Deep learning maximalism: a dialogue

A: So these guys are very confident that they will develop superhuman AI within a few years

B: Cute, I guess that neuropsychology degree is really paying off for them

A: No, no, they don& #39;t have any background in neuroscience or psychology
B: Oh ok. I guess they must be the mathematician or physicist type then. That would be *so* in character!

A: No, they don& #39;t really have any strong math background either
B: Oh. Well, it& #39;s not like you really need a PhD in any of these fields, you could just read a lot of books and papers, and catch up...

A: No, here& #39;s the thing: they don& #39;t read books. Not even AI/ML books. They only read ArXiv preprints

B: ...

A: ...are you getting it now?
The moral of this story is this: if you set out to solve ambitious problems, what you need most isn& #39;t "billions of dollars in funding", you need an attitude of doubt & perpetual learning. You need to read books. You can& #39;t be supremely confident in the things you think you know.
Especially if you happen to know nothing and to be horribly wrong in your key assumptions.

Being ignorant and wrong isn& #39;t the end! It could be the beginning. But if you don& #39;t read and if you display absolute intellectual arrogance, your journey is already over before it started.
You can follow @fchollet.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: