Thinking and out-thinking

Cristian Ispir
2 min readApr 14, 2021

A recent business education ad noted that ‘there is no great though leader who can out-think the internet’. And OpenAI’s GPT-3 is the most powerful language-based A.I. because it’s been trained on 45 terabytes of text data extracted from the unfathomable depths of the internet. The quantity problem seems to have been resolved, we have enough data, the information is there, always more plentiful. If infinity is an imaginable point — which shouldn’t, really — then the internet is approaching it every day. It is generally agreed that we’ve long passed the point where we can make sense of the available data. Make sense, that is, on our own.

But information and data are, let’s face it, merely information and data. As important as, if not even more important than information is what we make of it. In other words, the key thing is understanding. And that can’t be outsourced to a machine yet. John Searle’s Chinese Room may have transitioned from a mere thought expriment to a live situation, but understanding and the application of meaning, at their most basic level, lie outside the room. There is no robust evidence to suggest that the crevasse between information gathering and meaningfulness will ever be filled. So far, it has been filled with faith. Faith that technology will close the gap, which is the belief that given the progress made so far, more of it will follow, right up to the point that the problem will be solved, and A.I. will finally be able to pick itself up by its own digital bootstraps. If you invent the wheel, it doesn’t follow that one day you’ll develop the automobile, even though we can see, with hindsight, that this happened indeed.

In the meantime, we remain the custodians of meaning, which is no small thing. We don’t out-think the internet, we are the ones who make the internet think. And this goes beyond making inferences or reducing errors through probability. I will know the A.I. is headed towards that Eldoradian region when it has asked the question ‘why’ before setting out to do what it does best.

--

--