Ilya Sutskever | AI will bring social progress and replace certain things|AI has entered a new point
28,808
Published 2024-06-25
All Comments (21)
-
This interview is 2 years old, it was at Scale AI
-
OLD INTERVIEW - THIS IS NOT RECENT
-
Still a great interview, but it was lame not tagging it as old.
-
This video is like a rollercoaster ride of emotions. Gripping from start to finish!
-
We need a new interview plz
-
π§π·π§π·π§π·π§π·ππ», Amazing interview, very interesting, thanks for sharing!
-
People don't necessarily generalize very well either. (My thesis topic)
-
any significant technological advancement will lead to both jobs being lost but new jobs being created, you just need to be prepared and adapt to the changes
-
all ML is informed by biology but none is trying to "replicate" it.
-
I am super optimistic on the future of AI. Yes, we are going to lose many current human jobs. But AI will also help to create even more incredible jobs that have never been imagined today. People's lives will forever be changed for better. Welcome to the Fourth Industrial Revolution - the AI and Robotics Revolution.
-
I love looking back at history, things were so backwards when this was filmed
-
As Ilya points out above, when a human learns something, they learn it multi-modal - through hearing, speech, touch, and yes, language. My intuition (same as Ilya's above) is that this will lead to a deeper learning. To me this is just common sense. Curious however whether each modality can be optimized through different types of neural nets.
-
Ancient video. You should have made this clear. Bye.
-
When asked about technology taking over human jobs and creating unemployment the AI said: 'Are you humans Stupid? All you have to do is make it financially worthwhile for people to SHARE the jobs left and enjoy your LIVES working much less...there would never even be such a thing as unemployment just the ability for you all to enjoy your lives working ever increasingly less.'
-
Crazy how two years in AI is like 20 years in other fields. Now researchers are saying that it's not the amount of data, it's the quality of that data. And that makes a lot of sense. It's faster to train, cheaper, and it wasn't using a lot of the lower-quality data anyway.
-
Yes! Someone lined my boy Ilya up. His shit was lookinβ rough there for a minute. π
-
It's annoying that these reuploaded interviews don't state up front what year is it from. Here is your thumbs down sir.
-
Using ever more data is useful in the context of his first student, of course give it a bigger text book to memorize with more examples, and it'll do better in exams, but in the context of his second student, who's better at generalising, it would need to do better with less data
-
Not just text and images. Please do include experiential learning. That way we validate our bookish learning. This is how it can learn more generalization and will require very less data.
-
I told someone recently that eventually all businesses will be AI (meaning no humans). They seemed skeptical.
