The drill bit was probably dead by the time you got to the bike piece. The expensive drill bits require low rpm and frequent cooling, even dipping in water a number of times for each hole extends the life tremendously. Get it hot with too much rpm and you can kill it in seconds. Typically a squeaking sound while drilling tells you the bad news.
I love epoxies and use them whenever I can, but there's just no way torque arms is an appropriate application. I did come up with one good thing to say about it. With good prep and using one on each side it's pretty unlikely that both are going to give at the same time.
The demonstration Doc did was like a 3M salesman's demo or those done to show how strong Superglue is, and I'm sure we've all seen superglue bonding failures. A more realistic test would be to bond a steel bar to aluminum, do a bunch of warming and cooling cycles, hang some weight on the bar, and then start banging on the joint with a big hammer hundreds of times to simulate each month of use. That's a lot closer to simulating what really goes on with a torque arm, especially with regen which is absolute torture with thousands of pounds of force at a 5mm radius alternating direction. Failure is all but certain.
Even simply considering the fact that a significant percentage of people will use a lesser epoxy and/or not adequately prep or cure. Almost no one appreciates the forces inside a dropout with a hubmotor until they've had a failure, and even then most don't appreciate the real risk. Spinning the axle and having to replace the motor harness is a real pain in the ass, but the real risk is if the spinning motion causes the the axle to climb right out of the dropout and suddenly the wheel is off the bike.
I got really lucky with my failure which was in the first 10ft of my first hubmotor ride. It was a front motor with a 72V20A controller. I fabbed 2 stainless steel torque arms with rectangular holes that fit the axle flats quite well. At the time I figured the 1/8th stainless on each side giving me .25" total was plenty strong. I was right that the stainless was strong enough, because it didn't deform at all. The axle torque just made it cut through the axle like butter, snapping the AL dropouts like butter. I only got lucky and avoided an asphalt face plant because it was only a couple of mph and the hose clamps securing the torque arms were able to hold the spinning axle in the slight indentation that remained at the top of the dropout.
I've used nothing less than 1/2" of steel with a tap in fit on each side since then, and regen deforms that fairly quickly. Clamping dropouts is the only solution, and I use it on every build. With Hubmonster I used 3/8" leaf spring steel to make these 20mm wide dropouts for each side. Even clamped firmly with 3/8" #8 bolts after less than 2 weeks with regen I heard the familiar clicking of rocking dropouts. It had all but disintegrated the lock washer, so now it's a hardened steel flat washer and red Loctite. If it happens again then I'll have to spring for the Norlocks.
If I get looseness with this with 14mm dropouts on a 1" axle and an extra strong clamping bolt holding over a 3/4" width of axle on each side, imagine what a joke glued on 3/8" pacman torque arms that loosely fit less than 10mm flats are.