François Chollet+ Your Authors @fchollet Deep learning @google. Creator of Keras, neural networks library. Author of 'Deep Learning with Python'. Opinions are my own. Jul. 13, 2019 1 min read + Your Authors

In late 2014, I said Go would be solved within the next 5 years (with the expectation that DL would play a role). AlphaGo certainly happened unexpectedly early, but not 20 years early. A few years early.

In the other direction, through 2015-2017 I was correctly tweeting that L5 self-driving was *not* around the corner. In 2018 I made a bet predicting robotaxis with no safety drivers would not be deployed commercially to the general public before 2023

If you know what *abilities* (potential) current AI techniques possess and what abilities are required for a given task, you can roughly time short-term applications

What you can't do is make 10+ year predictions, because what new abilities will be available is v hard to predict

You can follow @fchollet.


Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Since you’re here...

... we’re asking visitors like you to make a contribution to support this independent project. In these uncertain times, access to information is vital. Threader gets 1,000,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Your financial support will help two developers to keep working on this app. Everyone’s contribution, big or small, is so valuable. Support Threader by becoming premium or by donating on PayPal. Thank you.