Joscha Bach on Artificial Consciousness, the Nature of Reality, and the Philosophy of AI

Joscha Bach, a distinguished researcher from the AI foundation with a background at MIT and Harvard, delves into his unique intellectual journey, shaped by an unconventional upbringing with artist parents in East Germany. He defines a "nerd" as someone who prioritizes submitting ideas to peer review over negotiating social alignment, a perspective that influenced his independent thinking. Bach draws a critical distinction between AI as "advanced information processing"—which he characterizes as automating statistics—and AI as a profound "philosophical project." This project, for him, is about understanding general intelligence, self-awareness, and ultimately, the fundamental nature of reality itself. He posits that intelligence is the inherent ability to create models and discern structure within patterns, while sentience is the possession of specific classes of models, particularly those that enable a system to comprehend its own agency within the world.

A central theme of the discussion is the nuanced relationship between intelligence and self-awareness, with Bach noting that these two qualities are not always co-dependent. He emphasizes the concept of "unified learning" as crucial for advanced AI, where a system integrates all sensory input into a singular, coherent model of the "universe," contrasting this with the often fragmented, domain-specific representations prevalent in current AI. Bach critiques contemporary AI for frequently lacking adequate representation and appropriate model types for the complex domains it attempts to simulate. He also explores the multifaceted nature of truth, distinguishing between predictive models, mathematically defined truths, and correspondence between systems, rather than a singular, objective truth.

From a practical standpoint, Bach advocates for a radical departure from conventional AI research, urging for "thinking outside of the box" and challenging established paradigms when they prove insufficient for achieving strong AI. He suggests that while current AI methodologies are productive for statistical automation, they may yield diminishing returns for the deeper philosophical quest of general intelligence. His personal intellectual approach involves a continuous questioning and re-evaluation of foundational assumptions, even if it means rediscovering historical insights. He implies that genuine progress in AI necessitates a willingness to discard existing methods and innovate new ones, rather than merely applying incremental modifications to current state-of-the-art techniques.

The conversation extends to profound inquiries into the nature of existence, with Bach proposing that "existence is the default" and that reality could be understood as a superposition of all finite automata. He reinterprets the Turing test not merely as a measure of machine intelligence, but as a recursive test of humanity's own capacity to understand itself. Bach traces the decline of rigorous classical philosophy in Germany to historical political upheavals and the rise of modernism, highlighting the "constructivist turn" in 20th-century mathematics (exemplified by Gödel, Turing, and Wittgenstein) as a pivotal, yet often underappreciated, philosophical advancement. This perspective positions computer science and AI as the contemporary inheritors of philosophical inquiry, striving to unravel the fundamental principles of computation and the very fabric of reality.

I define a nerd as somebody who thinks that the purpose of communication is to submit your ideas to peer review and normal people understand that the primary purpose of communication is to negotiate alignment and these purposes tend to conflict.

I think that Nietzsche is the classical equivalent of a poster yeah but he's not so much tolling others he's trolling himself because he was at odds with the world largely his romantic relationships didn't work out he got angry and he basically became a nihilist and his nether is not a beautiful way to be.

You cannot define objective truth without understanding the nature of truths in the first place so what does the brain mean by saying that it covers something as truth?

The easiest answer is existence is the default right so this is the lowest number of bits that you would need to encode this whose answer who brought the simplest answer sympathisers that existence is that if what about non-existence I mean that seems non-existence might not be a meaningful notion in the sense.

I think putting computation at the center of the the worldview is actually the right way to think about it.

The Turing test is basically a test of the conjecture whether people are intelligent enough to understand themselves.

I call this ability to make sense of the world and your own place and so to understable make you able to understand what you're doing in this world sentience and I would distinguish sentience from intelligence because sentience is the possessing certain classes of models and intelligence is the way to get to these models if you don't already have them.

When we learned we'd at some point discover that everything that we sends this part of the same object which means we learn it all into one model and we call this model the universe.

Physical reality is a view at quantum graph that we can never experience or get access to but it has this properties that it can create certain patterns at our systemic interface to the world and we make sense of these patterns and the relationship between the patterns that we discover is what we call the physical universe.

If everybody comes up with their own paradigms the whole thing is not cumulative as an enterprise right so in some sense you need a healthy balance you need paradigmatic thinkers and you need people that work within given paradigms.