Some Good News About A.I. and Climate?
The conversation about "artificial intelligence" (which my wife and muse has long insisted should more accurately be called "simulated intelligence") is full of apprehension and possibility. A tool more significant than fire. Computers taking over the world. Concerns over trust, hallucination, IP, privacy, recursivity (as AIs inevitably start training on AI output), energy consumption. Oh, and that other kind of power, as new tech giants emerge on the economic landscape…this time not just monetizing our attention but our very words and thoughts.
Fear and excitement are physiologically identical, we're told, and we seem to have plenty of both. AIs are coming for your job, we're told. Or as Scott Galloway puts it, "People who know how to work with AIs are coming for your job."
Much of the conversation in the sustainability world has focused on the voracious and rapidly growing energy appetite of current AI technology. (Somehow blind to the __ that the price, performance, and energy footprint of new technologies often changes dramatically as we ride the learning curve and as penetration grows.)
Peter Leyden, a confessed techno-optimist, hosted an evening at the San Francisco Ferry Building two weeks ago to explore a different question “Can AI accelerate progress on energy and climate?” What might be possible, in other words, if we turn the unprecedented pattern recognition capabilities of AI, and its ability to iteratively test millions (trillions?) of hypotheses at blinding speed, not only to challenges like medical diagnosis and pharmaceutical innovation but to challenges like Factor 10—or or Factor 100!—innovation in data center design, electric grid efficiency and stability, transportation system optimization, building design, and more? (Including, ahem, AI energy consumption.)
The atmosphere was a bit breathless at times, as folks from Google, Microsoft, Stanford, MIT, OtherLab, and a bunch of startups you've never heard of laid out their visions and their wares. But it opened a window to a very different sort of exploration than what we're seeing in the popular press. Some very big things are moving very fast. (Peter hasn't posted the videos yet; I'll share them with you on my LinkedIn once he does.)
Architect John Picard put the evening to bed:
“As humans we’re not gonna get it done. We need tools that can see things we can’t see.”
So, yes, I've stepped into the fast-moving river! (To be continued…)