Culture programs a society to operate in a desired way. Social norms, memes, and even the very language we use, shape how we think and therefore act in the world.
Newspapers, TV, social media – these are all powerful engines of cultural message delivery and curation. This is why we have PR, brand, and influencers – if I can present a specific view of the world and convince you of something, I can manipulate you into acting in a certain way.
But even more powerful than culture is environment. We are literally products of our environment – the resources and tools that we interact with determine our actions and outcomes.
A prevailing facet of western culture is the rejection of censorship – we believe that a centralised party controlling what media makes it into the public forum is wrong in most cases. We flinch at the idea of Google or Apple modifying how they operate based on the physical location of the user.
If I can’t find a record of an event that occurred via Google, nor 3rd-hand accounts on community-driven sites, the quality of my decision making is impaired.
And now we’re building “foundational” models (which itself is a beautiful bit of PR) that will serve as the guiding lights of learning and research for a new generation. While social media has offered a depressing glimpse into attention capture writ large, ChatGPT hints at a potentially more dangerous outcome.
With social, ideas must be diffused and remixed. There is clearly a lot of abuse of this system to maximise for manipulation and monothink, but GPTs are currently being positioned as oracles. Even though the type of intelligence that they represent is so fundamentally different to the panoply of human intelligence, we’re teaching the world that they’re to be trusted.
The foundations of monothink have already been laid, my main worry is that the starting point has the biggest impact of the destination and direction. What will distilled monothink do once diffused across the world?