With the amazing amount of stuff coming out of Google at the moment I think I need to start getting a handle on this stuff. It seems that it’s going to be key feature of software development over the next decade or two to be able to work with these tools.
Mostly looks like part-of-speech parsing which has been available from University servers for decades. While there are a lot of compelling aspects of English conversational interfaces such as dirt cheap hardware: mic + speaker + processor + embedded linux + wireless…
Prolog in the 1980s didn’t produce many cool toys to play with. Pre-trained neural networks (and the tools to train your own) on the other hand are making impacts on the way online services are being run today. We all expect it from Google services for them to make intelligent inferences from our queries. This will become standard practice.
John Hardy Developer 
Heh. Do you actually plan to use it? Could be useful for certain chat assistants perhaps.
LikeLike
With the amazing amount of stuff coming out of Google at the moment I think I need to start getting a handle on this stuff. It seems that it’s going to be key feature of software development over the next decade or two to be able to work with these tools.
LikeLike
Mostly looks like part-of-speech parsing which has been available from University servers for decades. While there are a lot of compelling aspects of English conversational interfaces such as dirt cheap hardware: mic + speaker + processor + embedded linux + wireless…
LikeLike
Michael Tufekci — reminds me of Prolog back in the 1980s.
LikeLike
Prolog in the 1980s didn’t produce many cool toys to play with. Pre-trained neural networks (and the tools to train your own) on the other hand are making impacts on the way online services are being run today. We all expect it from Google services for them to make intelligent inferences from our queries. This will become standard practice.
LikeLike