Hello there, friend!

In Activity theory, the concept of mediation is central – it means that objects are not simply "out there" to be perceived and interacted with. Instead, our relationship with them is always simultaneously instrumentally and socially mediated.

Social mediation is about us always being embedded in various social structures, which always shapes our options. This has several implications, the chief of which is the "primary contradiction of capitalism" between the use-value and exchange-value of the outcomes of any productive activity – we are always balancing between doing something for its own sake, because its right, and doing it so that we can exchange the results for something else later, i.e. doing it to increase our income or status or something like that. This obviously has many consequences, but exploring those is a topic for a different week. Instead, I want to look a bit closer at the other way of mediation.

Instrumental mediation is about us using various tools to interact with the object, which shapes what the object actually means to us, since the tools shape what we are able to do with the object. However, our most prominent tool isn't material – it's our language, and through it various concepts that we use to... well, conceptualize the objects.

I understand the importance of this in three ways. First, we perceive the world through our mental models, meaning that we usually end up finding what we expect. This is probably the most "well d'uh" point, but I think most people (me naturally included) systematically underestimate its importance. All knowledge work is increasingly moving towards the state where the key challenge is not solving a problem, but defining it. But all our pre-conceptions about the problem are going to shape what we see.

Second, the concepts and others cultural artifacts have their own cultural-historical context. That means that at some point, someone created them... but because the mediating relationship goes both ways, the tools then created their creators. In this way, the creators may have developed far beyond the context that first led to the creation of the artifacts. But because the artifacts are externalized, they endure, and they embody the earlier conditions, which in the present conditions might be creating problems. This means that studying the instruments and their history can help us understand why people and organizations that use them act the way they do.

Third, the concepts are static. They are re-presentations, they are not the thing itself. But the thing itself, the object – that's living reality, it's dynamic and cannot be captured by any concept. The best the concepts can aspire to is capture a partial truth of the object. But it's never going to be complete, and it's likely going to become obsolete fast too, as the object keeps transforming.

This is of course the briefest of sketches of the concept of mediation, but I do find it very powerful, because it makes me think about the instruments mediating my interactions with the world. It also connect very well with the concept of the Worldview that I mentioned last time – it seems to me that our Worldview is our key instrument (along with language), but I'm not sure what the value of conceptualizing it as such could be.

I'll keep you posted about my findings.

Last week’s dig-ups

Entrepreneurship

Philosophy & Sense-making

Looks like this week it will be all about John Vervaeke. Well, he is one of my main models, so I guess that's to be expected.

• Here is John's take on the dangers of AGI. Some of the most interesting points for me were:

  • ChatGPT clearly demonstrates that propositional knowledge is completely orthogonal to morality, i.e., intelligence is very different from wisdom

  • AGI will face the same trade-offs of self-organizing systems as our cognition, meaning that it will also have to dream and have the risk of going mad

  • The proposed path towards alignment is to make the machines aspire to the truth and long for enlightenment. Which sounds rather close to Musk's solution. But John acknowledges that it would require us to become much wiser first. Which to me is a big ask, and also very much on-point.

• Here's the second part of the discussion with Jordan Hall on governance. They talk about a combination of both educating for new skills and values required for sovereignty, and using blockchain-like tech (in the way that it can't be captured and changed) to enable a much more decentralized approach.

• An older conversation between John and Jonathan Pageau – they talk a lot here about Attention, e.g., about it being an act of sacrifice and prioritization, and how attention simultaneously excluding and including is similar to how Roman law was based both on laws (what not to do) and models (what to do, what to model oneself after). Also, the better we can internalize the perspective of others, the better we can transcend ourselves.

Max Tegmark wrote a piece for Time on how the challenge of AI alignment is pretty much like the movie Don't Look Up: "Superintelligence isn’t a “long-term” issue: it’s even more short-term than e.g. climate change and most people’s retirement planning."

Reflection

• This week’s edition was veeery rushed, so I’m glad it’s not as empty as I was worried it will be. It also didn’t help that my snipd-readwise-logseq integration broke somewhere, making the curation take more work and also showing the inherent fragility of such PKM systems. I don’t think I will be changing anything though, I just need to fix it.

• One thing that made this week quite challenging was me joining a Startup Weekend run here in Oulu. It was great, I think I learned a lot, as I was able to work on my idea for self-management consulting business, and even found 2 teammates. So I’m now more ready than ever to get started… as soon as I finish the thesis

And that’s it! Let me know what resonated, and I’ll see you here next week.

Take care

Chris