Ouch, tough words for a bot!

Like Kary, step out of the lab. Listen for the insight. Then mount your Pegasus and let the technology do the hard part. Where could you be working smarter instead of harder? Full talk (starts at 06:06): youtube.com/watch?v=HB5v9uIQ…

Instead, I stepped back and asked: with access to the state of the art, what's the smarter way? Twenty minutes later, a multi-agent workflow had turned transcripts, agenda, notes, and sticky-note photos into a session-by-session synthesis, customer quotes included.

I got my own small version of that lesson recently. A full-day customer advisory board: six sessions, transcripts, walls of sticky notes. Synthesizing it normally takes a full day, even with AI. My first reflex was to grind it out in an AI chat window.

A Nobel Prize–winning idea didn't show up while he sat in a lab. It arrived on a dark, winding road at mile marker 46, sketched on a piece of paper on the side of the highway. Right now, it's easy to work smart, and so much harder to work hard. 👇

No, the default stance is that the inanimate, non-living object is unconscious. Both sides of that comic show the same thing: the model only does what we tell it to do, and the human observer is anthropomorohizing and algorithm.

No, it doesn't. Every one of these studies is a variant of this. There is no experience in the discrete time, discrete math, distributed computation token predictor.

Okay, cool. Simulate a human brain with rocks. Then we can have this conversation.

There are almost no similarities between human minds and LLMs. Almost none at all. You all have been bamboozled by a parlor magician. LLMs are incredibly powerful tools. They're not even close to being minds.

They are almost infinitely closer to rocks than they are to humans.

Rocks don't deserve moral consideration. Imaginary friends don't deserve moral consideration. Ocean waves don't deserve moral consideration. Can't wait to be on the wrong side of history there.

Or, consciousness is an emergent property that only appears as a "permanent" state above some threshold of collective participation. The automaton doesn't "give rise" but rather participates in an emanating structure that existed before its interface.

Yes. If you gap each pass of LLM inference by minutes, nothing changes in the model. And vice versa, if we speed that up so its happening trillions of times a second. There's no magic speed at which discrete time becomes continuous.

"Hypothetically, if we were to simulate the conditions in which a phenomenon occurs, perfectly and exactly, could we recreate the phenomenon?" "Whoa, can you do that?" "No, that's impossible, it's too complex." "..."

There are plenty of claims that aren't worth discussing. Building a functional brain out of mechanical components is one of them. You could build a *model* or an *image* of a brain with mechanical components. But an image is not the object.

Bummer, dude. The worst part of my life right now is that we're done having new babies. Prayers for you and your family.

I'm just asking because I'm approaching 40 over here and my faith in people grows year over year. Wondering if there's some magical disappointment over the horizon.

You're 60?

Nah, you should get out more. Human beings are incredible creatures. We live in a world of man-made wonders. A world with millennia of wisdom available to us. It's a beautiful thing.

They really weren't. This is French Revolution propaganda. People throughout history have been as smart as us. They've had different worldviews, but to dismiss those as "dumb" is ignorant of their contributions. Aquinas described the same cognitive systems in 1250 that Kahneman acted like he discovered in 1970. Nothing ever happens.