What current are you trying to drive the LD with? The problem I see is that the variable resistor has to take the full current of the LD (since the current into the JFET gate is zero), so it's going to have to be a low value, like 10's of ohms. At that point you may as well not even use the JFET, becuse the voltage and current won't swing that much when you have over 2v of headroom anyway.
I don't see what the capacitance is for. My guess is 10uF should be okay -- at 50mA, it will charge and drain in about 1 ms, which is plenty short not to be annoying. I'd guess you could go as high as 100uF without really affecting turn-on/off delay, but I'm not sure why you would want or need to.
I've seen all designs incorporate capacitors due to battery voltage spikes.
The problem with lithiums (as well as all batteries to some degree) is the voltage drops with usage, and I'd rather the output of the laser didn't drop with usage (as would be the case with a single resistor design)
I built it today and the JFET is dropping .8V and the resistor dropped a little less than 2V with the resistance value at about 15 ohms to achieve 120mA
Is there any chance of using some sort of regulator? Driving a diode directly like that is never the best idea - either you'll be overdriving it on the high end of battery charge, or underdriving it on the low end (or both!). With a regulator you could stop worry about the cap, and your output power would remain stable independent of your input (within reason and the specs of the regulator of course)