Read the literature, but don't read too much of it.
Trust your intuitions, because if you don't trust them then there's no point
'This person is either drunk or stupid' - feedback on one of his papers that went on to become significant
For creative researchers; read a bit of the literature and notice something that you think everybody is doing wrong. (I guess GH is contrarian in this sense)
Look for a problem that doesn't feel right. Then figure out how to do it right.
When people tell you that your (contrarian) approach is just no good, just keep at it.
Either your intuitions are good or they're not. If they are good then you should follow them and you will eventually be successful. If they are bad then it doesn't matter what you do.
You might as well trust your intuitions.
When you try to replicate a published paper you discover all the little tricks needed to make it work.
Never stop programming.
If you give a student something to do and they are a bad student then they'll come back and say it didn't work. They'll say this because of some little decision they made which they didn't realise was crucial. But if you give it to a good student, you can give them anything and they will come back and say it works
Read enough so that you start developing intuitions, and then trust your intuitions and go for it! - Don't be too worried if everybody else says its nonsense.
If you think it's a really good idea, and others tell you its complete nonsense, then you know you are really onto something!
One example of this is when Ramford and I first came up with variational methods, I sent a mail explaining it to a former student, who showed it to his colleagues. He told me that they said 'Either this guy is drunk or he is just stupid'. They really really thought is was nonsense.
Maybe that is partly because of how i explained it - I explained it in intuitive terms, but when you have what you think is a good idea and other people think it is complete rubbish, that is the sign of a really good idea.
See if you can find an advisor who has beliefs similar to your own, because if you work on stuff that your adviser feels deeply about, then you'll get a lot of good advice and time from them.
Read enough so that you can trust your intuitions
Notes From An Interview With Geoffrey Hinton