Two things to consider:
1) They claim they tested 100 algorithms over 2 years before they settled on one to bring to market. That's pretty much the definition of 'harder than it looks'.
2) This product requires sufficient cheap, powerful, and small accelerometers, processors, batteries, and sensors. All of which have really only come together in the last 4-5 years. Any engineer who made the connection in the 90s might have at most whipped together a prototype before doing the calculations and realizing that that it just wasn't viable.
Conceptually, it's simple. I've personally seen 2 senior year engineering student design projects that more or less do the same concept - do feedback control on the spoon. They even had -ok- looking prototype demos. You could look at their spoon vibrating less, and you could look at their metrics, and see their claims, but then you would really look at it, and think 'man, that 40% reduction in amplitude sure looks sweet, but does it actually do enough?'.
I can straight up tell you that nearly all 'simple' ideas become retarded complicated when they moosh up against a human being (or mother nature...), and 'complicated' ideas become even worse.