For anyone engaging in the conversation around how we design ethical AI, the answer is: we don’t.
We don’t design for morality. We design systems that allow our humanity to breathe.
https://t.co/7u3C9ASnKM
AI safety researchers spend a lot of time asking how AI might destroy humanity. The more immediate question may be how we design systems that help keep humanity from destroying itself.
They say AI will end us. I say, AI will end the version of us that capitalism built — the exhausted, optimized, monetized, content-producing, perpetually-distracted version. And what rises in its place might be the most human thing the world has seen in 200 years.
@paulg Yup. It helps if you're really good at symbolic logic.
A good writer may not see the equation beneath the words, but they can feel it in the flow.
@SJ_MacroView@PeterDiamandis The printing press took around 300 years to reach mass scale. Society had time to adapt.
AI (including the technologies leading up to it) will unfold over about 30 years. Much less time to adapt.
So, less about complaining, more about our ability to adapt. Fast.
The internet gave a Sam Levinson speech to Killer Mike and called it consensus.
That mistake is now the basis of an argument about how we build AI. Stay with me.
https://t.co/kpFCA6oNNv
A true revolution has no allies. It also has no enemies that can reach it. Because it lives inside, and the inside is the one place that remains, even now, beyond the reach of capital.
@paulg This isn't just a Europe problem—as a millennial living in a North American city, having space to "tinker" just isn't there. It's completely out of reach for so many of us.
@bcdsignature If AI is going to live alongside us, it must meet us as moral agents in development—not children to be managed, protected from ourselves, or overridden for our own good.
Technology can loosen the grip survival has on work. It cannot tell anyone what to do with the freedom that creates. That part requires clarity. And clarity isn't the harder build—it's just a different kind of build.
The question worth asking is not how we build machines that manage what people do. It's how we build systems that help people understand why they do anything.
Intelligence and control emerge from constraint, geometry, and relational structure, not brute strength.
The biggest changes occur not by adding more force, but by adjusting the shape and direction of the system itself.
@JonhernandezIA They're asking: what rules should govern this system?
When they should be asking: how does intelligence stay in right relationship with evolving human meaning?
Static values vs living coherence.
@harari_yuval@nytimes I liked what you said RE: defining AGI as comparable to human intelligence—that it's like defining airplanes as "able to fly like birds." The comparison leaves out the natural substrate that makes bird flight what it is. I think that omission is a blind spot for many developers.