It is defined as one disintegration per second, it has nothing to do with mass.

So a terabequerel is 10^12 disintegrations per second. Not that that is such a huge amount; 1 mole of pure C14 (14 grams) contains 6.02 x 10^23 atoms, so that level of radiation means a mole of pure C14 would be radiating (to start with) at a rate that would last 6.02 x 10^11 seconds. Actually, its much less than that, since every disintegrating nucleus subtracts from the total amount available. C14 has a half-life of several thousand years.

The idea that C14 could be recovered in large quantities from radioactive waste graphite solves the availability problem, but you still have the dilution issue to contend with. There are only 14g (1 mole) of C14 in 10^6 g of waste graphite (1 tonne)--a tiny fraction. and it still is very difficult to separate from the other isotopes of C due to their identical chemical properties.

This sounds like a promising power source, but only one with very limited applications: long-term low energy requirements for very low-mass devices. A good application would be in chip-sized nanosatellites or other tiny bits of machinery, such as medical nanobots.

As nanotech progresses, no doubt we will be building tiny microscopic powered devices, but it would be of little use for macro gadgets like radios, electric motors, heating or cooling coils, pumps, light fixtures, computers, servos, sensors etc.