I enjoyed this conversation with Rajat about the past, present, and future of TensorFlow. Here's the outline: 0:00 - Introduction 1:17 - Google Brain early days 4:47 - TensorFlow early days - open sourcing, etc 12:53 - TensorFlow growth 22:00 - Keras 26:24 - TensorFlow project management 37:10 - Competition and PyTorch 39:48 - TensorFlow 2.0 51:20 - Building a good software engineering team 1:03:48 - Search ads and paying for content 1:08:43 - Using TensorFlow on a budget 1:10:16 - How to get started with TensorFlow
Thank you for taking the time to annotate the topics in your talk. I'm sure someone will use it to do some topic classification from the segments in the talk.
Thank you Lex Fridman for making the best AI podcast! Your guests are fantastic and I believe you bring an amazing outlook at AI and ML from within the core of ML itself.
I use tensor flow for nearly all of my network models. It’s proven extremely valuable for image and language models. This insight behind development and ideas behind TF was fascinating, you’re really incredible at what you do. Thanks for such a good podcast!
Can you have a podcast with Roger Penrose? The books he wrote on the quantum physics aspects of consciousness are really interesting. It would be interesting to know what he thinks now on these topics.
The TensorFlow 1.0 had some drawbacks. But I have even using 2.0-beta and I am very happy with it. I liked how much of the bloat was eliminated and Keras is now used to define models in an higher-level manner while the low-level stuff is still accessible using PyTorch-style eager execution.
Very cool intro to the history and future of tensorflow.... What resources would you recommend to start a journey in AI and ML algorithms?...which show practical applications ( paid or free)... Many thanks for your amazing channel...
Lex is the man! Inspired by Lex's work, I have started this channel to help distribute tech contents to Nepali communities. Our next episode is coming next week, please subscribe us and help this community grow.
@@vinaydebrou83 thanks. what else advise? and I was told that fastai will have a course for creating a deep learning framework that you know something about this?
AI is the field the moves faster because it's open source. And that goes to show you that they were lying to you when they told you that predatory capitalism is the best we can possibly dream of. Thanks Rajat Monga. Thanks Lex Fridman. Thanks to all the beautiful people who everyday spend time making top-quality code available for everybody.
I enjoyed this conversation with Rajat about the past, present, and future of TensorFlow. Here's the outline:
0:00 - Introduction
1:17 - Google Brain early days
4:47 - TensorFlow early days - open sourcing, etc
12:53 - TensorFlow growth
22:00 - Keras
26:24 - TensorFlow project management
37:10 - Competition and PyTorch
39:48 - TensorFlow 2.0
51:20 - Building a good software engineering team
1:03:48 - Search ads and paying for content
1:08:43 - Using TensorFlow on a budget
1:10:16 - How to get started with TensorFlow
Thank you for taking the time to annotate the topics in your talk. I'm sure someone will use it to do some topic classification from the segments in the talk.
This is great, thank you!🔥
Thank you Lex Fridman for making the best AI podcast! Your guests are fantastic and I believe you bring an amazing outlook at AI and ML from within the core of ML itself.
I use tensor flow for nearly all of my network models. It’s proven extremely valuable for image and language models. This insight behind development and ideas behind TF was fascinating, you’re really incredible at what you do. Thanks for such a good podcast!
This podcast is absolute gold!
Good Interview. Interviewing Jeff Dean would be great.
I can't even believe you are still so good at electrifying me with those interviews.
Your job in AI is priceless.
Can you have a podcast with Roger Penrose? The books he wrote on the quantum physics aspects of consciousness are really interesting. It would be interesting to know what he thinks now on these topics.
The TensorFlow 1.0 had some drawbacks. But I have even using 2.0-beta and I am very happy with it. I liked how much of the bloat was eliminated and Keras is now used to define models in an higher-level manner while the low-level stuff is still accessible using PyTorch-style eager execution.
Good person to interview!
thanks for this talk
Very cool intro to the history and future of tensorflow.... What resources would you recommend to start a journey in AI and ML algorithms?...which show practical applications ( paid or free)... Many thanks for your amazing channel...
Hi< I'm from Iraq, i hope in the future have English subtitle for understanding some words> thanks
Appreciate the thoughts on tensorflow.js wish more appropriate what JavaScript can do for machine learning.
Ay, another one?
I think next podcast will be with andrew NG.
Lex is the man!
Inspired by Lex's work, I have started this channel to help distribute tech contents to Nepali communities.
Our next episode is coming next week, please subscribe us and help this community grow.
Nice.
What do I need to know to create my own Python deep learning framework? What are the books and courses to get knowledge for this?
I'd suggest you to start from Andrew Ng's Machine Learning course (it's freely available on internet).
@@vinaydebrou83 thanks. what else advise? and I was told that fastai will have a course for creating a deep learning framework that you know something about this?
Why is it that whenever I find an Indian person in tech and I look him up, its always an IIT grad?
They work harder.
Because JEE filters the best. That's where it starts. IIT just make sure many of them retain their mettle.
AI is the field the moves faster because it's open source. And that goes to show you that they were lying to you when they told you that predatory capitalism is the best we can possibly dream of.
Thanks Rajat Monga.
Thanks Lex Fridman.
Thanks to all the beautiful people who everyday spend time making top-quality code available for everybody.
There is a scar on his head where they installed the RAM.
Same as earlier😑
Get Paul Smolensky on your podcast!
Not many know him now, however long after he is gone , his name will get carved into history as the father of AI (For the better or worse) .
This guy sounds half American and half Indian.
LUL
He is indian man
Tell him Pytorch is better than Tensorflow.
But is it really?
would love to see bill gates on this channel
Abhijit chavada guy found here great 👍
Brother 💪