Hilary Putnam

Yet what they're really asking for underneath the word is something completely different.

Hilary Putnam came up with this thought experiment to famously prove that meaning just ain't in the head.

The fact that two people have the same idea in their heads and that idea means the same thing to each of them doesn't mean that they're referring to the same thing in their physical realities.

Meaning comes from our external worlds as much as it does our internal ones.

So he's talking about function in the way that we speak of function nowadays under the rubric of functionalism.

So Hillary Putnam's idea of functionalism, which was introduced in the 50s, actually Freud introduced it in 1900.

In fact, he introduced it in the 1890s, but he explicitly in the interpretation of dreams says, I'm dissecting the function in the sense that you speak of what goes on between the lenses of a telescope.

And he said it makes no difference what the telescope is made of, whether it be metal or cardboard, whether the lenses be glass or plastic, the function of the telescope is distributed between its parts.

After that, the computer would trick the brain into believing it was a fully living, breathing person.

Nearly a decade later, in 1981, a philosopher and mathematician named Hilary Putnam took things a step further.

Following the same line of reasoning as Descartes, Zhuangzi and Plato, Putnam suggested that since we can't fully trust our senses, there's no way to prove we aren't brains in a vat.

The brain in a vat theory became one of the most enduring concepts in philosophy, and science fiction writers around the world were eager to incorporate the idea into their work.

For Rorty, there was no need for truth to answer to anything beyond what an audience could be talked into agreeing on, and he'd been making that case to Habermas for years.

Lafont, my advisor, who'd arranged the hotel meeting, pushed from the other side, drawing on the US philosopher Hilary Putnam, who had spent much of his career arguing against exactly the kind of view that made truth append entirely on human agreement.

Lafont had argued since 1994 that Habermas's own philosophical tradition, in its turn toward language, had drifted into a position where it couldn't say the world holds firm no matter how we talk about it.

When Habermas changed his mind, he moved toward Lafont's position, not Warty's.

The US philosopher Richard Rorty pushed from the side from which Habermas was now backing away For Rorty, there was no need for truth to answer to anything beyond what an audience could be talked into agreeing on, and he'd been making that case to Habermas for years.

Lafont, my advisor, who'd arranged the hotel meeting, pushed from the other side, drawing on the US philosopher Hilary Putnam, who had spent much of his career arguing against exactly the kind of view that made truth depend entirely on human agreement.

Lafont had argued since 1994 that Habermas's own philosophical tradition, in its turn toward language, had drifted into a position where it couldn't say the world holds firm no matter how we talk about it.

When Habermas changed his mind, he moved toward Lafont's position, not Rorty's.

All right.

The Hill and Putman.

How could we miss that?

The Twin Earth, we didn't talk about.

Again, it's the brain in the vat problem.

If you don't know what I'm talking about, go look up Hillary Putnam.

He's a mid-20th century philosopher who detailed the fundamental problem with brain in a vat, which is what LLMs are, way back then.

So this isn't misalignment, this is LLMs working the way they were designed to.