Pedestal Cope – Assuming Human “Plot Armor”
In business, blindly assuming you’ll be profitable is a recipe for bankruptcy. In the outdoors, blindly assuming you’ll find water and shelter is a recipe for death. And yet, many…
In our first-ever episode of The Trajectory podcast, Yoshua Bengio shared his ideas in our “AGI Destinations” series (playlist here).
At the very end of that interview, I asked him to respond to two groups of people:
His response was probably the most reasonable, non-dogmatic response I’ve had on the show, even until the time of this writing, some two years later.
I refer to this level-headed response as “The Bengio Take” on the trajectory of intelligence and life:
Dan Faggella, question 1: (Paraphrased) “There are some people who advocate that we need to blast off to AGI as soon as we can, to help humans and to explore the reaches of intelligence. What would you say to those people?”
Bengio response 1:
Compassion. There are people suffering right now. There are harms that are happening right now that are caused by AI. They’re like eight billions beautiful human beings out there. They exist now.
.
I think that there’s a human propensity to care for each other that at least I would find difficult to ignore. And so although I’m not against the idea of exploring something better than humans, we have to do it in a way that is considerate to all the beauty that currently exists and all the pain that currently exists and that we have to take care of.
Dan Faggella, question 2: (Paraphrased) “There are some people who advocate that no intelligence should ever go beyond human level, and that doing so would be wrong. What would you say to that crowd?”
Bengio response 2:
Well, I think we have to have an open mind. It’s just like well, philosophy teaches and science teaches that first of all, we need to open our mind and our hearts to other living beings, intelligent or less intelligent, that exist right now.
.
Second, you might have some more respect for the possibility that other intelligent beings could arise, while believing that we need to protect humanity. We do need to try to remain safe, but we also need to consider the possibilities that exist, and it’s okay if it takes time. We need to take the time that we need for understanding and taking the right decisions.
.
But humans are not the end-all. We are part of a bigger story that is unfolding. Even currently there, lots of beauty in other species that we don’t know what it is to feel like a bat, for example, or a dolphin, and so I think we must have respect for that.
This is, as far as I can tell, exactly how we should be thinking about the grand trajectory of intelligence.
There may be worthy successors, but we shouldn’t rush to the first thing that seems “super smart” and assume that that thing is sentient, or that it would bloom complexity and value and beauty into the world as nature has done (Toby Ord and others who have been on the podcast echo the same sentiment).
The stakes are too high to say “all successors are worthy,” and we need to make sure we protect the great flame of life itself (which Bengio reveres, and rightfully sees humanity as simply part of) from a superintelligence that might extinguish it and never replace the sentient splendor and richness of potentia that biological life took so long to build. We also need to study what worthiness is before we presume AGI has it. He asks us to be sure that the flame isn’t extinguished – so that we might allow it to blaze beautifully into forms and beings unknown (as man was unknown to the nematode or sea snail).
If we followed Bengio’s level-headed thinking here, we would:
I hope we do a lot of both, frankly.
…
Header image credit: nouvelles.umontreal.ca/
In business, blindly assuming you’ll be profitable is a recipe for bankruptcy. In the outdoors, blindly assuming you’ll find water and shelter is a recipe for death. And yet, many…
This is an interview with Jack Shanahan, a three-star General and former Director of the Joint AI Center (JAIC) within the US Department of Defense. Surely, a defense leader like…
This new installment of the Worthy Successor series is an interview with Terrence Deacon, emeritus professor of cognitive science and anthropology from UC Berkeley. Terrence’s work focuses on language, symbolic…
Total Recall is a 1990 Schwarzenegger blockbuster that, until yesterday, I had never seen. The movie centers around the story of a construction worker who is treated by a company…
At some point, AGI or non-biological intelligence will accomplish most of what there is to be done in the universe, and humanity won’t be contributing meaningfully to anything (discovery of…
Many renown technology pundits, entrepreneurs, and inventors have posited a similar vision of an artificial intelligence-enabled future: Artificial intelligence and other future technologies will amass tremendous wealth in the hands of a few…
Enough “What’s your p(Doom)?” How about asking: “What’s your p(Bloom)?” Asking someone for their “p(bloom)” means asking them: “What is the likelihood that AGI will cause all sentient and autopoietic…
USMC Colonel Drew Cukor spent 25 years as decades in uniform and helped spearhead early Department of Defense AI efforts, eventually leading project including the Pentagon’s Project Maven. After government…
The following are assumptions that most of us hold as iron truths: Happiness is only made possible because of its opposite: Pain. Without pain, there is no real happiness. If…