You're probably thinking of heat engines. Extracting 100% efficiency from a heat source is impossible. That doesn't mean that near-100% of non-heat energy can't be converted to another energy type. Eg resistors pretty much convert electricity to heat at a 100% "efficiency" rate..

Hm clearly I've got some more learning to do about this. Thanks!

In fact because of thermodynamics over 100% efficient LEDs are possible. This was practically demonstrated some years ago: https://phys.org/news/2012-03-efficiency.html

Converting light to electricity on the other hand cannot reach 100% efficiency.

It needs to be pointed out that the diode isn't converting over 100% of the electrical power into optical power. The extra optical power comes from Peltier heat drawn from the semiconductor lattice. The LEDs only had ">100% efficiency" when heated to 135°C. At 25°C the electrical efficiency was less than 1%.[1]

[1] Fig 1 (a) https://journals.aps.org/prl/pdf/10.1103/PhysRevLett.108.097...

That still counts. Ambient temperature and pressure are necessary for just about any 'engine'. In fact, extracting heat from the environment to do something useful is more impressive than just converting one form of energy into another.

Please explain how this is. It seems like something could convert all energy to light without generating heat and while conserving total energy & mass...

Well, the law they're referring to states that rather than just conserving energy and mass you also can't have a process that lowers entropy. Although it's unclear how this is relevant because, if anything, converting kinetic energy to light increases entropy, or at the very least doesn't decrease it.

This rules out superconductors, does it not?