Machine learning is not the future – Google I/O 2016

AI and machine learning have been a topic of science fiction for years, and they’ll continue to be. But this talk isn’t about science fiction—it’s about all the amazing new things that machine learning is already doing for us today. It is already around us, in many many products. See where machine learning is used today and how you can use TensorFlow to make it a part of everything you do.

See all the talks from Google I/O 2016 here:
Watch more Android talks at I/O 2016 here:
Watch more Chrome talks at I/O 2016 here:
Watch more Firebase talks at I/O 2016 here:
Subscribe to the Google Developers Channel:

#io16 #GoogleIO #GoogleIO2016


  1. maninm says:

    So interesting field presented by so boring Speaker. I think with this type of presentation we need someone who is really excited about innovation and knowledge and not someone who is there only to speak.

  2. kupruser says:

    So i sat through 40 minutes of monotonic speech with a strong accent just to hear “ML is the future” at the end. Oh well…

  3. Dustin Rodriguez says:

    One question I have about the contemporary approaches to voice recognition through machine learning: Once the network is trained, why can’t you produce a ‘frozen’ version of that network which can be loaded onto end user devices so that voice recognition can be done without sending all of the audio over the Internet to servers owned by the company? I understand the training requiring tremendous amounts of storage, compute power, etc, but once the network is trained, all of that goes away. Is it purely business motivations that prevent offering voice recognition that doesn’t require submitting audio to third parties?

  4. Jeff Kriske says:

    I’m not sure I understand the comparison with CuDNN when CuDNN is required during the compilation of TensorFlow itself. The graph is a bit misleading. Maybe he is suggesting that using TensorFlow with CuDNN is faster than CuDNN alone.

  5. Syafiq Firdaus says:

    Hats off to those audiences. The actual talk should be about ‘Patience Learning – it’s not the future, but it really is’.

  6. Ed Hendel says:

    This is such a misleading title. It makes it sound like machine learning is overrated and fading, but really the message of the talk is that machine learning is the *present*, not merely the future.

  7. Steve Bergman says:

    It’s not “click bait” since “Machine Learning is NOT the Future” is the title of the talk.

  8. arun satyarth says:

    This video was not a complete waste of my time. It taught me about the new technology in google called ‘Google ClickBait’ !!

  9. HIMANSHU SINGH 14BCE0166 says:

    The only takeaway from this video is
    Found it to be useful.

  10. L.J. Huang says:

    It is a designed contradictionary topic, i.e,to say ML is not the future is to emphasize it will be the future……

Comments are closed.