One may say that predictions are dangerous particularly for the future. If the danger involved in a prediction is not incurred, no consequence follows and the uncertainty principle is not violated.
Secrecy, once accepted, becomes an addiction.
Two paradoxes are better than one they may even suggest a solution.
There is a time for scientists and movie stars and those who have flown the atlantic to restrain their opinions lest they be taken more seriously than they should be.
I hate doubt, yet I am certain that doubt is the only way to approach anything worth believing in.
In our educational institutions applied science may almost be described as a “no-man’s land.”
The best weapon of a dictatorship is secrecy; the best weapon of a democracy is openness.
Today’s science is tomorrow’s technology.
Physics is, hopefully, simple. Physicists are not.
Society’s emissions of carbon dioxide may or may not turn out to have something significant to do with global warming-the jury is still out.
My experience has been in a short 77 years that in the end when you fight for a desperate cause and have good reasons to fight, you usually win.
Life improves slowly and goes wrong fast, and only catastrophe is clearly visible.
I tried to contribute to the defeat of the Soviets. If I contributed 1%, it is 1% of something enormous.
When Columbus took off, the purpose was to improve trade relations with China. That problem has not been solved to this very day, but just look at the by-products.
If anyone wants a hole in the ground, nuclear explosives can make big holes.
I believe in excellence. It is a basic need of every human soul. All of us can be excellent, because, fortunately, we are exceedingly diverse in our ambitions and talents.
There is no case where ignorance should be preferred to knowledge – especially if the knowledge is terrible.
The eyes of childhood are magnifying lenses.
No, I’m the infamous Edward Teller.
Could we have avoided the tragedy of Hiroshima? Could we have started the atomic age with clean hands? No one knows. No one can find out.
A state-of-the-art calculation requires 100 hours of CPU time on the state-of-the-art computer, independent of the decade.