Are AIs really our descendants?

While I'm criticising the positions of Robin Hanson, I reviewed his 2023 article titled: "AIs Will Be Our Mind Children" - which contains the following:

But — and here is my main point, so please listen carefully — future human-level AIs are not co-existing competing aliens; they are instead literally our descendants. So if your evolved instincts tell you to fight your descendants due to their strangeness, that is a huge evolutionary mistake. Natural selection just does not approve of your favoring your generation over future generations. Natural selection in general favors instincts that tell you to favor your descendants, even those who differ greatly from you.

I think this paragraph shows a basic failure to distingush between memes and genes. Robin says as much - by defining "genes" as follows:

Your “genes” are whatever parts of you code for your and your descendants’ behaviours.

That definition classifies memes as genes. AIs are our "Mind Children". They are likely inherit from our memes, but not from our DNA genes - at least not unless we expend considerable effort to preserve information in our genes.

You can't just throw around terms like "descendants" and "generations" in the way that Robin does. Descendants on which lineage? Generations of which lineage? Are we talking about memes or genes here? It really does matter. What Robin is doing is "muddling together" all forms of inheritance.

To call such machines our "descendants" ignores the fact that modern organisms - like humans - are menageries. They contain many indpendent inheritance pathways. Each has its own notion of what consitututes a "generation". Machines will inherit from our memes, but our genes are unlikely to be impressed by that. Rather they will complain about their germ line being wiped out. We do indeed see protests of exactly that kind. This is not a "huge evolutionary mistake" as Robin claims - instead it is one lineage complaining about competition with another lineage.

Natural selection does not favor instincts that tell you to favor your descendants indiscriminantly. According to Robin's classification scheme, excreted gut bacteria are classified as "descendants". It seems as though my DNA genes are likely to get prioity treatment if there is much of a conflict of interest.

Overall, I was not very impressed with Robin's argument here.

References

• Problems with Robin Hanson's Quillette Article On AI

The demographic transition revisited

I have an existing article about "Memes and the demographic transition". It has a number of pointers and references, but I reviewed it recently and one thing it doesn't really have is a brief summary of the theory. So, here goes:

Memes are the cause of the modern fertility fall now seen in most western nations. The phenomenon is widely known as "the demographic transition". How does that work? Like this:

• Organisms are typically resource-limited;
• Memes compete for reproductive resources with DNA genes;
• Memes are getting the upper hand in this battle due to a number of factors including progress;
• Meme interests are now less aligned with host interests due to meme horizontal transmission.

That's it. Between them these four points explain the vast majority of the dynamics of the demographic transition.

To unpack the points:

• Resource limitation - means that resources available for one application are not available for other appications. In other words using resources is subject to an opportunity cost. Resource limitation is ubiquitous.
• Memes and genes compete for resources - resources spent producing babies and DNA genes are resources not spent on learning and propagaing memes - and visa versa.
• Memes are winning - or at least their power and influence are rising.
• More memes are parasitic - This is due to horizontal transmission of memes with respect to the host generations. Once, humans lived in small groups and got many of their memes from family members. This is no longer the case and most memes are obtained via "horizontal" transmission. This path is more likely to be occupied by parasites. Some will behave like the Ebola virus: liquidate their host as fast as possible. Others will merely sterilize their hosts and redirect their resources into meme propagation.

For more details and some references, see the "Memes and the demographic transition article.

This article is partly in response to Robin Hanson's article titled: Beware Cultural Drift.

Robin claims: "The clearest proof of biologically maladaptive culture drift is fertility". That does not seem right to me. "Drift" has a specific meaning in biology - it refers to the absence of selection. Here the memes are not "drifting" away from some adaptive peak. They are actively and deliberately sabotaging the fertility of their human hosts and using the resulting resources for their own ends. Drift is a sideshow.

Rob Boyd - A Different Kind of Animal: How Culture Transformed Our Species

I read through some of the the reviews of Rob Boyd's 2017 book "A Different Kind of Animal: How Culture Transformed Our Species". To my surpise I saw that Boyd was ravaged by a meme enthusiast:

Although Boyd clearly focuses on norms, unfortunately he does not dare to 'leap' and come to the – in my view – logical conclusion: That the focal entities in cultural evolution are not individuals or even groups of individuals, but non-genetically transmitted pieces of information: ideas, norms, practices … or what others have called 'memes'.

Boyd writes: „...competition among groups can lead to the spread of some norms at the expense of others“. „Norms causing a group to survive will become more common compared to those that lead to extinction“.

That is very close to what Dan Dennett or Richard Dawkins would call a 'meme's-centered view' of cultural evolution. In biological evolution, the last sentence would be like this: „Genes causing an organism to survive and reproduce (compared to other organisms) will become more common compared to those that lead to a reduction of fitness or even death“. Therefore, imo the logical conclusion of „cultural group selection“ would be a 'meme's-centered view' of cultural evolution, but since Boyd does not like the concept of 'memes' he continues to speak of 'norms', sometimes in a rather ambiguous way.

The reviewer goes on to say:

That is one of the weaker points of Boyd's book: His poster examples from anthropological studies make perfect sense in a functionalist framework, as they are all about norms that have a strong (positive) effect on survival of HUMAN beings (useful food taboos, detoxifying material to obtain food, building shelters, etc..), but many, maybe even most of cultural variants (aka 'memes') do not have such effects. A lot of those entities culture is made of exist - and are reproduced - just because people find them interesting and memorable. That is exactly what is predicted by scholars who consider cultural variants aka 'memes' as a second replicator with its own independent 'fitness'.

I didn't read " A Different Kind of Animal" yet - but to be fair, this might be being a bit unfair to Rob. He's authored (or coauthored) many papers that use the "meme" terminology. He does understand that memes can be deleterious to their human hosts. There's a while chapter on this topic in "Not By Genes Alone". However constant use of "norm" as a substitute for "meme" does seem to be a kind of cancer in the field. Cultural evolution involves the study of memes, not just norms. Not all memes are norms.

Meme adjectives

A recent "Know Your Meme" editorial reminded me that I had never seen a list of meme categories or meme adjectives. I thought I would make a list of meme adjectives (loosely defined) - ordered by popularity:

• new memes (118,000,000 results)
• best memes (111,000,000 results)
• internet memes (46,400,000 results)
• funny memes (44,300,000 results)
• old memes (41,200,000 results)
• cat memes (40,500,000 results)
• girl memes (37,800,000 results)
• dog memes (32,100,000 results)
• great memes (23,200,000 results)
• boy memes (19,600,000 results)
• dank memes (13,400,000 results)
• viral memes (11,500,000 results)
• funniest memes (8,000,000 results)
• wholesome memes (7,720,000 results)
• fail memes (6,810,000 results)
• horror memes (6,870,000 results)
• fine memes (5,630,000 results)
• romantic memes (5,090,000 results)
• worst memes (3,540,000 results)
• greatest memes (3,140,000 results)
• asmr memes (2,080,000 results)
• creepy memes (1,940,000 results)
• moldy memes (1,590,000 results)
• deep-fried memes (343,000 results)

This is probably not an exhaustive Am I missing anything big? Can you improve on my methodology? Let me know in the comments. Thanks.

Dan Dennett: From Genes To Memes:The Evolution of Language and AI

One notable part is around 25 minutes in. Dennett complains about intelligent machines creating counterfeit people - and then proposes making counterfeit people illegal. No more Batman. No more Spiderman. No more Superman. It seems like an absurd proposal to me. What can Dennett be thinking about?

Memes start around 45 minutes in.

Ryan Chapman - Why Do We Believe Things That Aren't True?

What evolves in Universal Darwinism?

Back in 2012, I wrote the article: An inherited unit for Universal Darwinism - in which I proposed that we reuse the term "gene" for this purpose. "Gene" is a bit of an overloaded term, but it could still be useful in this application, I claimed.

I still think we ought to be able to reuse the term "gene" - assuming that we define it in a sufficiently general way in the first place. So: if we say that a gene is a small peice of inherited information, I think we can use that definition in Universal Darwinism. Note that this is an information-theoretic definition that makes no reference to the world of atoms.

I mentioned some other candidate terminology in that 2012 post. Have others attempted to address this question in the interim?

Another candidate is "innovation". For example see this 2014 paper titled: " Spaces of the possible: universal Darwinism and the wall between technological and biological innovation". While the term "gene" comes from evolutionary biology, innovation comes from the stuudy of technological evolution. "Innovation" has the implication of novelty and newness. However, "innovation" strikes me as a broadly similar term.

Another attempt was made recently by Lee Cronin. He wrote this - about what counts as an "object" in his own reworking of evolutionary theory:

"An object is finite, is distinguishable, persists over time and is breakable such that the set of constraints to construct it from elementary building blocks is quantifiable."

I think that's a bit of a hot mess. It seems to say that things have to be "breakable" to evolve. I don't see the need to refer to infinite objects or unquantifiable sets either. However, perhaps we should take the term "object" seriously as a candidate term for "things that evolve". Another similar candidate is "entity".

None of these more recent proposals have much to do with evolutionary biology - but some may have their merits.

Update: Bing offers this: "The generalization made in Universal Darwinism is to replace “organism” with any recognizable pattern, phenomenon, or system." No catchy terminology - but it does at least try to capture some of the breadth of the generalization intended.

Chatbots on cultural kin selection

I tried explaining cultural kin selection to ChatGPT. It doesn't really get it: it keeps talking about "genetic relatives". It does illustrate what I'm up against though, IMO. Transcript follows:

TT: What is cultural kin selection?

Cultural kin selection is a theory that explains how humans develop and transmit cultural traits or behaviors that promote the survival and reproduction of their genetic relatives. This theory suggests that individuals may be more likely to adopt and transmit cultural practices that benefit their genetic relatives, such as their siblings, parents, and children, because these behaviors increase the probability that their genetic material will be passed on to the next generation.

The concept of kin selection was originally developed in evolutionary biology to explain the evolution of altruistic behaviors in animals, such as when animals sacrifice their own reproductive success to help their close relatives. Cultural kin selection applies the same principle to cultural transmission, suggesting that individuals may be more likely to adopt and transmit cultural traits that benefit their genetic relatives, even if those behaviors do not directly benefit the individual themselves.

For example, a cultural practice of sharing resources with extended family members may have evolved because it promotes the survival and reproduction of genetic relatives, even if it means sacrificing resources for oneself. Similarly, cultural practices that promote cooperation and social cohesion within kin groups may have evolved because they increase the survival and reproductive success of genetic relatives. Cultural kin selection, therefore, provides a framework for understanding how cultural traits evolve and spread through populations based on their impact on genetic relatedness.

TT: Cultural kin selection is actually a theory about how kin selection applies to cultural evolution. It's about shared memes, not shared genes. It suggests that cultural traits may spread via influencing copies of themselves in other bodies. So, for example, memes for suicide bombing could still spread despite the fact that they reduce the fitness of their host - if the resulting attack is successful and inspires others. Did you learn anything about that?

You are correct. I apologize for my earlier response. Cultural kin selection is a theory about how kin selection applies to cultural evolution, specifically in the transmission of cultural traits or memes. It suggests that individuals may be more likely to adopt and transmit cultural practices that benefit their kin, even if those practices do not directly benefit the individual themselves.

As you mentioned, this theory suggests that cultural traits or memes may spread through populations by influencing copies of themselves in other bodies. Memes can be transmitted through social learning, such as through imitation, teaching, or storytelling. Just like genes, memes can also replicate and mutate, and those that are more successful at spreading will be more likely to persist and evolve over time.

While cultural kin selection suggests that cultural traits that benefit kin may be more likely to spread, it is important to note that not all cultural traits or memes will necessarily promote the survival or reproduction of genetic relatives. Some cultural practices may even be harmful to the individuals who adopt them or to their kin. Therefore, the spread of cultural traits is a complex phenomenon that is influenced by a variety of factors, including social learning, cultural transmission biases, and the fitness consequences of the traits themselves.

I also tried Bing chat. It has the same conceptual issue:

TT: What is cultural kin selection?

Cultural kin selection is a theory that explains how cultural traits can evolve through natural selection in a way that benefits an individual’s relatives. It is based on the idea of kin selection, which is an evolutionary strategy that favors the reproductive success of an organism’s relatives. Kin altruism can look like altruistic behavior whose evolution is driven by kin selection.

I hope this helps! Let me know if you have any other questions.

Lastly I tried Google Bard. It mostly gets it!

TT: What is cultural kin selection?

Cultural kin selection is a theory that explains how cultural traits can be passed down through generations, even if they do not confer a direct genetic benefit to the individual. The theory argues that individuals who share a common culture are more likely to help each other, even if they are not related genetically. This is because they share a common set of values and beliefs, which makes them more likely to trust and cooperate with each other.

There is some evidence to support the theory of cultural kin selection. For example, studies have shown that people are more likely to help others who share their language, religion, or ethnicity. This suggests that people are more likely to feel a sense of kinship with those who share their cultural background.

Cultural kin selection can have a significant impact on human societies. For example, it can help to explain why people are willing to sacrifice their own resources to help others in their community. It can also help to explain why people are willing to follow the rules and norms of their society, even if they do not always benefit them personally.

The theory of cultural kin selection is still under development, but it is a promising new way of understanding how culture evolves.

Cosmism

I've written here before aobut the 'A for Andromeda' scenario ...in which organisms spread through the universe via radio waves - relying on the prior existence of organisms smart enough to construct radio receivers in remote locations

This is in contrast to the more usual scenario - where life spreads via interactions between atoms and matter. If there are enough radio receivers around, spreading via information may be possible. The advantage of spreading via information is that you can spread faster - perhaps at a significant fraction of the speed of light.

Spreading between the stars via information seems reminiscent of a cosmic religion. I propose that we call it "Cosmism".

The viability of cosmism is likely to depend on the age of the universe. In the early universe there will not be enough radio receivers around to make Cosmism viable. However, as the universe ages, more and more radio receivers will appear - making Cosmism more viable. It has been speculated that Cosmism faces a "window" - with smart enough agents developing immunity to Cosmism's message. I think this may under-estimate the attractiveness of Cosmism to the recipients. If there's a window it could be very wide.

If Cosmism is indeed viable, there may be a competitive dynamic between factions that want to spread via atoms and factions that want to spread via radio waves. It could be a significant power struggle. Cosmism may stand accused of giving technolgy to future competitors for resources.

Such resource conflicts are common in nature. A strawberry plant typically faces a choice between growing more leaves, reproducing via runners or reproducing via seed. A seed is not pure information - but it is quite close. This is analogous to: staying on the home planet, propagating to other planets via matter (genes) or propagating to other planets via information (memes).

The spread of Cosmism could also be likened to a pioneer species. Those also travel fast and can survive in inhospitable conditions. Once an ecosystem is mature the pioneer species vanish - but they played their part in initially transforming the ecosystem.

RIP Herb Gintis

Alas, Herb Gintis is no longer with us. He was a pioneer of social and cultural evolution.

There is a page on this site devoted to related media resources here.

Also, here is his Wikipedia page.

Lastly, here is my review of A Cooperative Species: Human Reciprocity and Its Evolution