Exponential View

Exponential View

Share this post

Exponential View
Exponential View
🎧 AI's Near Future

🎧 AI's Near Future

My conversation with Jürgen Schmidhuber

Azeem Azhar
Aug 21, 2019
∙ Paid
1

Share this post

Exponential View
Exponential View
🎧 AI's Near Future
Share

Hi,

I’m sharing here the transcript of my full conversation with Jürgen Schmidhuber, whose pioneering AI research has enabled breakthroughs in commercial AI.

We discuss:

  • The development of long short-term memory (LSTM) and its role in machine learning, including Google and Facebook translate functions.

  • How LSTMs overcame a problem in previous feedback network models, allowing machines to ‘remember’ previous events and build up an understanding of their environment based on previous observations.

  • The role of LSTMs in facilitating recent AI breakthroughs in video game-playing bot learning.

  • Meta learning (learning to learn) and transfer learning (learning from one situation and applying what was learned to another situation), and the role of each in AI development.

  • The shift from the Cloud to local, edge computing.

  • The next AI wave will see machines that not only passively recognize data, but also actively shape the data that they perceive (including creating and adapting their own internal and…

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2025 EPIIPLUS1 Ltd
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture

Share