Like most technologies, improvements in lighting seem to be following a sigmoid curve. Here’s how that looks:
Basically, things start out slow and accelerate upward, getting better faster and faster. Toward the middle of that curve, dramatic change can happen pretty quickly. But then things start slowing down, because eventually the realm of possible improvements is smaller than what has already been done.
For example, flashlights just a couple decades ago used incandescent bulbs which produced only 5 lumens per Watt. Then LEDs started getting good, and pretty quickly we had lights which emit 100 lm/W. That’s a 20X improvement, and it happened in a short time. That’s the steep part of the curve shown in the graph. Some lights even get 150 lm/W now.
However, that will never happen again. The exciting part of the curve is behind us.
The theoretical maximum efficiency for white light is 300 lm/W. So even if a flashlight improved to the absolute limit, it would only be 2 to 3 times better than what we have now, not 20X. And we won’t actually be able to reach 300 in a practical sense, because nothing* is ever 100% efficient.
With lighting, we’ve already reached the point of diminishing returns. There is still more improvement to be had, but it’s slowing down. There simply isn’t as much room to improve, compared to what has already been done. The low-hanging fruit is already picked.
(*) Nothing except heaters. Heaters** are 100% efficient.
(**) Except for heat pumps, which can technically be more than 100% efficient***, because they move heat instead of just generating heat.
(***) Heat pumps are cool. … Literally.