essentially a "grid of random mutating programs which may interact with their neighbor in the grid" the POSSIBILITY of affecting other programs and the possibility of affecting them in such a way that they copy the original, then eventually, replication will happen, obviously. The fact there is no pressure to do anything in particular actually just makes it more likely, not less likely, because there is no force that could prevent such a replicating program to occur. It just becomes a matter of time until the randomness creates a combination that will replicate. This research so abstract I don't understand how it has any impact on our understanding of reality - a reality in which we already KNOW life can happen through Abiogenesis (we are here after all!), but need to understand how it can happen chemically, under which circumstances (because unlike this abstract research, life in reality is subject to many pressures and dangers, and possibly requires elaborate chemical and physical circumstances to even have any chance occur at all), and how such self-replicating chemical structures could eventually become more and more complex while surviving over billions of years of planetary and stellar evolution.

09:51 you can say that again 😂 length is very important when it comes to replication

It’s incredible to see such kindness and love. Thank you all!

Thanks again for your "Above and Beyond" most excellent reports and even MORE awesome Graphics, your graphics are the absolute Best !! Love watching , thx so much Anton !!!

OMG, this changes everything! Well, it probably doesn't, but I've always wanted to say that.

The game if life is interesting but I don't buy the assumption of emergence from simple rules. Sure, there are simple rules but the infrastructure on which those rules run is quiet complex.

Between this simple program and its emergent properties; and the experiment combining primordial atmospheric gasses and lightning producing biochemicals, some self-replicating; life arising on a lifeless planet being an emergent condition is both unsurprising but fascinating and awe-inspiring!

I am glad someone finally created a program for this, I have often pondered a simulation of this nature. Imagine what could come from running a larger simulation with our most powerful computers allowing for faster computations. Digital evolution on fast forward. Incredibly exciting.

7:52 “billions of years” i think life arose on earth within less than a billion years after its crust cooled enough.

I've only read the title so far but I'm going with 'Just because we can doesn't mean we should' - The updates would be a nightmare.

A blind guess would be that in order for more complexity to appear, a kind of predator needs to evolve, like, every pattern is selecting for faster replication because there is a limited space (resources), but a predator isn't competing directly with fast replicators because its "food" are the fast replicators.

In cybersecurity we deal with these "self-replicating life forms" all the time.

PSA: Don't go into the von Neumann rabbit-hole. The guy was an absolute intellectual unit. Apart from creating the concept of cellular automata (predating Conway by three decades) as mentioned here, he was a major contributor in the Manhattan Project, invented the mathematical field of game theory, created the modern computer architecture and in general was to most mathematical researchers what they are to us. "When George Dantzig brought von Neumann an unsolved problem in linear programming "as I would to an ordinary mortal", on which there had been no published literature, he was astonished when von Neumann said "Oh, that!", before offhandedly giving a lecture of over an hour, explaining how to solve the problem using the hitherto unconceived theory of duality." Of note, the George Dantzig telling the story is the mathematician who as a student did the "solve unsolved blackboard problems" of Good Will Hunting fame, who was later awarded the National Medal of Science by President Ford.

Small issue: your introduction of John Von Newmann is really underselling his contribution, and made him sound like some science fiction writer. He made huge contributions to foundation of quantum mechanics, quantum information, chemistry, functional analysis, game theory... and so on. He is often considered as one of the smartest person and polymath ever lived. I know the Von Newmann Probe is relevant here.... but only talking about that really misleads the audience in thinking what kind of person Von Newmann is.

limited space would correlate as a fitness problem as soon as self replication commences

Consciousness/Life Always finds a way... Reminds me of one of my favorite Stories by Theodore Sturgeon "The Dreaming Jewels" about Sentient Crystals that can Manifest into Human Forms.

This study basically suggest that if a bunch of mundane particles in primordial soup produces self-replicating life over here, then it also does it over there. And aliens exist.

Hahaha, BRAINF**K, huh. Half an hour on a laptop? The 3D versions of Conway's Game of Life are really neat. Maybe someday we'll have Conway's Game of Thrones! Thanks Anton.

As long as conditions are the same, the fastest replicator takes over. As soon as you introduce variable conditions , complexity becomes important.

'If you leave hydrogen out in the sun long enough, it will start thinking about itself'

I may be missing something, but isn't DNA made up of only 4 items, arranged in a complex string of 4 item blocks... ?

I would guess this doesn't happen anymore on earth because once a self-replicator appears, the environment becomes pressured by it to the point another one has no time to appear.

9:53 "The length plays a very important role in determining if self replication is possible" 😭

Have a wonderful day Anton.

An infinite number of monkeys on an infinite number of typewriters will eventually write Shakespeare. So if 4 billion years ago trillions of random proto-lifeforms were doing their random thing, as soon as one of them figured out self replication, it took over everything and then competed with itself and bettered itself until here we are. And the process is still ongoing.

Just in case anyone has time to look him up, John Von Neumann is one of the most impressive minds t ever walk this Earth. He is ‘Your favorite rappers favorite rapper’ kind of mind

The missing variable perhaps seems to be 'Environment' or 'Primordial Soup'. As it was, the code was running in an either absent or neutral environment and as such the code did not have any stresses to evolve differently. A good next test would be to re-run the simulations in separate instances where environment variables are introduced. Those different environments may then result in different self-replicating code that is best-suited for each environment. Then introduce the different environments with their individual limitations. See how the different evolutions cope when they interact with the new environments and differing evolved code.

Avi Wigderson received the Turing award, the “Nobel Prize of computing,” partly for formally connecting randomness with mathematical functions that are hard to compute. He and his colleagues created a process that takes a suitably complex function and outputs “pseudorandom” bits that can’t be efficiently distinguished from truly random bits. Randomness, it seems, is just computation we cannot predict.

One might infer that biological life / evolution is a natural course of events with enough time and increasing entropy.

Noticed the the people who say we can't trust our memory also discard history for hypothetical models

The general rule is quite simply: systems and patterns which propagate more effectively will always surpass those that don't. This applies to biological systems, businesses, ideologies, black holes, literally everything. The form can vary wildly depending on the system and environment, just as the RNA and computer program experiments demonstrated (rewarding different mechanisms of propagation), but the underlying rule is still holds true. In essence, it's the same reason x⁴ is always bigger than x². In any universe where you have _stuff_ that can change and causality, you're likely - if not guaranteed - to observe this rule.

The Von Neumann error still haunts us to this day. Life is not binary, but we are still trying to model it as such. Complexity Theory in binary ignores the material realities used to do the computation.

It sort of makes sense to me. If something self replicating can emerge, then eventually it will. Natural selection is an emergent property of self replicating systems, so long as the rules aren’t too rigid.

what is the probability, for each "language" that it is self replicating when randomly written.. ? In the extreme case, if the language has only 2 instructions ("replicate" or "don't replicate").. and the thing is random, of course it will appear and conquer. the idea of "physically existing" and "ability to copy information" creates a loop.. and space imply competition... of course something more complex will emerge.. if it means more ability to copy..

Watched an interview with denis noble on the KZfaq channel "Variable Minds" and this video reminded me of some parallels discussed. As far as I understood noble suggests that there exists potentially inherent function of random events in a system, which continually develops eventually resulting in 'containers' within which continued interactions are constrained. The inherent functionality evolves as more and more complex containerized objects appear in the primordial soup, which eventually could lead to what we describe as life, and evolution from "single cell" organisms towards complexity, multi cellular organisms as time goes on. The experiment here seem to result in self replication within the limited number of population, and without any competition. Not sure if that makes any sense but I feel such experiments will eventually further our understanding and are fascinating. Shout out to Variable Minds as each interview I've listened to was excellent and I'd highly recommend 😊

Reminds me of a Greg Egan short story where someone uses self replicating AI to create intelligence wildly different from our own

I think thermodynamics is a major barrier which is completely missing from this computer simulation. In the real world, molecules or even quasi-cellular structures, will need energy to survive and reproduce. That is the reason why the biological experiments favor simplicity. Because, it is less energy-consuming to build simple structures rather than complex structures. On the computer, there is no need for external energy input to be simulated. So, the programs can grow unaided. But on a planet, not only the conditions will need to be favorable (liquid water, atmosphere, temperature etc) but the thermodynamics need to be favorable as well. The energy flow patterns should be able to support complex structures.. On earth, Sun's visible and UV radiation can do both the things, form complex structures, as well break them down. Our atmosphere blocks enough UV light to prevent cellular damage, but still allows enough sunlight to pass through to allow photosynthesis and other 'constructive' processes. This right balance may be critical for early evolution too. Too much radiation or too little radiation may prevent early abiogenesis. On planets with underground oceans (like Europa), there is no sunlight at all. Their liquid oceans may remain sterile despite favorable conditions.

and the humans were created from clay gets a whole new meaning 11:56

I have Conway's Game of Life on my phone and it's fascinating!

It stopped evolve due to a limited and to static environment. In the next step there need to be external factors that change the environment, and that the self replicating life also have effects on the environment. Greate video.

Short answer is no. Longer answer is no.

It's impressive that self-replication emerged without direction. The problem is that life arose here from the planetary geochemistry. It didn't arise out of nothing. There must also have been some sort of selection for more complex and fitter lifeforms, at first microbial, then eukaryotic, then multicellular. There were a lot of energy inputs on early Earth which drove the development of life in general. And, life altered its environment. This system doesn't get altered by anything that arises.

This wouldn't be the first time this has been done... I remember a scientist did this back in the early 90ies where he created a code resembling a single cell organism and introduced artificial environmental living conditions.. and over time it multiplied and evolved creating entirely new pieces.

The motivation of each cell is to make your neighbours accept most of your properties and to accept as little as possible from your neighbours properties. Strengthen your output and weaken the input to ensure you won't change while others will. If the goal of a cell is to force surrounding cells to be like itself., then there is cellular motivation.

I fully agree that life may need more time to emerge than most scholars think. In other words, we strongly underestimate the force of evolution.

why is this a big shock? Obviously if there is 1 replicating code happening by random chance, it will start replicating, what else do people expect? Eventually there will be a random algorithm that wants to survive, what will it do? Well obviously it will try to survive

In fact, they rather found a good a argument that self-replication might be a natural, spontaneous process for information, rather than for biology.

3:10 things i didn't have on my bingo card for this century ... Someone stumbling across a genuinely practical usage for Brainfu@£

Who watched this and then immediately thought, this proves we are in a simulation ran on a laptop?

It's like no one has seen Stargate 👀

I wonder if Von Neumann's work may have inspired the Star Trek TNG episode _Home Soil_ (S1E18). The crew discovers that a terraforming project is disrupting a crystalline form of intelligent life. Individual crystals aren't intelligent, but in their natural environment (threatened by the terraforming) intelligence emerges in spectacular fashion.

It may blow your mind, but evolution didn't start with life. The whole "dead" universe also undergoes huge and really slow "survival game". Some stars burn for longer, some planets can host life, some can grow more, some get stripped of materials, some stars did better in the early universe, etc. etc.

Holy shit doesnt this like completely refute all religious arguments against evolution?

... is it self aware ? ... no "then it aint life"

I think key to complexity is competition. Life needs energy, Space and other resources which directs evolution into other things than fast replication and simplicity.

I was homeless, got into drugs, went into prisons, then i got to know Jesus, He changed my life.. Now i make 22k weekly. have a home, a wife, a lovely daughter... A child of God. HALLELUJAH

Very strange! 😮

Emergence and self-replication is interesting. But when those agents become significantly more complicated from one generation to the next, THAT will be profound.

Great first step! For the next they need to vary the landscape. Essentially this is currently a simulation within a flat world, or alternatively within a single drop of water. In the real world there are areas of high pressure, low, temperate, humidity, solar radiation exposure and so forth. To do this perhaps the team could add initial rules to regions of the world. E.g. a region where random characters are corrupted (radiation), another in which code executes more slowly (low temperature), another in which a command is suppressed (nutritient restrictions) and so forth. Simulated life is then met with challenges beyond simply space for replication and infinite resources to do so. I suspect adaptation and synergies would emerge in this level of simulation. For the subsequent simulation, it would be great to find a way for the organisms to influence the rules of the regions. E.g. to somehow enable moving the region, lessening the rules, whittling down some of the regions. I.e. akin to a moss creating local humidity on a rock, or algae in a body of water reducing the temperature.

DNA is roughly binary where three "digits" (beginning almost anywhere including reverse) are a codons which might be analogous to the eight symbols of brainfk. Also like LISP, DNA is both the code and structure, amino acids that generate amino acids that build proteins.

Big fan of most of your videos, but really surprised you picked out this study. The setup and results are extremely underwhelming. This is the kind of program I'd expect a student to cobble together. Despite the supposed grandeur parallels to life replication, all the complexity has been baked into the "environment" control code. That said, this is a very interesting area of research. I hope they do something more sophisticated with something like alphafold in the future.

These are called Boid Phenomenon. They are the subject of AI hallucinations and the way shapes form when swarms of birds and herding animals move. They follow simple rules but are highly complex in an open environment

Replication is a fundamental property of the physical basis of our Universe. Replication was at the beginning of all beginnings, at the moment of the emergence of the Universe, and it was replication that created our Universe.

…raises a lot of questions, for example, life from clay… And my mind blows off! 🤯 “In Aztec mythology, the gods created clay men (!!!) but later abandoned this idea because the clay men were fragile and unstable and failed to praise them. Then, the gods created the first human beings out of wood, but they were imperfect and lacking feelings, so they were destroyed and turned into monkeys. Finally, the gods created man from corn, mixing yellow corn and wood to build the skin and red corn to make the blood. These beings were perfect, but they were so similar to the gods that they felt threatened and decided to send them fog to counteract their greatness.”

Imagine inserting this as a payload with the ability for it to escape sandboxing or buffer overflow. I think someone just discovered what could be a very nasty attack vector that is easy to hide and run.

7:35 we don't have to imagine, just look around

There was a fitness function. It just wasn't explicit. They were competing for space on the grid, and probably execution time.

Why did they expect nothing to happen? They gave random code the ability to interact. It was only a matter of time until random inserts would create code capable of self-replication. They basically created a cross between bacterium and virus. There are billions of possible code configurations capable of self-replication. This was like infinite monkeys, with infinite typewriters, randomly recreating someones business card. While I never would have expected a 100% success rate, expecting a 0% one was foolish.

Amazing results but I agree with some of the criticism: the emergence of self-replication doesn't imply the eventual emergence of evolution.

Kolmogorov Complexity theory in a self referential system. Makes sense, then certain amounts of complexity give rise to emergent properties of consciousness, which in turn, collapse the wave function within the self referential system. Consciousness acts as a product as well as a driver of the system

Those scientist that say that this cant evolve into more complex life completely dropped the ball. Is like saying buhu cool cell simulation but why isn't it evolving into elephants? - Well because emergence happens when you scale up the simulation universe by maybe millions times. or trillions times.

It is amazing how you can dive into a paper on a subject that isn't exactly your area of specialty and explain the hell out of it so well!

but now we can see that a background of various self replicators is kind of the substratum, the environment where something emerges

"Nothing magical happened here." While I understand his point the fact that it happens at all IS the magic to me.

This is what I thought about when I was defragmenting my PC always wondered if the random blocks of data one day result in a self replicating data string that could catch enough data to amass into a bot of some kind like bacterial growth

Conway's Game of Life CAN actually self-replicate, but it has to be very carefully engineered.

Programs will stop evolving because randomization was applied only initially. Who knows what will happen if researchers figure out how to properly add randomness during whole simulation session.

There is a game from the late 80's or early 90's. The game is called SimLife and was published by a software/game company named Maxis. You may know them as the originators of the game The Sims.. before Electronic Arts bought them out. The manual for the game asked the question "What is life" and went on to describe life as "the process by which genes attempt to make copies of themselves" It's an older DOS-based game but worth looking into if you have the time and enjoy this subject.

Cool experiment! Would be nice to see if complexity does evolve eventually given enough space, time and Brainfuck 🙃

its consolidation. the informed changed. the reactive change. The scenario for self replication is mathematically defined. at some point the most prevalent process will emerge and replicate. if you take a bunch of jagged rocks put them together in a bag and shake the bag randomly at random speed and intensities at random intervals, eventually all the rocks will resemble one another in smoothness and maybe even size.