Shultz was another to succumb. His grandson, Tyler, joined Theranos but left when he understood the bluster behind its promises. Rather than heed his warnings however, his grandfather turned him away, communicating with him only through lawyers. As Shultz continued to side with Holmes, Tyler was not invited to his 95th birthday party.

It was a far cry from Holmes’s own birthday party – her 30th, in 2014. There the guests included Shultz and Kissinger; the latter even composing a limerick which he read in her honour.

By backing her, such men were also helping to right a palpable wrong, championing a female entrepreneur in a tech industry long damned by its scarcity of women leaders. With her pale skin and pale blue eyes set off by her black polo-neck, she even dressed to suggest she could become a female visionary to match the founder of Apple. “You start to realise you are looking in the eyes of another Bill Gates, or Steve Jobs,” sighed one male professor on Theranos’s board.

At the heart of her trial, then, was the conundrum of whether Holmes was so convincing, so driven, so messianic on behalf of her product because she was a con-artist, pure and simple, or whether at some level because she, too, was sucked in by her own delusion, with hope morphing into certainty in her mind that even if Theranos wasn’t quite ready to change the world today, it would tomorrow.

The latter was the heart of her defence. Day after day on the stand, Holmes – still only 37 – presented herself as a starry-eyed do-gooder, a little naive perhaps, but in the best possible cause: the “really big idea,” as she branded it, of miniaturising and automating blood diagnostics. Prototype machines, she insisted, were created after “years [of research] with teams of scientists and engineers”. Those prototypes were “really good” she recalled. But crucially, they could do little that was not already available.

As with conventional machines, they struggled with small samples like blood drops rather than blood taken from a vein. Indeed, Theranos ended up processing some samples using existing commercial technology. But, Holmes insisted, the promise of radical benefits never evaporated in her own mind. As such, she was not misleading investors, she maintained, but simply talking “about what this company could do a year from now, five years from now, 10 years from now.”

This potential future payoff seemed to excuse any deception, even that of plastering the names of big-pharma companies including Pfizer over marketing material, as if they endorsed Theranos, which they certainly did not. “I wish I had done it differently,” she said.