Gene told me the opposite.
Anyway, it doesn't matter what he says in private conversations. What matters is what he recorded in his report.
"""
A third attempt was made wherein the capacitor bank was allowed to
charge fully before the load was applied. This took less than a minute. In
this attempt, the efficiency on connecting the load to the capacitor bank
was about 150%. It slowly degraded over a period of about 5 minutes to
unity, and then on down to 70%. When this test was repeated with an
adjustment to the output switch, testing was forced to a stop.
"""
So was the energy applied during the time the capacitor was charged up more or less than the energy during the 100%+ period that followed? Did the decay from 150% to 70% follow the pattern you would expect from a capacitor discharge? To answer either of these questions, one would need to see the readings of power over time, which Gene does not report. Why did he neglect to provide this data? Because the gizmo broke while he was using it.
The entire basis for optimism in this gizmo is based on it breaking.
Edited to add: This is a general rule. When money is involved, never believe anything that anyone says off the record.