5 ms·
I suspect Peter should have left Google long ago. Google's approach to ML/AI probably is not what Peter really liked, judging from his past writings and books.
by justicezyx 5y ago
I suspect Peter should have left Google long ago.
Google's approach to ML/AI probably is not what Peter really liked, judging from his past writings and books. There is certainly a strong sense of symbolic and rule-based intelligence, based on my limited reading and even more limited understanding.
But I guess Google paid very well...
In reality, Google probably should sponsor Peter and let him work at Stanford...
- nl 5y agoGiven that Norvig was one of the authors of "The Unreasonable Effectiveness of Data"[1] and his famous statistical spell checker[2] are both about showing how rule and symbolic intelligence doesn't work I don't think your thesis is quite correct. [1] https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/35179.pdf https://static.googleusercontent.com/media/research.google.c... [2] https://norvig.com/spell-correct.html https://norvig.com/spell-correct.html
- vl 5y agoWhen I was at Google I never heard Norvig’s name in connection with anything. At any internal AI conferences Jeff Dean would be present, Kurzweil, etc., but never Norvig.
- nothrowaways 5y agoRule based and all was long ago before data was so much abundant. Or in other words before big data era.