- There really isn't anything nefarious in trying to find a good case to set precedent. That's actually the EFF's (and ACLU, etc.) business model: find the most sympathetic plaintiff for the issue you want to litigate to improve your chances in court (and the court of public opinion).
- A method that is "90% ready" just isn't ready. Not only are such numbers meaningless in such a project. It just doesn't matter, legally, if you gain such capabilities in the future. A good faith interpretation of the FBI's motivation must also accept that if(!) the FBI had any interest in the phone's content, that interest was urgent.
I also consider the EFF's and tech community's technical arguments to be rather weak and transparent attempts to skirt the core issue: we don't want backdoors for governments, period.
By suggesting technical difficulties they are trying to shift the debate from the political sphere to the technology. That's probably good for the cause, but it circumvents the mechanism we have established for policy debates, namely democracy.
The difficulties also seem to be far exaggerated: a 1-of-2 encryption scheme should be possible. There are several such schemes (like 2-of-3) working flawlessly for Bitcoin already.
Safekeeping of master decryption keys also seems to be a solved problem. Certificate Authorities, as just one example, seem to do so mostly without any hitches.