...that's pretty much how I designed the MetaTech MT-5X prototype which is in my Matrix Specialist, Violet's (#86) head.

I'm here.
Let's see if I can keep up with your math:
I'm with you on the calculation of the Caliban.
Okaaaaay. I'm not getting 11k. Now see, first off, I probably have you multiplying the base signal 5 price (1,000) by 4(the response), but it makes little sense for your signal to be dependent on the response, so it should probably be multiplied by 5 (it's own rating), which would bring the total to 13,000 , but that really doesn't change the issue at hand. That's still much cheaper than buying a Caliban.
Of course, using the standard upgrade rules you can get a Caliban for 3,000Y instead of 8,000Y, so I'm not creating a new problem, so much as I am making an existing problem worse.
Off the top of my head, maybe customization costs should be doubled again, on top of what I've already suggested. It'll still be cheaper than buying hardware off-the-shelf, but then again, do-it-yourself should be.

I added these rules into an already-running campaign the first time I used them, so this didn't really come up. I suspect the problem could be fixed with some constant multipliers. I could've put everything into a massive spreadsheet and tried to rebalanced the costs, but I didn't. Partly because I wanted them to be cheaper anyway. I was just trying to get someone other than the "hacker" to a point where they could feel technically savvy, like someone who lives in a computerized world, without them having to go to extra effort, 'cause my players won't. I posted my rules a couple times in relevant discussions, but I'll freely admit they haven't had as much playtesting as they should, and they could be better.

To be perfectly honest, I'm having a bad week. Bear with me if I'm a pain in the ass, and not exactly 100% mentally/gramatically. If you have any suggestions, I'd love to here them. I know my rules have room for improvement. Maybe I'll play with that spreadsheet.

This is somewhat true. But that was part of my inspiration for the rules. What I was seeing at my table was the hacker buying every program, cracking them all sooner or later, and eventually giving them to every member of the team. Now everyone on the team has the same programs, and only one character invested BP into them. I saw that as a bit of a problem. By putting all the costs into the hardware, you actually get what you pay for.

On the one hand, you're noticing that commlinks have gotten cheaper overall, in a lot of cases. On the other hand, people are complaining that they've gotten too expensive. It's a compromise. I tried to keep my house-rules as low-impact as possible, but some things have to change a teensy bit.

Overall, it makes being a hacker cheaper in terms of equipment, but you're not actually saving on BP as much as it seems, because now you're forced to buy a higher Logic than you would under core rules, so the total BP cost doesn't change as drastically as it looks from looking at the equipment, because of the new implicit need for high logic.
And if the character has a genius IQ to begin with, then yeah, I don't really have a problem with them being able to branch into hacking cheaply/easily.

If the issue is that people want to play "non-hackers" with no skills, no brains, no programs, but a SOTA super commlink to give them uber-worry-free information security with no investment on their part, well, I'm a tad unsympathetic.

You get what you pay for. If you don't want to be a hacker, get a cheapo commlink and don't put anything important on it. I assume that a commlink can be subscribed to another commlink to let the team's dedicated hacker watch over everyone's commlink anyway, so it's not that big of a problem. (sure, I let them put all their eggs in one basket like that, if they want to.)

One thing I will say that is nice about my system: I wrote this before Augmentation came out. When it did, every single bit of hacking-related gear fit into the new system without any modifications of the new gear or of the house-rules. I certainly don't expect 100% success with Unwired, but I expect to have to alter very little.

Okay, that's enough defending my system, let me switch sides:
The costs of things are all messed up.
I've never been 100% happy with what you do when a program rating is called for. Initially, the rule was "use Response or System, as appropriate". As appropriate? What the hell is that supposed to mean? You're replacing real rules with house-rules-of-thumb!? Of course, System is limited by Response, and software is free, so System will always be equal to Response, so why is it even there?

Okay, so what was I thinking there?: I was thinking that the OS's in the book are free, but they're all low-mid range, there's nothing super high-end, so people are still going to be paying for those high-end custom System and Firewall. Which means that Response and System may not be even. Also bear in mind, just because Runners habitually play the numbers game and choose optimal solutions, the real world frequently doesn't. Finally, I decided to leave it alone in the interest of forward-compatibility. If I start doing things like removing System I stand very little chance of working with any new material. Since Response is the #1 limiting factor on everything I left it alone and made required program ratings always go to System, to keep things simple and to help restore the importance of System.

But that's probably the #1 iffy thing about my system, the way I see it.

Ryu: Interesting solution. Personally, I don't desire the effect of no one saves money, some spend more, but if I did I think that's a good fix.

Dang, that post got long and ranty. Am I making any sense?

One possible way of implementing SOTA rules that would be consistent with both methods would be to periodically require some kind of test. This test would represent the time and knowledge the character puts into keeping their things up to date. I'm thinking datasearch would be appropriate.
If they roll well, they don't owe any "SOTA fees". If they roll poorly, they have to contribute a little money into keeping themselves up to date. (Perhaps higher lifestyles should give one or two free hits on this roll) Finally, anyone who doesn't pay whatever they're left owing has a non-SOTA penalty of -1 (or more, if you like) applied to all matrix actions until they get their stuff updated.
Just a suggestion, I'm not going to work out the details 'cause I don't want SOTA rules, so this is probably as far as I'm going to go with the idea.

As for your second idea, I think it's a fine idea. As for why I didn't do something like that in my rules: in an effort to keep changes minimal I wanted to keep the number of programs running vs. response degradation mechanic in place and unchanged. (more of my forward-compatibility policy)
If you're doing a new system from scratch, I think what you're suggesting (if I'm understanding it, which I'm not sure of) would be a really interesting way to do it. I think something like that would result in non-hackers having either a fairly balanced 'link, or perhaps one with a heavy bias towards defense, with hackers tending towards having multiple links for multiple types of jobs. It's a neat idea.

Link? I tried a search, couldn't find it.