Every time I hear Stephen Wolfram speak, it's revelatory. The man has a unique vantage point for looking at where we've been, and where we’re going with AI. At a recent event, Wolfram broke this down in philosophical ways that made me think about the big picture. I think this is worth writing about, and thinking about, particularly at this moment that we are all in. “There's more and more AI,” he said by way of preamble. “There's AI everywhere in the world. What does it feel like? It's the continuation of a story that is gone through history … we start off as pure biological humans, and we invent devices. We also have ideas – these are things that kind of extend beyond our pure biological character.” AI, he points out, is an example of computation, where you set systems up to rules, and let them run…the principle of computational irreducibility means that you can't predict what these systems are going to do. A vintage tin toy
Robot collection belonging to Anthea Knowles,
UK, 16th May 1980. (Photo by ... [+] Fresco/Evening Standard/Hulton Archive/Getty Images) “That's something that we have not yet gotten used to,” he said. I thought it was instructive when Wolfram mentioned the Industrial Revolution, and hierarchical thinking, as tied to deterministic programming. Think about it this way – the software programs of the Web 2.0 era used text boxes and controls for input – input directly from human users. Then, much like Babbage's mechanical engine, the computer took that input and calculated results. There wasn't really any black box issue with that! But if you move beyond industrial/hierarchical thinking, into a world where software results are randomized, you have something that looks a lot like where we are now. That’s some of what I was thinking listening to Wolfram analyze the likely outcomes of the AI age. “We are surprised,” he said, as he enumerated challenges around trying to constrain these systems to be deterministic. “If we (constrain these systems) we will not get the power of computation is available to us.” Then Wolfram talked about this in another way that was inspiring to me. Think about AI is a new frontier of expiration, sort of like deep space… “There are many things out there that we have not yet explored,” he said. “Is there more science to be discovered? Are there more mathematical theorems – more inventions?” Then he threw out this interesting observation: “There’s sort of an infinite set of things that can be possible in the computational universe…the issue is: do they relate to anything that we humans care about?” This rings true to me in the age of the new “attention economy” – but there’s something deeper to think about here, too. Wolfram also made an analogy to the natural world, where, for example, we might find new stores of an obscure element or material, and ask: do we have a use case for it? “What will the AIs eventually do?” he said. “In the natural world, there's a lot of computation that we don't understand.” Thunder, lightning and rain during summer storm at night. Wolfram made the analogy to weather, which, as we’re fond of saying, has ‘a mind of its own.’ Will our interaction with AI be kind of like that? Another point that Wolfram made was around comparing computational models – for example, our use of language where we might use, say, 40,000 words, and computation systems that might use of millions of possible conceptual units, going far beyond what we see as normal. He asked us to think about this: what if we thought a million times faster than we currently do? Fast forwarding a bit, Wolfram also talked about the speed at which AI experiences the world around it. “As we look at the history of jobs,” he said, “systematic things that got done by humans were eventually automated - that automation ‘zeroed out’ that area of work, but it tended to open up many other possibilities. What you typically see is that things become more and more fragmented, there are more and more
Job categories.” As a corollary point, he noted that we should focus on the use of goals, which, he said, are essentially a human construct. “It takes humans to be inserted, there, to do the job to choose those things (that should be done)” he added. He talked about people having bots of themselves, and how you might say ‘talk to my bot’, sort of like we used to sarcastically say ‘talk to the hand’ … in this case, not facetious, but in earnest! In further assessing our visions of the future, he added, it might be daunting to think of the VR age as “a trillion disembodied souls playing video games for the rest of eternity,” but for the whole picture, we have to look back to history. “At any given time, there was a sort of a local view of what the purpose of life was, and people were satisfied or dissatisfied with that local view,” he said. “The things that we find important he said can change. I'm somewhat optimistic that as long as people are prepared to modernize as the world changes, that it will always feel like it's a world that's worth being in.” There you have it! I think this gives us a lot of fodder for thought, as we guide ourselves and our technologies – into the future. Stephen Wolfram of Wolfram Research in his library on May 7, 2009. Wolfram is the creator of ... [+] Mathematica, and the author of "A New Kind of Science". (Photo by Rick Friedman/Corbis via Getty Images)