The thing is, most people can't. In fact, if you reinvented calculus yourselves, you should be very proud of your mathematical ability.
The thing is, most people can't. In fact, if you reinvented calculus yourselves, you should be very proud of your mathematical ability.
Another example is that the article says "It would be hard for an AI to compose this image as it has lots of different elements/emotions" - but see the Chat-GPT example below:
> Me: Can you generate a prompt for an AI to generate an interesting image?
> GPT3: Sure! Here is a prompt for an AI to generate an interesting image: "Create a surreal scene featuring a giant, humanoid robot standing on a desolate, alien planet, with a small group of humans huddled fearfully at its feet. The robot should have a sleek, futuristic design, and its glowing red eyes should convey a sense of intelligence and power. In the background, there should be a faint hint of a cityscape, suggesting that the robot has traveled far from its home world. Use bold colors and contrasting lighting to create a sense of drama and tension in the scene.
I can then pipe this into Dall-E 2.
Sure, it's not GPT3 drawing the image, and Dall-E 2 might struggle to draw that, but most humans would struggle to draw that too as drawing is hard. Hard things are sometimes hard.
It’s not like asking if Stable Diffusion or ChatGPT can do it. Maybe there is specialized commercial AI that aren’t so widely known.
There exist right now people with low levels of physical and mental capabilities that current state of AI knowledge surpasses it by miles. We don’t go around saying “AI is better than humans”, even though - technically - it can be in some selected cases.
LLMs solve natural langage and common sense, which are not necessary for discovering calculus.
Additionally, without a construct, like a body, in which to store these memories, AGI will continue to disappoint those who do.
As ChatGPT can contradict itself within a single reply, I doubt that this is the only issue.