On the Titan’s second deep test dive in April 2019—an attempt to reach 4,000 meters in the Bahamas—the sub protested with such bloodcurdling cracking and gunshot noises that its descent was halted at 3,760 meters. Rush was the pilot, and he had taken three passengers on this highly risky plunge. One of them was Karl Stanley, a seasoned submersible pilot who would later describe the noises as “the hull yelling at you.” Stanley was no stranger to risk: He’d built his own experimental unclassed sub and operated it in Honduras. But even he was so rattled by the dive that he wrote several emails to Rush urging him to postpone the Titan’s commercial debut, less than two months away.
The carbon fiber was breaking down, Stanley believed: “I think that hull has a defect near that flange that will only get worse. The only question in my mind is will it fail catastrophically or not.” He advised Rush to step back and conduct 50 unmanned test dives before any other humans got into the sub. True to form, Rush dismissed the advice—“One experiential data point is not sufficient to determine the integrity of the hull”—telling Stanley to “keep your opinions to yourself.”
he did one test to determine if the hull could withstand the pressure, it failed, and then he says one data point inst enough to determine the integrity of the hull of of one failed data point.
Back when this incident first happened, there was a few articles posted that day, and someone posted a reply on one of the threads, but I’m too lazy to find it now.
Basically they said they worked at a university research lab where Rush was conducing pressurized testing on the capsule (seals and such). They’d conduct half a dozen tests in one day, have one “success” and Rush’s team would call it a “success” and move on to the next test.
i remember seeing something to that effect as well. move fast and break things isnt a bad way to do things, til you put the lives of others on the line. making many iterative designs and testing them mightve proven him right long term, highly unlikely but whatever. instead he went with the first that would never have been safe.
he did one test to determine if the hull could withstand the pressure, it failed, and then he says one data point inst enough to determine the integrity of the hull of of one failed data point.
Back when this incident first happened, there was a few articles posted that day, and someone posted a reply on one of the threads, but I’m too lazy to find it now.
Basically they said they worked at a university research lab where Rush was conducing pressurized testing on the capsule (seals and such). They’d conduct half a dozen tests in one day, have one “success” and Rush’s team would call it a “success” and move on to the next test.
i remember seeing something to that effect as well. move fast and break things isnt a bad way to do things, til you put the lives of others on the line. making many iterative designs and testing them mightve proven him right long term, highly unlikely but whatever. instead he went with the first that would never have been safe.