Gary -- I think it is not correct to accuse Andrew. Some his statements are over the top, but he is doing the noble work of training thousands of AI engineers, that'll eventually in a couple of decades translate this hype to reality. Until, that time paychecks will be attractive.
overhyping is perhaps getting to be a habit for @AndrewYNg, in radiology, cars, and AI more generally. this thread below is consistent with what I have noted elsewhere with respect to his misleading statements about http://drive.ai and what AI can do in a second.https://twitter.com/guzmanjorge_/status/1000581060059639808 …
-
-
-
i think the overhype is potentially dangerous, and i may be writing more about why soon.
- 1 more reply
New conversation -
-
-
Perhaps the entire conversation *is wrong*, in the sense that radiology is not something that humans do in a second. Reads on even routine mammography scans for example can take several minutes, and often require detailed comparisons across time and bilateral symmetry
-
Nah, it is definitely something humans do in a second (radiologist speaking). The problem with most radiology is a field of view issue. Humans use 30+ 1 sec views to read a single view mammo, since we need to look at small 2-5cm squares at a a time. Models parallelise attention.
End of conversation
New conversation -
-
-
This Deep Learning propaganda is cutting off great startups from funding. AI does not equal deep learning. :(
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Gary, go easy buddy. Andrew's a LEGEND.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
This Tweet is unavailable.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.