They mention the product lifecycle being short as a primary reason for not wanting to bring on new employees. If it takes someone 12 months to get up to speed in an 18 month product lifecycle, it is too costly.
I've got to wonder what kind of industry takes 12 months to get up to speed unless you've got workers coming in with absolutely no training and education. For most programmers I know, the worst case ramp up time is around 3 months. That is often with a project that will never be profitable, much less make it for 18 months.
There are other fiscally attractive reasons to hire young so I'd think if training was the only problem, we could solve that problem. The problem I see with regards to education in most companies is that there is simply no one on staff who can do the training effectively or is given the time to do so. Perhaps this will open the door for the return of a mini trade school in the form of an app.
Hardware engineering can easily be 12 months out of 18. There are a sundry of reasons:
a) Moore's law. The technology your school is using for hardware design is easily 5-6 or 10 years out of date. This makes it easier to teach and cheaper to work on, but the "real world issues" like heat, leakage current, etc just aren't addressed at the same level.
b) Intel has its own RTL that is not taught in schools. Some technology companies also have proprietary in-house technologies (like Cachet, or Wasabi).
c) Ripple effects. Analog components might change/fluctuate and this impacts the entire design. Or a new antenna placement might merit new design, etc.
d) Scale. An undergrad in hardware engineering typically builds a 5 stage pipeline processor that supports 2-4 hardware interrupts and a memory controller. This is enough to run linux on an FPGA. You've engineered a computer! But a Core 2 Duo has 23 pipeline stages. Vector units. Out of Order execution. 96% accurate branch predictor. Each one of those things I've mentioned could be the focus of a masters or PhD thesis. Getting hired, you'd be expected to pick up the logic of all of them in about six months.
e) Minimization. More systems are being done by less chips. This is what is called System-On-A-Chip design by some. Previously, you'd need to know how to design a CPU. Now you need to know how to integrate a CCD into the CPU.
This is off the top of my head but i'm sure you get the point.
I've got to wonder what kind of industry takes 12 months to get up to speed unless you've got workers coming in with absolutely no training and education. For most programmers I know, the worst case ramp up time is around 3 months. That is often with a project that will never be profitable, much less make it for 18 months.
There are other fiscally attractive reasons to hire young so I'd think if training was the only problem, we could solve that problem. The problem I see with regards to education in most companies is that there is simply no one on staff who can do the training effectively or is given the time to do so. Perhaps this will open the door for the return of a mini trade school in the form of an app.