Bits of AI history and the current state. But no current or future applications of AI in industry.
The guy who discovers the first killer application in the space will be rich and famous. Until then, most efforts will go the way of scheduling theory.
Theoretical results are nice. But 90% of effort for building applications goes into domain and project specific details. So every new client project is starting nearly from scratch.
This is a consultant project-business. Not a startup product-with-market-fitness business.
Tech company stocks are going crazy. Nvidia more than doubled it stock in one year. I think it's still early but the premise holds true.
Once we thought human flight is impossible. Now we can fly 500+ people half way around the planet in a single aircraft because we discovered fundamentals of flight.
We're discovering some fundamentals of the human brain. Atleast the audio and visual cortex. Being able to achieve human level imagenet recognition is an amazing milestone.
Go was thought to be a human only skill. A skill of reasoning and intuition that could never be achieved by computers. We did it.
Technology grows exponentially. With better sensors, labelled data and cheap faster computers we can achieve parity with some parts of the brain.
The big question is how can we use it to build a better world. Privacy, security, income disparity and joblessness are very sensitive subjects that need to be tackled in a way that benefits everyone.
Oof this article is atrocious, ill-informed hype. Most of it is so high-level that it's essentially a paraphrase of Wikipedia, with some ridiculously specific details like gradient boosting and sigmoid activation functions sprinkled in for extra confusion.
>The breakthrough in deep learning is to model the brain, not the world... Artificial, software-based calculators that approximate the function of neurons in a brain are connected together.
This comparison is criminally incorrect.
And don't even get me started on figures 2+3 - 2 adjacent figures, with different horizontal and vertical scales, that use highly technical terms without introduction?? It's really hard to see how a lay-reader could get anything useful out of those plots.
Despite the author's promise to "get behind the headlines" and "cut through the hype", this is just a regurgitation of "intro to machine learning" cliches.
I have no expertise in this area, so I have no idea if you are right or wrong.
That said, I think you comment would be more useful to people like me if you said for several specific points what is the truth that the author got wrong, and also gave a link to an article for non-experts that gets things right.
7 comments
[ 4.4 ms ] story [ 26.5 ms ] threadBits of AI history and the current state. But no current or future applications of AI in industry.
The guy who discovers the first killer application in the space will be rich and famous. Until then, most efforts will go the way of scheduling theory.
Theoretical results are nice. But 90% of effort for building applications goes into domain and project specific details. So every new client project is starting nearly from scratch.
This is a consultant project-business. Not a startup product-with-market-fitness business.
Once we thought human flight is impossible. Now we can fly 500+ people half way around the planet in a single aircraft because we discovered fundamentals of flight.
We're discovering some fundamentals of the human brain. Atleast the audio and visual cortex. Being able to achieve human level imagenet recognition is an amazing milestone.
Go was thought to be a human only skill. A skill of reasoning and intuition that could never be achieved by computers. We did it.
Technology grows exponentially. With better sensors, labelled data and cheap faster computers we can achieve parity with some parts of the brain.
The big question is how can we use it to build a better world. Privacy, security, income disparity and joblessness are very sensitive subjects that need to be tackled in a way that benefits everyone.
>The breakthrough in deep learning is to model the brain, not the world... Artificial, software-based calculators that approximate the function of neurons in a brain are connected together.
This comparison is criminally incorrect.
And don't even get me started on figures 2+3 - 2 adjacent figures, with different horizontal and vertical scales, that use highly technical terms without introduction?? It's really hard to see how a lay-reader could get anything useful out of those plots.
Despite the author's promise to "get behind the headlines" and "cut through the hype", this is just a regurgitation of "intro to machine learning" cliches.
That said, I think you comment would be more useful to people like me if you said for several specific points what is the truth that the author got wrong, and also gave a link to an article for non-experts that gets things right.