I doubt you would notice if there was any difference in distortion. Grid-leak works perfectly well, the grid/cathode form a parallel diode anyway.
It doesn't state what the B+ voltage is but I would be cautious about the rectifier diode types, personally I would go for ones with higher PIV rating, maybe 1N4006 or 1N4007.
The anode supply to the first stage is unconventional, normally the load resistor goes at the top end of the choke rather than the anode itself. If you move it you might get a marginal increase in gain. Also, I can't see any advantage in the first stage cathode using that 10uF capacitor. The purpose is to allow the choke to block RF and let the resistor do the grid biasing but it would be simpler to just place the choke in series with the resistor and it would avoid a potential resonance peak as well.
Note that in the second link, the 'power amplifier' only consumes 1.5mW (excluding heater) so don't expect to drive a loudspeaker!
Brian.