Aubergine Society

In the second year of the Eggplant book club we’ll delve into Chapman’s source material. All are welcome to join at any time.

We will meet at 11am ET Saturdays on zoom

These zoom sessions will have no wait room, no host required, and will not be recorded. We’ll plan to spend one hour per meeting, but no strict time limits if participants want to continue.

Meeting schedule

Discord channel

2021.11.06 2.01

Unpacking The Meaning Crisis

We kicked off the new season with a discussion of the exchange between Vervaeke and Chapman. We generally agreed that it was worthwhile reading but it would have been more interesting if they got around to disagreeing about anything rather than just uncovering the depth of their parallel convergence. When speculating about where they might disagree @dglickman suggested the topics of neoplatonism and esoteric Christian theology, and @red_leaf was able to confirm with Chapman’s own recent quotes from the eG slack.

It seems hasn’t seen much traction since they were turned down by YC and subsequently acquired by substack (I think). I participated in a few exchanges myself, including one with Peter Limber on The Metagame.

We discussed the Joker movie since that was a theme Vervaeke brought up in the exchange. Like Chapman, I haven’t seen it, but unlike Chapman I’m not refusing to see it :slight_smile: . This led to a long digression on pop culture and the atomization of subcultures over the last couple decades, facilitated by the internet. I suggested the the atomized subcultures (microcultures?) have augmented subcultures rather than replacing them. For example, there are hundreds of subgenres of electronica and metal music now, but they are all still part of the larger subcultures, arranged in a hierarchy.

One of the newest subcultures I’ve personally joined is also one of the largest, namely the bitcoiners (only around for the last 10 years or so). The bitcoin subculture is part of a larger blockchain/crypto subculture with bitter rivalries between various factions, now morphing into a web3 subculture including DeFi, NFTs, and DAOs. Within the bitcoin subculture there is a faction of bitcoin maximalists (maxis), and within that there is a Christian faction, see the Thank God for Bitcoin book. Quite strange.

On the topic of art movies @red_leaf recommended:

We revisited the meaning of metamodern, differentiating the political variety (Hanzi-style) from cultural (a la, the braiding of irony and earnestness.

Apparently Valeria has extended family members that “actually talk to spirits”. I thought the phrasing was funny because it can be construed to mean “talk to actual spirits”, so we got into a discussion about what that could mean. Are the models of other people we hold in our heads spirits in a sense? If we continue conversing with our internal models of people after they die are we talking to their spirits?

I took this opportunity to offer my new definition of “supernatural” as anything that exists in the world of ideas, supervenient on physical reality but not itself physical. I pointed out that everything we currently consider to be in the category of supernatural, e.g. ghosts, gods, vampires, faeries, and the like, are also supernatural by my new definition.

Fodder for future reading selections:

2021.11.13 2.02

I started with a proposal for a new reading selection mechanism (in the schedule worksheet). The selection rules:

  1. Anyone may add an option (a title linked to the text under “reading”)
  2. Each option is voted on with a +1 (acceptable), a -1 (not acceptable) or 0 (neutral).
  3. Abstentions are counted as 0 (neutral).
  4. Each option is assigned a score equal to the sum of votes.
  5. The option with the highest score is selected and moved from the options to the schedule.
  6. If there is a tie for high score among the options, David has the right to decide which one is selected.

Note that when you add an option, it must be linked to the specific text to read (as per rule #1). Feel free to add a placeholder for later, but those won’t be selected until specified.

The discussion began with a critique of the article. I suggested that the initial agent model represented by Alexei the robot was a bit of a straw man, if that was being contrasted with the MIRI embedded agent approach. For example, what if Alexei was playing a chess game against a human opponent, then there is no way it could represent the game in its internal model. Initally I thought the embedded agent model was being presented as an improved (as in more powerful) approach to AI, but thanks to input from Evan and Daniel I revised my perspective substantially. Embedded AI is (obviously) not more “powerful” than AIXI (for example) but more constrained and realistic. That makes a lot more sense in retrospect.

The discussion turned to comparing the actual results of DeepMind vs MIRI, and recent restructuring at MIRI. I was about to bring up the Logical Induction paper but Daniel beat me to it.

There was some disagreement about whether societies has a lot of knowledge of the world before the invention of math. No doubt humans can get along fine without math, building houses and getting enough to eat (like beavers, I suggested), but it surprised me that it wasn’t obvious to others that our knowledge of the world has exploded since we’ve been developing mathematical (i.e. scientific) models.

Evan recommends:

(added to the list of options)

I mentioned I had a related bit string theory proposing that what we perceive as the world is embedded in a recursive function. Daniel dissented but Evan agrees were are likely living in a Tegmark level-4 multiverse.

Evan mentioned a theory of suicidality as a eusocial behavior and I related that to how soldiers essentially sacrifice their lives for the collective good. I mentioned that I observed that I am significantly more individualistic than average (I see collectivism as the practice of farming other people, to use Evan’s phrase), which brought us back to ob. Kegan stages. (h/t Valeria)

Vassar’s theory was revisited, that Kegan stages were tied to trauma levels and that stage 4 and 3 were reversed in that 4 implied more trauma. We discussed definitions of trauma and I suggested that trauma was a defense mechanism, drawing on a book I listened to last week about how IFS therapy was used to heal trauma:


2021.11.20 2.03

How to Make Our Ideas Clear by C. S. Peirce
Popular Science Monthly 12 (January 1878), 286-302.

It was quite interesting to pivot to an article from the American civil war era. For example, when talking about his definition of “doubt” and “belief”, Peirce illustrated with this example:

If, for instance, in a horse-car, I pull out my purse and find a five-cent nickel and five coppers, I decide, while my hand is going to the purse, in which way I will pay my fare.

I had to look up what a “horse-car” was on wikipedia…

The 3 of us had quite an animated discussion, in roughly 3 acts. Act 1 was all about the nature of counter-factuals. I suggested that when we are talking about counterfactuals, we are referencing different Everettian timelines. @dglickman was very quick to object, a theme would return to in Act 3.

We agreed that technically counterfactuals shouldn’t work at all since logically anything follows from a false premise (in other words, a counterfactual). But we do use them extensively and they are meaningful, so what is going on?

I suggested that counterfactual situations have a distance from the factual situation we observe. To illustrate I mentioned two fictional worlds depicted in film:

In some sense, the 2nd is closer to reality because we share a history before 1939, the beginning of world war 2. It is at least physically possible, as opposed to Reign of Fires that includes real dragons flying around battling modern helicopters. Dragons could not evolve on earth (though I concede we did have very large flying reptiles a 100 million years ago on Earth, and it may be possible that an alien lifeform on a lower gravity planet could look a lot like what we would call a dragon. Not sure about breathing fire, maybe an acid attack like the bombadier beetle’s defense, except it ignites in the alien atmosphere?

Counterfactuals are imaginary. I recalled a message I posted to the neurophilosophy channel on The Bridge discord arguing that imagination is an extension of memory. Once a species evolves memory (which are imaginary in the sense that the experience is no longer true), the next step is to take control of the memory models and tweak parameters to generate other possibilities:

A hierarchy of powers. No choice without imagination. No imagination without memory. No memory without awareness. Make sense?


Act 2 was about the nature of belief. I argued that beliefs were only part of agent models and used to explain and predict conditional behavior, which seemed to be much in line with Peirce’s definition, especially the third part:

And what, then, is belief? It is the demi-cadence which closes a musical phrase in the symphony of our intellectual life. We have seen that it has just three properties: First, it is something that we are aware of; second, it appeases the irritation of doubt; and, third, it involves the establishment in our nature of a rule of action, or, say for short, a habit.

I produced a list I’ve collected of words I equate with belief:

  • bias
  • desire
  • preference
  • disposition
  • impetus
  • predilection
  • proclivity
  • propensity
  • penchant
  • inclination
  • predisposition
  • prone
  • leaning
  • tendency

The common theme here is the agent will tend to behave in some way under some conditions.

Act 3 was about the nature of probability and was the most contentious. I suggested that probabilities can be equated with measures over Everettian timelines. Like for a fair coin flip, 50% of the timelines (more or less) would contain a version of me that observed heads and the other 50% would contain a version of me that observed tails. The fact the Daniel objected strongly to this characterization took me by surprise since I thought it was obvious and non-controversial. My understanding of Daniels position is he thought that 50/50 split in timelines would only apply to quantum level observations like observing the spin of a single particle, and the classical events like a coin flip would have all (or almost all?) timelines collect in the observation of heads or tails.

We tried to find the crux of our disagreement with various modifications of the experiment, such as a biased coin, a double-headed coin, and a robot that flipped the coin in very controlled circumstances to fix the outcome. In all cases we would bet the same way, so were unable to distinguish our disagreement despite some fairly heated discussion :grinning_face_with_smiling_eyes:

For better or worse, I was called away to work before we reached any resolutions. To be continued?

2021.11.27 2.04

After a brief discussion of Pinker’s new book on Rationality (which I’m currently enjoying) and Julia Galef’s recent book The Scout Mindset (which I recommend, though @dglickman didn’t finish it) we got into Jessica Taylor’s article on counterfactuals and how acceptable (or not) it is to entertain logical inconsistencies. @Valeria’s description of how we only observe projections of thin slices of reality, and how two projections of the same thing from different perspectives can look inconsistent on a first pass reminded me of the cover art from GEB:


I would still argue here that reality is always internally consistent, and any apparent inconsistencies could be resolved at least in theory, with a deeper understanding.

I was aware that there were many competing decision theories, namely:

I speculated that they were invented to resolve decision paradoxes like Newcomb’s, the smoking lesion, and Parfit’s hitchhiker. Daniel ventured that the paradoxes were more likely invented to illustrate the different approaches to decision theory.

Somehow we got on to a long tangent about how our current technology stack in computing is a terrible mess (largely due to leaky abstractions) and we are in dire need of a complete rewrite (like urbit is attempting but not necessarily the way urbit is attempting).

Other topics that came up:

Next week: