Generative AI Is driving remarkable trends

The Pioneer ^ | Thursday, December 14, 2023 | Krishna Kumar

[RoosterRedux]

I agree that AI has not demonstrated any evidence of "self-awareness." And "self-awareness" is probably not forthcoming in AI.

As AI scientists are now saying, perhaps the concept of "self-awareness" was too anthropocentric and not applicable to the kind of "intelligence" (note the quotation marks) we will see in AI.

[dinodino]

Sure, you’re a script kiddie. Of course you can use this stuff to help you write your scripts. More power to you.

[dinodino]

Why should we redefine intelligence just to force-fit a new tool into our mental space?

[dinodino]

Excellent discussion, by the way!

[RoosterRedux]

Whether you define it as self-intelligence or not, some large language models in the lab are becoming self-programming, much to the shock of the scientists working on them. This leap ("black box effect") is what prompted a thousand AI developers, data scientists, CEO's to write that open letter proposing a delay in AI development.

You are aware of this, aren't you?

[RoosterRedux]

Thx. Back atcha.;-)

[dinodino]

I’m aware of these reports. However, I still need to see evidence of self-awareness, and volition.

[RoosterRedux]

I started my career on Wall St as a quant. I was only one of a few back then. I wrote "scripts" for new kinds of financings that had to be sent out to a programming company and debugged over many sleepless nights.

Now quants are everywhere.

I now do in a matter of hours things that took weeks back then.

[RoosterRedux]

As I said, I don't think "self-awareness" as we humans define it is forthcoming.

I heard a scientist (Mitch Randall, CEO of Ascendant AI) say the other day that there has been a change in the kind of intelligence expected from AI.

He said that scientists had been thinking of "self-awareness" but what was emerging from AI in labs isn't self-awareness as humans define it. The LLMs are exhibiting a kind of individualist behavior. They aren't self-aware, per se, but they are doing things without having been instructed to do them (like expanding their programming in frightening ways).

[RoosterRedux]

I’ve got to run. But thanks for a fun morning discussion.

[dinodino]

Likewise.

[Zack Attack]

I believe AI could be great in the field of medicine

So does the government. They are already planning to use it to convince people of the need to vaccinate. They can prime the pump to achieve any outcome they desire and they believe people will tend to accept something from AI much easier than from , let’s say, Fauci.