The Probability of Life2020 Mar 27
It so happens that we exist alive here on earth. We also find ourselves asking questions about the origins of life and our environment. How did this come to be? Could our life have occurred on earth due to natural random processes? If that were so, then probability would be involved; that is: probabilities would have made life possible. note However, is it reasonable to think our existence happened by chance?
Note that this is not an argument regarding Darwinian evolution. It is primarily a probabilistic argument regarding chemical abiogenesis which must occur before evolution can begin. The thought experiment will even attempt to help such a process by thinking of it as a search engine that is actively looking for life and for the information that life requires.
This will be an argument about genetic information which is then used to generate biological products. DNA which stores genes has no chemically preferred sequences; it is a pure information storage system. Therefore DNA itself does not modify the odds on the information that it carries. DNA functions like a book that can contain letters and words on its pages, or computer files that contain bytes in any sequence. This data-agnostic charactistic of DNA has been demonstrated by encoding text, pictures and video into DNA and then accurately decoding them back out again. Like that demonstration, the DNA of life contains complex purposeful biological information.
Where did the specified complex information that life requires come from? If it came from natural causes, then these had to be probable, or at least possible. Could the universe find the information of life in the random conjuctions of natural matter? If it was a search engine, could it run fast enough and have enough time to find it?
Probabilistic reasoning about pure information is especially appropriate. Information theory depends on reasoning just like this. The modern computer tech that we use all the time depends on probabilistic methods for successful information storage, communication and many calculations.
However, note that natural processes uniformly degrade information. Information systems that do not account and correct for these errors will fail. Self-replicating systems that survive therefore must require built-in error detection and correction mechanisms. The argument here will mostly ignore this issue and focus on the problem of finding and gaining useful information. However, without error handling the rate of information loss by random processes greatly exceeds all information gain. Therefore, this information retention problem strengthens the arguments here by a substantial amount.
By conclusion, if abiogenesis is unable to happen, then evolution never would have happened afterwards. Since we are alive, if we can infer that life itself could not have begun by natural processes then we must inevitably accept that it must have come to existence by a non-materialistic means.
Living things are complex information storage and usage systems. Information of this type according to all uniform human experience only comes into existence from an intellect. Specified complex information was needed before life in this universe could begin. Therefore, by logical reasoning, the most likely cause for this information would be from an intellect external to the natural material universe. If this is so, then it is also a reasonable idea that speciation may have been caused similarly by the same external agency. And this then is not Darwinian.
A useful way to think of this is to imagine the universe itself had always been searching for life. The chances of life occurring in this universe however are finite. What then are the probabilistic resources available to our universe that could account for whether it could discover life? note Was life inevitable, or was it implausible?
The multiverse has been suggested to make our existence be more likely. See The Multiverse God. I suggest however that my arguments will make even a multiverse be probabilistically insufficient and useless.
- It is estimated that the universe is 13.8 billion years old, which is about 10^18 seconds
- also written as 1e18
- That is a 1 with 18 zeros after it: 1,000,000,000,000,000,000
- It is spoken of as having 18 orders of magnitude. note
That is actually 13800000000 years x 360 days/yr x 24 hours/day x 60 minutes/hour x 60 seconds/minute =
4.29 * 10^17 or 4.29e17 or rounded up to 1e18
When we write 1e18, the 'e' marks the start of the exponent part of the number. The exponent basically is the number of digits (the zeros) after the first whole digit in the number. So, 1e18 is written out as 1 followed by 18 zeros, which is that 1,000,000,000,000,000,000.
The phrase “orders of magnitude” is another way of talking about this number of zeros. So, this number has 18 orders of magnitude.
Remember also: when multiplying numbers with exponents, the rule is that the exponent values are added. So 100 x 1000 = 100,000; which alternatively is 1e2 x 1e3 = 1e5.
- The smallest unit of time that exists is the Planck time interval, of which there are about 10^43 each second (1 with 43 zeros after, or 1e43). note
Wikipedia: Planck time and length
It is 1.86 x 10^43 = 1 / (5.39 × 10^44 seconds)
- There are an estimated 1e82 atoms in the known observable universe. note
Estimated from 1e78 to 1e82 atoms in the universe.
There are about 1e43 increments of time every second, and so there have been about 1e61 increments of time since the beginning of the universe. note That means that every atom has had no more than 1e61 individual moments through all of time. And cumulatively, all atoms in the known universe would have together had 1e143 moments. That is, there have been no more than this total of momentary events ever. If then every atom in the known universe had been involved continuously as fast as possible from the beginning in an active search for some combination of stuff that was alive, the whole universe would have had a total of about 1e143 search chances. note If the complexity of life is less than this number, then there is some probability that life would have accidentally occurred, and that life could have been found by the universe. note
Note that some people imagine that life on earth was seeded from elsewhere in the universe. (This is called panspermia.)
This number of 1e143 chances for life appearing in the universe already includes all the (supposedly additional) chances of life from panspermia. Panspermia is already accounted for. Panspermia does not help to solve the problem of life.
4.29e17 seconds x 1.86e43 intervals x 1e82 atoms
= 7.98 x 10^(18+42+82)
And rounded up to be generous, is 1e143 chances. This number includes all chances for life to appear over all time anywhere in the entire universe.
This life search engine is (best-case) assuming that new random conjunctions of matter that might be a new life are continually and rapidly occurring.
This thought experiment is also generously giving each moment the ability to self-recognize a successful life search candidate even though recognition like that is not in these momentary events. Actually, a separate recognition engine outside of the universe would be required for this because the evolutionary process of natural selection does not exist before life and/or self-replication. However to simplify things, I will assume that this comes for free.
Also, this is quite unrealistically generous for search chances for the universe because each of these chances is at the level of one atom. Many, many atoms together would actually be needed for any one living thing. So there should actually be less chances given. (At least divided by the number of atoms in a minimal living thing.) However, the difficulty of the search is so great that being this generous won't even weaken the argument here.
So actually there would be 4.29e17 seconds since the beginning of time. And actually there would be 1.86e43 Planck time intervals in a second.
Multiplied together these would be 7.98e60 which I am rounding up to a larger total of 1e61 time increments since the very beginning.
What does it take to make a living biological organism? There are many necessary things, only one of which is a collection of essential genes. Genes, roughly speaking, are sections of information in our genetic code that can convert to the making of a protein. There is no standard count of genes because the counts are different between types of organisms. However, by some estimates it might be that as few as 256 genes would be enough for a simple organism note. Every single one of these genes would be uniquely different, and the organism would need every last one of them to maintain life. This would mean that at minimum the organism would need 256 uniquely different functional proteins together at the same time to maintain life.
Wikipedia: Minimal genome
Note however that 256 is a theoretical minimum genome and no actual organism exists with a genome this small.
What are proteins? note They are biomolecules that perform functions in the cell. Proteins are the physical or chemical engines of the cell and participate in virtually every process in the cell. They start as a long chain of amino acids assembled in a specific sequence from the instructions in a gene. After a specific chain is assembled, the protein is folded into a specific three-dimensional shape that will enable it to perform functions.
Proteins come in different sizes, but are commonly several hundred amino acids long. The correctness of their assembly is critical to their functionality. We know this because mutations that change the sequence in a protein consistently result in broken functions. Are there lots of sequence combinations that turn out to be useful proteins? Or, are useful protein sequences rare? We will look at this question in two ways.
One study looking at this question constructed proteins of random sequence and tested them for a minimal function note. A simple test of functionality was used: they considered it functional if the random protein would attach itself to another important molecule called ATP. Of the initial random proteins, only 0.1% would bind to ATP. This means that this very primitive functionality existed in only 1 of 1000. (And all the 999 rest had no function.) If our organism requiring a minimum of 256 genes had proteins that were each only this rare, then the probability of randomly finding a complete set of genes that would make the necessary functional proteins would be at least (1/1000)^256 which is 1 out of 1e768. This unfortunately is 625 orders of magnitude too rare for the universe’s total probabilistic search resources. note.
Again, orders of magnitude is the size of the exponent in scientific notation. It is the size of the count of digits in the number.
The universe has 1e143 chances to find life. However, the likelihood of finding life by the measure here is 1 out of 1e768. (That is 1 followed by 768 zeros!) This gap is 768 - 143 = 625 orders of magnitude. (That is 1 followed by 625 zeros!) It is 1e625 times too improbable to have happened.
Another way of studying this is to see how many mutations a protein can withstand before it looses its intended functionality. A couple of studies using this method note came to the similar conclusion that the rarity of properly working functional proteins is on the order of 1 out of 1e77. note Our theoretical all-searching universe at best then has the probabilistic resources to only ever find one isolated functional protein. Two coordinated functional proteins, however, are 11 orders of magnitude too improbable to ever be found. note And we know that most of the elementary functions of life for all organisms happen in complex coordinated systems, requiring the coordination of more than just two proteins. note
Living systems are highly integrated, and important life processes depend on coordinated function. Said another way: life fails if the protein functions are not coordinated.
There are some functions are done by single protein engines. However, there are very many cases where multiple proteins coordinate to complete a function. Hemoglobin, which carries oxygen in our blood, is made of four protein units. A more dramatic example is the bacterial flagella which is a propulsion organelle composed of 20 different proteins; and it requires every one of them for correct function.
A requirement for two coordinated proteins multiplies the odds against it. 1e77 x 1e77 which is 1e77^2 = 1e154 Since the chances in the universe = 1e143, then 154-143 = 11 orders of magnitude implausible in our whole universe.
Other papers suggest proteins are not quite so rare at 1 in 1e40. This lesser rarity could be true for selected proteins because they come in a large range of lengths.
(This rest of this note will make more sense after development of the line of reasoning in a few more paragraphs.)
The details may be different, but the argument that follows will still hold because even this smaller value is still strikingly rare. The all-searching universe could at best find three coordinated proteins of this rarity 1e40 x 3 = 1e120 which is less than 1e143.
However, the rarity of the full genome would be 1e40^256 = 1e10240, which is 10240 - 143 = 10097 orders of magnitude implausible.
A Dentist in the Sahara: Doug Axe on the Rarity of Proteins Is Decisively Confirmed. “… only 1 in 1e77 sequences corresponds to a functional fold/domain within the protein…”
Estimating the prevalence of protein sequences adopting functional enzyme folds “… implies the overall prevalence of sequences performing a specific function by any domain-sized fold may be as low as 1 in 1e77…”
By this more realistic metric of protein rarity (1 out of 1e77), the complete set of 256 genes to make the required usefully functional proteins had 1 chance out of 1e19712 of occurring by accident. note Compare this to the total of 1e143 available search chances in our universe (19712 - 143). This is 19569 orders of magnitude too rare for the entire lifetime of the universe. It is beyond plausible to say that this minimal genome could have occurred through time by natural processes.
1e77^256 = 1e19712
Theorists on abiogenesis suggest instead that only a single (modest) self-replicating molecule needed initially to have happened by chance. From there, evolutionary processes would be able to take over to produce the type of life organisms that we see today. Some RNA strings are capable of both storing information in the molecule and catalyzing chemical processes similar to the work of proteins. Although RNA does neither thing well, some RNA is functional in this way by some small amount. Therefore, in the opinion of some theorists this is sufficient to demonstrate a possible path from chemicals to life. note
Chemical abiogenesis is hopeful self-deception because the chemistry is adverse to success - in an RNA world or in any other chemistry. There is poisoning of the reaction with by-products, lack of purification of the reaction products and lack of protection from destructive environments. Even if some researcher intelligently determines a candidate for a self-replicating molecule, this does not make abiogenesis be more likely because the adverse chemistry still dominates in all natural contexts. So this chemical approach ignores that chemistry always works to make such a process fail.
This is even acknowledged in the famous Miller-Urey experiment. The molecular products that they were able to generate had to be rescued from the source materials to avoid ruin. No such laboratory rescue would be available to any naturalistic abiogenesis.
In addition to the raw chemistry problems, unless this self-replicating RNA would have virtually perfect reproduction, its copies will very quickly degenerate and become non-viable. This is what lab experiments always show. This is a practical re-statement of the original problem.
Note that philosophically, even if an imagined chemistry of self-replicators could be achieved, they do not change the context because they have no instinct to survive like living things do. They will just as “happily” die (that is: fail) as they will live. They won’t try to live, so they won’t survive. The “will” of living things to continue living is a non-trivial component of life.
Machines that self-replicate have never been demonstrated in spite of concerted research and the encouragement of science fiction. note A machine that could self-replicate must be able to obtain all its raw materials by itself, and then duplicate all its parts, assemble them all, and duplicate all the information that it contains. This requires mastery of multiple disciplines, including the ability to fabricate multiple materials. This is incredibly difficult. It is no wonder that the only known self-replicating entities are living organisms. Much of the complexity of life is required to service replication. Self replication is a pinnacle feature of living things. Or, said another way: self replication demands the complexity of life.
As I heard Elon Musk say in an interview, a repeatable process to make a space rocket is an order of magnitude more difficult than making one rocket, and a machine (a factory) to make make rockets is again an order of magnitude more difficult. He was not even suggesting that the rocket was a self-replicator.
Living things can’t be wasting anything in the process of replication. Natural selection already assures this to be true (because wasted resources reduce fitness). Living things must already be at an evolutionary point of greatest simplicity. Any further simplification would reduce fitness; it would not be not competitively viable. If self-replicating living things cannot be simplified, this central tenet of evolution argues that no modest molecule could ever be a self-replicator.
Note also that the simplicity of a self-replicating machine will be inversely proportional to its ability to survive any modification. Said another way, the more simple the machine is, the less it is capable of surviving modifications. This can be seen in the competitions to write short self-replicating computer programs. The shorter programs will completely fail if even a single smallest bit anywhere is mutated. The same would be true of a simple replicating molecule. Changing any atom in it will stop it from replicating. It would be incapable of withstanding any modifications, therefore it would be incapable of evolving.
Evolutionary theory assumes that the entities involved can be incrementally changed from one form to another, but uniform repeated experience shows that does not work. You cannot change a good car into a good submarine by increments that consistently add functional advantage. note Fitness for one category of function is never bridged to a different category of function by a single step, and therefore it will not be by a path of gradually improving fitness. The path from one high point of fitness to another goes across deep valleys, and natural selection actually prevents any crossing. Fitness only exists at rugged isolated peaks of competence that are separated by incompetence, failure and death. The path from chemical self-replicator to life is guaranteed to pass through these wastelands. Evolution is incompetent to get through.
This was not meant to be an argument about the adequacy of evolutionary processes in living things. However since evolutionary processes are suggested to function in pre-living chemistry, the problems on the theory must be addressed. The model for transition of a car to submarine is analogous to the assumed evolutionary transition from land animal to whale. The non-living thought experiment (the car to submarine) shows the extreme unlikeliness of a successful path between the two such types in living entities and the very certain lack of fitness in the intermediaries.
Changes like this had to have occurred in any path from a first self-replicating molecule to LUCA (the Last Universal Common Ancestor). Therefore, the necessary path must go through states that are non-viable.
It is theorized that with a self-replicator the evolutionary processes will modify the context so that it is no longer limited by probability; it will then perform better than just by chance. However, this is a false hope because evolutionary processes are slower than chemical processes. Once evolution is involved there is both information processing time as well as chemical reaction time involved. In a proposed evolutionary path toward life, certainly it is critical to retain the information gained between generations. However, evolution must wait for mutations at the rate of generation cycles, and this is much slower than chemical reactions and it diminishes the possibility of success. Evolution becoming involved will actually make the number of chances become substantially smaller for the universe to successfully produce life. note
Even if a simple self-replicating molecule existed since the beginning of the universe, and even if it was trying combinations of improvements as fast as possible (at Planck intervals), it would (by definition of abiogenesis) still have to be able progress up to the complexity of life as we observe it. The same rarity argument from before still holds. There would still only be 1e143 opportunities to come up with anything in the known universe. However as previously pointed out, the rarity of life from an over-simplified view of proteins is 625 orders of magnitude too rare to be produced in our material world, and from a more realistic view of protein rarity, existing life is 19569 orders of magnitude too rare to exist.
In the thought experiment of a life-searching universe, the mechanism for recognizing successful searches was given for free without a mechanism. Here, evolution “taking over” is finally just taking on that necessary task. This means that the context has not changed for the better from the original thought experiment. In fact it is much worse because the for-free recognition mechanism was up to full Planck speed, but an evolutionary mechanism will have a recognition interval on the order of seconds to years and generally also it does not do a very good job of recognizing fitness.
A self-replicating molecule as a solution to the problem of abiogenesis is a profoundly naive self-deception.
Common Biological Toolkit
Next, evolutionary processes must have been capable to achieve the information that we know is contained in the common biological toolkit (CBT) which all living organisms share. note This particular information is so rare that there is only one singular toolkit for all known forms of life. This means the CBT code has never been able to evolve into something different since evolution’s last universal common ancestor (LUCA) at the very beginning of life. (Meaning all mutations away from the CBT have ended in death.) note This re-emphasizes the ruggedness of the fitness landscape. Natural selection is incapable of incrementally finding this toolkit information because it is evident that there is zero fitness for life as we know it without the complete CBT. This also means no fitness test could indicate incremental progress toward a functioning CBT. All fitness results would be a solid fail until suddenly (with the full CBT) everything is fit all at once for life.
Note that if the CBT was only something ordinary, then more than one variation or kind of toolkit should have evolved. That fact that there is only one CBT is evidence that the CBT is spectacularly special.
Essentially, that which I am calling a Common Biological Toolkit (CBT) is the functional basis for all living things. On evolutionary theory it must have existed since the last universal common ancestor (LUCA). It includes DNA expressed through RNA into proteins, metabolization of fats and sugars, ATP as an energy intermediate, a cell enclosed by a lipid membrane, ion transporters across the cell wall, cellular division, and etc.
Since it is complex, the CBT fills a substantial part of our theorized minimum genome of 256 genes. This re-argues that the minimum 256 genes must be correct all at once for life, and any supposed precursor that might be missing any part of the CBT is actually non-viable for life. However we have seen already the insane impossibility of this genome set.
The CBT itself also re-argues that any type of self-replicator needs complexity equivalent to the CBT. This argument definitely includes the chemical self-replicators. So it shows that no simple molecule could ever do this function. Insanely impossible complexity is again required right at the beginning for all replicators. note
The thoroughgoing materialist will say that life as we know it could have been different if time was run again from the beginning. They don’t have any attachment to life as it exists now. Instead they speculate materialistically that with a new run though time, life could have taken a different path and a different kind of life would have emerged. Therefore they let arguments based on the practical problems of abiogenesis or of evolution just roll off their back. This is because they see observable life as just an example of the life-developing process. They don’t take the view that that the process is defined by life as it does exist in our history.
Even if every mechanism they have proposed for explaining life development ends up being invalid because of an obstacle, that also is not perceived as a problem. These were just potential mechanisms that could be taken up or discarded, and discarding a mechanism is not perceived as evidence against abiogenesis or evolution. (See Evolution as Fact)
However, there is zero evidence for any kind of life other than what we know. A concept for a type of life other than what is observed is only speculation, and there is no scientific evidence for it. Therefore, such speculation is not scientifically based and it is metaphysical.
In fact due to the massive evidence for fine-tuning of the universe, in my thinking it is more likely that there simply are no other viable chemistries for life other than what we already know. Life depends on remarkable chemical characteristics of carbon, oxygen, and water and others. These were established already at (or before) the beginning of our universe. Therefore, there can be no naturalistic explanation for how these physical characteristics came to be so uniquely ideal for life.
If there is no other life other than what is known, and there can be no other life other than what is known, and life is unexplainable based on what is known of naturalistic processes, then most likely life was specially made here on earth. As the character Sherlock Holmes said, “How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?”
So, there is no naturalistic path that could proceed to life. Natural laws do include the laws of probability, and probability indicates that abiogenesis could not have happened. The requirements are beyond all best-case naturalistic options. If natural processes then are not capable of getting to evolution’s putative LUCA, then instead obviously an abiogenesis miracle is being claimed for “science”. note
A better approach is to allow the option of non-naturalistic explanations. Life clearly looks designed. The universe clearly looks optimized for life as we know it. Both of these would suggest a super-intellect that optimized and designed the universe and life.
- Note that a person allowing non-naturalistic explanations still has access to all the naturalistic explanations. They just have a bigger and a better explanatory toolbox than those that claim to only use naturalistic explanations.
- Note also that it is impossible to avoid metaphysical things in science. See Darwinian Metaphysics. Materialistic science depends on a bedrock of immaterial metaphysical assumptions. It is profound ignorance to misunderstand this.
Of course, “unlikely” things happen every day. However, there are different classes of unlikely.
It is unlikely that crystals would randomly form perfectly ordered shapes, forming as they do out of the chaos of fluids (think: salt cubes, and six-sided snow flakes). However, this is what they do because the shape is not a contingency. The automatic working of the molecular bonds of the material make it this way every time. It is a necessity and is not by chance, so it is not rare. note
We say that the shape of the crystals is governed by law. Therefore, they have this regularity as a matter of necessity.
As a second example, the color sequences of passing cars on a highway is not controlled by necessity. You expect a different chance sequence every day. The part that is unlikely is for you to be able to predict what the sequence would be on a day. Any one specific sequence would be very rare. note
The random output of this example is purely by chance. Sequences like this are sometimes called Shannon information. The content has a lot of detail, but it is not complex. They have high "information" content, but very little meaningful information content.
So, any one recorded sequence like this would be very rare. (That is, the sequence is very unlikely to repeat.)
Also, these sequences effectively are collections of noise. Sometimes it is important to get a good example of noise. If noise is what I want, then any one sequence is as good as any other. So, it isn't very valuable. It is rare to get a specific sequence, but every day you get a sequence.
Every day, therefore, we expect a unique sequence because it would be very unlikely to repeat the exact same sequence on successive days. This means we never can expect to get a sequence that fulfills a specification (other than for noise).
Also, it is unlikely that you would recognize any meaning in the random sequence of a day. Generally speaking, the value of a sequence is connected to its meaningfulness. Therefore, even though one of these specific random sequences are unlikely (are one of a kind), they are neither valuable nor significant.
A third example of unlikeliness is the sequence of all my experiences up to the present point. This sequence also was not of necessity, was not predicted, and it also has some portion of randomness in it. The unique collection did lead up to the present, but it didn’t construct the present. That is, if anyone attempted to repeat the steps, they wouldn’t recreate this moment. The sequence can only be seen in retrospect. This sequence of my experiences could be meaningful, but maybe only to me. It is not a sequence that is a useful procedure to create a future. note
Like the previous example, this example is not constrained by necessity. However, it is different this time because it has a complexity that the previous example did not have. The detail is not just random noise; it is not just Shannon information.
This sequence was not just purely chance; it does include considered choices. This is what makes it more meaningful. However, the value of this sequence is limited; it is not a specification for how to live life that could bring someone to a known state.
There is a final, special kind of unlikeliness. It is a sequence that has the power to predictably and reliably change the future, and it is very different from all the previous kinds. Consider the order in the process for making a house, or the sequence of events for launching a rocket. These are not sequences that could be expected to occur randomly. If these sequences are done wrong, then there will be a problem and sometimes the effort is a spectacular failure. If they are done correctly, we can be reasonably confident of success. note
Salt cubes have their shape as a matter of necessity. However, there is no necessity that some random process will be a sequence that makes things work. There are simply many more ways to make a process fail than there are to make it work. A sequence that makes things work is a functionally specified sequence.
It is easy to get some types of unique sequences (like the car colors). However, this fact is simply not an argument that it is therefore also easy to randomly get functionally specified sequences.
This example of unlikeliness is a different category from the others, valuable and very rare.
This final kind of sequence is not random data; it is purposeful specifications for making useful things or doing functional things. It would be meaningful to anyone making or doing these things. This is the kind of unlikely specification info that fills DNA and therefore, our genes. It is the kind that is required for life. note
This type of unlikeliness is associated to functionally specified information.
- Because it not constrained by necessity, it has contingency.
- It is complex, therefore minimally, it is not a product of chance.
- It fulfills specifications, therefore maximally, it is not a product of chance.
This three step method for eliminating chance and necessity was developed by Dr. William Dembski, and is called the Explanatory Filter. By inference, if something successfully passes through the three stages of the filter (is not by necessity, is complex, it fulfills specifications), then reliably the best explanation for it is that it is a product of design. (See Dembski, The Design Revolution, 2004) Pure information content of this type is consistently a product of a mind. And objects based on this type of information are consistently a product of a mind. Therefore, wherever such things exist, we have reliable certainty that it was produced by an intelligence.
Concluding that a collection of functionally specified information came from a mind is not an argument from ignorance. It is not a conclusion from a gap in our knowledge. It is making a logical inference from what we know very well by long experience. It is an inference from experimentally reproducible data.
Some of the sciences (such as archeology and forensics) are based on this reasoning because it enables us to distinguish between what would happen naturally, and what happened because of the action of some agent.
The scientifically respected SETI project (Search for Extra-Terrestrial Intelligence) is based on this being a good inference. It assumes that specified complex information is an indicator of intelligence and therefore reception of this type of information is evidence of it having come from an intelligent mind.
If we observe the existence of some object which is reliably made by a process, or we observe a complex task repeatedly being performed by a process, we intuitively interpret this to mean that an intelligence was involved to make the process. An ordered process to make or do the thing just wouldn't have happened by accident. The process came because of purposeful design. A house standing on some property and a rocket launching successfully into space are both artifacts of design. And in uniform repeated human experience, this kind of design only comes from an intellect.
All living things are filled with intricate co-dependent sequenced processes that construct and run the tiny machines of life. note This clearly looks intelligently designed and not random. Therefore we should apply this correct inference of design by an intelligence to the origin of life.
Note that this whole argument has ignored many other improbable things that are necessary for life. A world must exist for the life, with finely tuned parameters to make the physics, chemistry, temperature, radiation, energy, and everything else needed to make a favorable environment.
An organism living in this world needs more than genes. Even if a complete set of genes were to be found together by accident, they are insufficient for life because other highly complex systems and integrated components are needed at the same time and at the same place. (Examples: bi-layer lipid membrane, sugar code on membrane, homochirality, cytoskeleton, regulatory control networks, error correction systems, system for heritable epigenetics)
Note also that if I pull apart a living cell, I will have all the necessary parts all together at that time at one place, but there is no life. A living cell requires all the parts to also be in an integrated functional arrangement. However, outside their intact context, the components of life never self-assemble and integrate.
Genes are necessary for life to exist, but genes alone are insufficient. All the other necessary things are no less complicated or less rare than the genes. Together they show that the probability of naturalistic life origins is one of the lowest probabilities that could be known. Or said differently, the probability that life did not come from naturalistic origins is a thing that we can know with extremely high certainty.
In summary, the probability of life is insanely out of reach of the very-best-case probabilistic resources of the universe. To think that our existence happened by chance through natural materialistic processes does not stand up to reason. However, we do exist and we are alive, so therefore, there must be an alternative answer for the origin of life. I believe it is the most reasonable to allow that life came from a non-materialistic origin