Why Gen AI may be much bigger than even the hype suggests

Kurt Gödel’s famous Incompleteness Theorem is a thing of beauty in two ways. First, it proved something really powerful and counterintuitive – that any system of logic has statements that are true but can’t be proved by that system. Secondly, the way he proved it was super-amazing. He turned mathematical expressions into numbers, then used maths to manipulate the expressions themselves. [Mathematicians – please excuse my loose language here.] This second part is really amazing. A strange loop. Like taking a whole country’s cuisine and turning it into an ingredient, then cooking with it. Now, cynics who look at many current uses of Generative AI might say “Whoop dee doo, a computer can write a story or make a pretty picture. So what.” But in the spirit of Gödel, the point is that Gen AI can operate on anything that can somehow be collapsed to information, i.e. numbers. Which is actually pretty much anything. Gen AI can create me an essay, or a movie, or a set of accounts, or a new engine design, or the DNA of a new creature, a new medicine or a new toxin,  a computer program to analyze brainwaves, a new economic or political system, or ten thousand websites that all refer to each other and propagate a deep fake lie. And for everything it creates, it can create one of the thing, or it can create millions and test which is the best through simulation. If we stick quantum computing underneath that in future, it will be able to design and build extraordinary everything. And when we hook Gen AI to things like 3D printing and CRISPR, it is not just able to generate informational payloads, but physical and biological ones too. The supply side of Gen AI will continue to evolve and improve over the next few years. Smaller, domain specific models. Explainability. Accuracy using neurosymbolic approaches. Etc. etc. But perhaps even more importantly, we will find and imagine ways to apply Gen AI to more and more domains. Demand-side innovation. This demand side point applies to some other forms of AI and other technologies too. As well as worrying about the dark side and the practicalities, let’s all reserve some time to wonder at the possibilities. To finish, let me shared a non-boring definition Of AI that I read recently (maybe from author Rudy Rucker): "The Goal of Artificial Intelligence is to create a machine so beautiful that a soul would want to live in it." #Unthink #Unask #Unlearn #AI #GenAI