'We thought it was not working because we didn't have quite the right algorithms, we didn’t have quite the right objective functions. I thought for a long time it was because we were trying to do supervised learning, where you have to label data, and we should have been doing unsupervised learning, where you just learned from the data with no labels. It turned out it was mainly a question of scale.'
Wolfram MathWorld on the tetrahedron.
Obliquely related to previous link.
Party City and the helium shortage.
Helium, like hydrogen, is a lighter-than-air gas - and consequently it's great for filling balloons. But unlike its neighbor on the periodic table, helium is non-flammable - so, no Hindenbergs.
That's good, except that helium isn't just non-flammable - it's a noble gas, which means it doesn't form any compounds at all. Hydrogen is everywhere and it's in everything; heck, even water is two H's with an O. So you can store hydrogen by storing its compounds. But helium? It's just helium. There isn't any chemical or molecule you can break down to get helium - and when it goes, it's gone.
This is a problem, because helium has far more serious uses than party balloons ...