Episode 2: Flow, Metaphor, and the Axial Revolution
Last time we were talking about what was going on in shamanism and the Upper Paleolithic transition. We talked a lot about the flow experience and how it integrates altered states of consciousness, on a continuum with mystical experiences and meaning making, enhanced insight and intuition, and how this resulted in an enhanced capacity for metaphorical cognition which greatly expands human cognition, makes it much more creative, much more capable of generating all of those fantastic connection in meaning that drove the Upper Paleolithic transition’s explosion in culture and technology.
Then we moved to consider some other intervening revolutions that also had an impact. We briefly talked about the Neolithic revolution and the beginning of agriculture and then the rise of civilizations. We got into the Bronze Age civilization and then that led us into the revolution we’re concentrating on now, which is the Axial Revolution, a period around between 800 BCE and 300 BCE following the Bronze Age Collapse.
The Bronze Age Collapse was one of the greatest if not the greatest collapse in civilization the world has ever seen. That facilitated much more experimentation in smaller scale societies and that experimentation resulted in the creation of new psychotechnologies. One was alphabetic literacy happening in the area of Canaan, and it’s eventually going to be taken up very quickly by the Hebrews and then taken up by the Phoenicians and taken to the Greeks.
The Greeks further improved it; we talked about how that psychotechnology, alphabetic literacy, makes literacy more effective, more efficiently learned, more powerful. Its operation greatly expands the number of people that can be literate, enhances the distributed cognition and how that psychotechnology gets internalized into our metacognition and produces second-order thought. We get an enhanced awareness of our own cognition, both its power and its peril. We get an enhanced awareness of its capacity for self-correction and self-transcendence. We also get an enhanced awareness of its capacity for self-deception.
We talked about the invention of coinage to help deal with the mobile armies of this time and how that trains you in abstract symbolic thought and more rigorous mathematical reasoning and that also gets internalized. It gets exapted right into second order thinking and people start to become aware of themselves in a different way. They start to become much more aware of the meaning-making nature of their cognition, its capacity to generate illusion and self-deception and also its capacity to break out of illusion and self-deception and to come into contact with a more real world.
This leads to some fundamental changes; people start to become more aware of their responsibility for the violence and the chaos and the suffering in their own lives and they start to become aware of how much the transformation of mind and heart (in the Axial age, often referred to in a singular manner) is the way to alleviate suffering.
One of the things I really like about this series is the way in which cognition is viewed as this double-edged sword, where it is specifically the things that make it good that also make it bad. The ability to quickly reach conclusions is both what makes intelligence useful—you need less sensory data / less time to decide things—and what makes it problematic—you jump to incorrect conclusions more quickly as well. This is, of course, also my view on AI alignment: the problem is not that people build robots and then foolishly decide to put guns on the robots. The problem is that we only know how to make the first-order cognition, where we know how to make optimizers that search across a wide possibility space for things that maximize some score, with no attention on whether or not they have the right score function. So the robots we build now are very susceptible to illusion and self-deception.
This also feels very tied to the spirit behind Less Wrong: intelligence and rationality are distinct things, where rationality is mostly focusing on the ways in which you personally are subject to illusion and self-deception, and need to rearrange your thinking such that your intelligence is helping you instead of an obstacle.
I didn’t understand the connection he was drawing between causal modelling and flow.
It sounded like he was really down on learning mere correlations, but in nature knowing correlations seems pretty good for being able to make predictions about the world. If you know that purple berries are more likely to be poisonous than red berries, you can start extracting value without needing to understand what the causal connection between being purple and being poisonous is.
I didn’t understand why he thought his conditions for flow (clear information, quick feedback, errors matter) were specifically conducive to making causal models, or distinguishing correlation from causation. Did anyone understand this? He didn’t elaborate at all.
It sounded like he was really down on learning mere correlations, but in nature knowing correlations seems pretty good for being able to make predictions about the world.
This also shows up in Pearl; I think humans are in a weird situation where they have very simple intuitive machinery for thinking about causation, and very simple formal machinery for thinking about correlation, and so the constant struggle when talking about them is keeping the two distinct.
Like, there’s a correlation between purple berries and feeling ill, and there’s also a correlation between vomiting and feeling ill. Intuitive causal reasoning is the thing that makes you think about “berries → illness” instead of “vomiting <-> illness”.
Did anyone understand this? He didn’t elaborate at all.
Try flipping each of the conditions.
Information that is obscure or noisy instead of clear makes it harder to determine causes, because the similarities and differences between things are obscured. If the berries are black and white, it’s very easy to notice relationships; if the berries are #f5429e and #f54242, you might misclassify a bunch of the berries, polluting your dataset.
Feedback that’s slow means you can’t easily confirm or disconfirm hypotheses. If eating one black berry makes you immediately ill, then once you come across that hypothesis you can do a few simple checks. If eating one black berry makes you ill 8-48 hours later, then it’ll be hard to tell whether it was the black berry or something else you ate over that window. If you ate a dozen different things, you now have to run a dozen different (long!) experiments.
If errors are irrelevant, then you’re just going to ignore the information and not end up making any models related to it. The more relevant the errors are, the more of your mental energy you can recruit to modeling the situation.
Why those three, and not others? Idk, this is probably just directed sourced from the literature on flow, where they likely have experiments that look into varying these different conditions and trying out others.
I was thinking that there were groudns to think that flow is an experience of lots of implicit learing but I was much more lost on why flow would be conductive to more. Like if I have a proof streak then there is going to be more fodder for more and more proofs but most of that is going to be irrelevant calculation and dead-ends that don’t lead to theorems. And there is no guarantee of success. At some point what is getting and enabling me the results is going to run out. Success doesn’t by itself generate success.
Episode 2: Flow, Metaphor, and the Axial Revolution
One of the things I really like about this series is the way in which cognition is viewed as this double-edged sword, where it is specifically the things that make it good that also make it bad. The ability to quickly reach conclusions is both what makes intelligence useful—you need less sensory data / less time to decide things—and what makes it problematic—you jump to incorrect conclusions more quickly as well. This is, of course, also my view on AI alignment: the problem is not that people build robots and then foolishly decide to put guns on the robots. The problem is that we only know how to make the first-order cognition, where we know how to make optimizers that search across a wide possibility space for things that maximize some score, with no attention on whether or not they have the right score function. So the robots we build now are very susceptible to illusion and self-deception.
This also feels very tied to the spirit behind Less Wrong: intelligence and rationality are distinct things, where rationality is mostly focusing on the ways in which you personally are subject to illusion and self-deception, and need to rearrange your thinking such that your intelligence is helping you instead of an obstacle.
I didn’t understand the connection he was drawing between causal modelling and flow.
It sounded like he was really down on learning mere correlations, but in nature knowing correlations seems pretty good for being able to make predictions about the world. If you know that purple berries are more likely to be poisonous than red berries, you can start extracting value without needing to understand what the causal connection between being purple and being poisonous is.
I didn’t understand why he thought his conditions for flow (clear information, quick feedback, errors matter) were specifically conducive to making causal models, or distinguishing correlation from causation. Did anyone understand this? He didn’t elaborate at all.
This also shows up in Pearl; I think humans are in a weird situation where they have very simple intuitive machinery for thinking about causation, and very simple formal machinery for thinking about correlation, and so the constant struggle when talking about them is keeping the two distinct.
Like, there’s a correlation between purple berries and feeling ill, and there’s also a correlation between vomiting and feeling ill. Intuitive causal reasoning is the thing that makes you think about “berries → illness” instead of “vomiting <-> illness”.
Try flipping each of the conditions.
Information that is obscure or noisy instead of clear makes it harder to determine causes, because the similarities and differences between things are obscured. If the berries are black and white, it’s very easy to notice relationships; if the berries are #f5429e and #f54242, you might misclassify a bunch of the berries, polluting your dataset.
Feedback that’s slow means you can’t easily confirm or disconfirm hypotheses. If eating one black berry makes you immediately ill, then once you come across that hypothesis you can do a few simple checks. If eating one black berry makes you ill 8-48 hours later, then it’ll be hard to tell whether it was the black berry or something else you ate over that window. If you ate a dozen different things, you now have to run a dozen different (long!) experiments.
If errors are irrelevant, then you’re just going to ignore the information and not end up making any models related to it. The more relevant the errors are, the more of your mental energy you can recruit to modeling the situation.
Why those three, and not others? Idk, this is probably just directed sourced from the literature on flow, where they likely have experiments that look into varying these different conditions and trying out others.
I was thinking that there were groudns to think that flow is an experience of lots of implicit learing but I was much more lost on why flow would be conductive to more. Like if I have a proof streak then there is going to be more fodder for more and more proofs but most of that is going to be irrelevant calculation and dead-ends that don’t lead to theorems. And there is no guarantee of success. At some point what is getting and enabling me the results is going to run out. Success doesn’t by itself generate success.
Assyrian Armies of the Axial-Age: Alphabetical, Arithmetic, and Affluent.