Yeah that seems rather ridiculous for tritium as I've just been testing, it's set so that they convert at a rate of 1 + heat/3000 progress each time they get hit with a neutron. However as they're set to requiring zero progress for some reason anything that hits them with neutrons will instantly convert them. So for example take any reactor design and replace a neutron reflector with a lithium cell, the reactor should be cooler so produce less energy but it will convert at a rate of one lithium cell to tritium cell per second per slot. That will be trivial to run and keep in a cool reactor with thorium being superior due to the higher durability and it will easily produce more tritium then you know what to do with, the limiting factor will become lithium as mentioned but that's easy to get from dirt as previously covered in the thread. 1 lithium = 2 tritium too so while you'll need 5 slots converting lithium to tritium that only requires 5 tiny lithium dust a second. So 4 MV Electrolyzers converting clay to lithium which will consume 4 clay dust a second, as it's 18 dirt per clay dust you'll be needing to centrifuge 72 dirt a second which is where the real brunt of the work comes in as that's 864 LV centrifuge recipes at once. Of course it can be done better obviously, a mere 54 IV centrifuges could do it for instance.
So probably best to sticking to just synthesising clay if you're using UU.