Most of the value of deep learning today is in narrow domains where you can get a lot of data. Here's one example of something it cannot do: have a meaningful conversation.
Interpretation
Deep learning is effective in specific areas with ample data, but it struggles with tasks requiring meaningful conversations.
Andrew Ng's quote highlights the strengths and limitations of deep learning technology. While it excels in data-rich environments and can perform spectacularly in narrowly defined tasks, it falls short in areas that necessitate nuanced, human-like interaction, such as engaging in meaningful conversations, which remains a challenge for artificial intelligence.
In practice
In a discussion about the limitations of AI, one might say, 'As Andrew Ng pointed out, most of the value of deep learning is in narrow domains, but it cannot have a meaningful conversation.'
AI has been making tremendous progress in machine translation, self-driving cars, etc. Basically, all the progress I see is in specialised intelligence. It might be hundreds or thousands of years or, if there is an unexpected breakthrough, decades.
It seemed really amazing that you could write a few lines of code and have it learn to do interesting things.
Imagine if we can just talk to our computers and have it understand, 'Please schedule a meeting with Bob for next week.' Or if each child could have a personalized tutor. Or if self-driving cars could save all of us hours of driving.
A single neuron in the brain is an incredibly complex machine that even today we don't understand. A single 'neuron' in a neural network is an incredibly simple mathematical function that captures a minuscule fraction of the complexity of a biological neuron.
None of us today know how to get computers to learn with the speed and flexibility of a child.
I've been to so many manufacturing plants. I've yet to walk into one where I did not think AI solutions wouldn't help.
There is a difference between what technology enables and what historical business practices enable.
Techno-humanism aims to amplify the power of humans, creating cyborgs and connecting humans to computers, but it still sees human interests and desires as the highest authority in the universe.
Machine learning is looking for patterns in data. If you start with racist data, you will end up with even more racist models. This is a real problem.
The only reason we don't notice how absolutely interwoven our thinking processes have become with older technologies - pencils, paper, electric light, penicillin, fire - is that they're old, so we've ceased to notice their effects.
The Internet was supposed to be the greatest tool of global communications and means of sharing knowledge in human history. And it is. But it has also become the most effective instrument of mass surveillance and potentially one of the greatest instruments of totalitarianism in the history of the world.
When something is such a creative medium as the web, the limits to it are our imagination.
Subscribe for the occasional hand-picked quote. No noise.