It is ludicrous to think of the current LLM based approaches as a stepping stone to General Artificial Intelligence. But even if it were, this would not be a magic bullet to save human civilization or anything like that, except the reverse.
The Problem of Human Civilization is not "lack of intelligence." The core of the problem is lack of wisdom, which is something altogether different and mostly orthogonal if not oppositional. One key part of wisdom, for example, is self-restraint. While Intelligence tells you how you can conquer the world, thereby enabling and encouraging you to do so, Wisdom tells you it would not be a good idea.
Humans developed to capability to harness natural forces, and as a result have been digging their planetary home into total disaster. This is a problem that is most accurately diagnosed as Too Much Power with Too Little Wisdom. Intelligence is merely a form of power...albeit a foundational one which makes most others possible. Throwing more power into our flaming cauldron will only make the flame hotter and likely melt the vessel.
And this is before we even get into another more commonly discussed problem: Who own the AI and what is it used for?
It's clear that AI is owned by the oligarchs and pathocrats and will be used, as every tool in their hands is used, to oppress and further enslave if not murder everyone else. The biggest first use is targeting "enemies" in protests and war through AI firms such as Palantir. It only goes downhill from there.
Our human "intelligence" is only small part of our set of long evolved capabilities. Nervous tissue is not necessarily superior to silicon for computation or anything else, except that nervous tissue as part of evolved organisms has been trained for hundreds of millions of years. Even our individual 'training,' being thrown into this world and having to somehow adapt to both it and our long-evolved selves, is something that cannot be replicated even by reading all the books in the world.
Both this evolution and this training can impart at least small amounts of wisdom.
Our intelligence and other capabilities have limits which effectively enforce their wiser usage, as more reckless usage is unsustainable.
Those are precisely the kinds of limits some hope to superceed with AI. But it is those limits which also require and therefore enable a degree of wisdom.
The wiser course is to embrace limits and live within them. Machines can help us get our dishes clean, and that's nice, but what the main course is is up to us.
To err is human, to really foul things up requires a computer.*
(*Commonly attributed to Limits-to-Growth luminary Paul Ehrlich, it may actually have been penned by columnist Bill Vaughan, who was paraphrasing Agatha Christie.)
What about Artificial Wisdom? There have been lots of attempts to get there through things like meditation, reciting phrases, reading religious books. Of course, there is no such thing, but you could hardly do better than randomly selecting a page from Tao Te Ching every day. It is not going to answer all of your questions, but that is in the nature of Wisdom. It is limited, but that is exactly what is required for limited beings like us anyway. Unlimited wisdom is something we can't process.
This post was inspired by a talk on transhumanism, which sounds abhorrent to me.
There is no need for this at all. As Uber has proven, human drivers can be cheap because people need paid work. Why is there such a great "need" to replace them with AI? You don't need to go everywhere you can think of going. Doing so is a waste of time and physical energy. Having to pay someone for something is a way of preventing over usage.
Liked.
ReplyDelete