Except the failure rate is some probability distribution. If making the laptop thinner means that the 1-year failure rate spikes from 0.2% to 0.5%, that just means Apple pushed the manufacturing too far, not that there is an inherent problem with pursuing a thinner design.