Ali Eslami

But it's not as scary as I think most people think it to be because...

as humans i think we're very comfortable with interacting with other humans and we don't fully understand other humans right so society works um through systems of trust through communication and so on it's the same with these ai systems we don't fully know what they can or cannot do we understand the mechanics of it but we don't know the capabilities

But one thing that is new is this idea of self-improvement.

So in the past, yes, we humans built systems that did cool things, but we didn't really fully understand what they were.

But now we're in a realm where we build systems and those systems can continue to work on themselves.

They can continue to change their capabilities.

In fact, we want them to continuously change their capabilities.

So as researchers, our jobs become slightly more like social scientists almost, or anthropologists, where you're kind of measuring and steering the evolution of complex system or society.

build and interact with ai systems what can we extrapolate from that about the future of human creativity right okay i mean future of human creativity is super broad so i can only share my own personal experience um i have been really struck by how much my job has changed uh in the past couple of years um

careers obviously normally change as you get more senior, but I'm talking about something different.

The nature of the job has profoundly changed, I would say at least three or four times in the past three or four years.

And the time span between the changes has only decreased.

So there was the era before deep learning.

Then there was the time between deep learning and chat GPT.

That was about 10 years.

Then between chat GPT and maybe two years ago, that was like two or three years.

Then there was a year when the tech was becoming much more public.

Then just since the beginning of this year with OpenClaw, that's been two or three months ago,

And now the capabilities keep going up and up.

Each one of these moments, as a practitioner, as a researcher, I basically had to completely set aside my previous tool set and pick up a new tool set in order to continue my job.