He also made a similar forum post here.
One of the frameworks he uses to value NMR is NPV. The formula for this is \mathrm{NPV} = \sum_{t=1}^n \frac{R_t}{(1 + i)^t} , where R_t is the return at period t and i is the discount factor which includes alternative investments. Because of the payout factor we are presently seeing R_t increasing linearly in time at best. In the meantime I would strongly argue that the key alternative investment for this crowd is bitcoin, etc. The factor, t, in the exponent of the denominator means that the linearly increasing numerator is getting crushed by rapidly increasing bitcoin prices.
OTOH, he discusses extremely speculative other use cases for Erasure in the far future. Those in principle could contribute to R_t. I say speculative because I have not seen anything to indicate that there is any effective use case for the Erasure protocol other than for staking NMR in the tournament itself; although I would say that even that use case is not yet entirely proven. The speculative nature of all of those other use cases must be discounted and compounded into i even before computing the sum or just simply ignored as a contribution to R_t in the numerator.
IMHO, the most convincing discussion has to do with the treasury running out and Numerai being forced to buy NMR on the open market. Yes, he does mention this. But throw a dart in the map estimates give a time period of 10 to 20 years before that happens. In the meantime some of you will have become bitcoin billionaires and NMR will just start to be picking up speed. Wouldn’t that be an enormous opportunity lost?
I think this reference is really a good argument for starting to purchase back NMR now. Even small buybacks would point the rudder of the NMR juggernaut in the right direction; strongly reducing the negative effects of the discount factor, i.