Frankly I think our conception is way too limited.
For instance, I would describe it as self-aware: it’s at least aware of its own state in the same way that your car is aware of it’s mileage and engine condition. They’re not sapient, but I do think they demonstrate self awareness in some narrow sense.
I think rather than imagine these instances as “inanimate” we should place their level of comprehension along the same spectrum that includes a sea sponge, a nematode, a trout, a grasshopper, etc.
I don’t know where the LLMs fall, but I find it hard to argue that they have less self awareness than a hamster. And that should freak us all out.
If you just read the tiniest bit of factual knowledge about how LLMs are constructed, you would know they don’t have the slightest bit of self awareness, and that it is literally impossible for them to ever have any.
You are being fooled by the only thing they are capable of: regurgitating already written words in a somewhat convincing manner.
LLMS can not be self aware because it can’t be self reflective. It can’t stop a lie if it’s started one. It can’t say “I don’t know” unless that’s the most likely response its training data would have for a specific prompt. That’s why it crashes out if you ask about a seahorse emoji. Because there is no reason or mind behind the generated text, despite how convincing it can be
Frankly I think our conception is way too limited.
For instance, I would describe it as self-aware: it’s at least aware of its own state in the same way that your car is aware of it’s mileage and engine condition. They’re not sapient, but I do think they demonstrate self awareness in some narrow sense.
I think rather than imagine these instances as “inanimate” we should place their level of comprehension along the same spectrum that includes a sea sponge, a nematode, a trout, a grasshopper, etc.
I don’t know where the LLMs fall, but I find it hard to argue that they have less self awareness than a hamster. And that should freak us all out.
If you just read the tiniest bit of factual knowledge about how LLMs are constructed, you would know they don’t have the slightest bit of self awareness, and that it is literally impossible for them to ever have any.
You are being fooled by the only thing they are capable of: regurgitating already written words in a somewhat convincing manner.
LLMS can not be self aware because it can’t be self reflective. It can’t stop a lie if it’s started one. It can’t say “I don’t know” unless that’s the most likely response its training data would have for a specific prompt. That’s why it crashes out if you ask about a seahorse emoji. Because there is no reason or mind behind the generated text, despite how convincing it can be
Yeah ask it about anything you know is false, but plausible, and watch it lie.