(QGI) Quantum General Intelligence

I predict a disruption - a major disruption. It will occur as GenAI's progression towards AGI (Artificial General Intelligence) converges with true Quantum programming capabilities. I am coining the term as "Quantum General Intelligence" or QGI.

Where is AGI?

Let's look at the staggered advancements of AI. From Expert Systems to Machine Learning to linear Neural Network models to Adversarial Networks to Large Language models. (Stefan Kojouharov does an excellent job explaining the early progressions in his article: Becoming Human: Exploring Artificial Intelligence & What it Means to be Human.)

And then we have the hyper breakthroughs with OpenAI predictive models being augmented with Q* ("q-star") mathematical processing models. Several enthusiasts such as Matthew Berman have explored this potential milestone in What Is Q*? The Leaked AGI BREAKTHROUGH That Almost Killed OpenAI.

Beyond the backend computational advancements, AGI's modes of operation will advance to achieve true sentient intelligence - which include "self-awareness", the ability to "feel" intangibly, and have "desires" that may conflict with more optimal calculated outcomes. Ellen Glover does an excellent overview of this topic in her article What is Sentient AI?

What's Holding AGI Back?

There are indeed great advancements on the horizon in various technological arenas.

However, if we look at the neural network computational models as well as the predictive models, we know they are in fact developed using linear programming languages on hardware with massive computational capabilities. But, in the the programming logic is linear.

This means as the number of parameters increase, our technical ability to compute all the potential outcomes is limited by hardware running these one-dimensional algorithms.

Where Does Quantum Programming Fit?

Quantum computing has the potential to reset Moore's Law with AGI.

Looking at early x86 processors, Intel had reached a limit to how fast their processors could execute instructions. They decided to house multiple processing units on a single processor that could run in parallel, which could theoretically increase execution throughput. But most computing logic is hindered by decision points, areas where logic diverges based on previous computed outcome. To combat this issue, Intel introduced instruction and branch pipelining across those processing units.

Pipelining is not limited to low-level instructions of a CPU, it is a tried and true method for graphical applications, such as loading video frames for real-time playback.

Today's AI algorithms, at their most atomic level, are modeled in the same fashion: Linear programs with logic pipelining to emulate the multiple outcomes from an enormous set of parameters. Quantum programming today is a bit of a misnomer, as they again use emulation to mimic the "spooky" behavior of outcomes from an set of parameters.

The "QGI Convergence"

I predict the most significant breakthrough in AGI will come when we can truly conceptualize how to define problems and solution sets in terms of quantum logic using 3rd-generation quantum programming language models, and run it natively on a quantum computer.

This will be known as the convergence of Q3GL and AGI; or more simply "QGI Convergence" ...

Visually speaking, the Convergence is the alignment and/or overlapping of quantum programming branch logic with neural network branch logic.

True artificial intelligence is not a series of massive linear decisions, but a conglomerate of parameters that may have multiple outcomes, all in parallel.

AGI is the poster child for indeterminate (aka "spooky") behavior. Native quantum programs can be built for AI-based problems, once we know how to define the problem and the quantum logic to address the problem.

Summary

Sometimes in humanity we must restart from scratch to rebuild a new paradigm.

I am not suggesting current AI models should be destroyed; it's a genie that cannot be put back in the bottle.

What I am suggesting is that the algorithms, models and transforms we are familiar with today are moot when it comes to quantum programming, Instead of trying to replicate a linear algorithm that uses faux-pipelining methods find deterministic outcomes, a 3rd-generation quantum programming language would have that capability inherently rethink the problem entirely.

Quantum computing brings with it a native programming paradigm shift that aligns perfectly with the conceptual pinning of AGI. We just haven't caught up with the mechanics of it.