In terms of interaction design, they pretty much have gotten it right, which really means something to me, because I know the TiVo folks have had some issues to work around. Most significantly, they have had to deal with Gemstar patents, and the desire to avoid them due to the cost of licensing.
The TiVo is already too expensive at $400 for the "80" hour model (which is actually 37 hours at an acceptable quality level due to the lossy mpeg compression). So, avoiding more cost due to Gemstar's patently evil business plan (which is to have patents, and then litigate value out of them) is a good thing. Sure, there's a $50 rebate, but I have to mail that in, and fill stuff out, and lowers the VALUE of the $50 and increases my annoyance at the TiVo product. It dirties the product, and TiVo should knock of this mail-in rebate bullshit.
This is ok though, but keeping the cost of the box down. This is one of those decisions that we usability dorks need to make from time to time...
Is this compromise on ultimate usability worth it in terms of cost saving?Those cost savings may be realized in MANY ways. You may be able to avoid a patent infringement, or avoid more work (which may be billable against your budgets), or reduce the amount of processing a system may need to do by making the user do a little more work. However, these desicisions are extremely difficulty to make as you get further away from obvious liscense costs to the more ambiguous realities of what your hardware can handle. So far, TiVo seems to be making the smart decisions.