Depending on how the code is written, disabling the number of tries function might disable the delay too. If they are using a "Switch and Case" type statement to set the delays, sticking the "number of tries" variable to zero will always take the no delay branch.
Hashes are designed to take time, proof-of-work is the proof of security. They want the alternative interface to enter passcodes. They want to bypass flash and run in RAM. The main problem is the new interface.
The FBI request said they could do it any way they wanted to do it. It is the Judge that said to do it a specific way, and even then left it open to Apple to use a different method so long as it could achieve the objective.
The order is not written in stone. It can be interpreted in such a way that Apple does not have to risk either hardware or software getting into any other hands than their own.
Another problem is that the code in RAM is susceptible to reverse engineering.
Reverse engineering is not a concern since Apple can maintain control over the hardware because it will be at one of their own facilities.
It is not a "request" it is an ORDER, DiogenesLamp. And it requires that the FBI/government approve of the way Apple decides to do it, which would mean they would have to show the FBI/government how they will do it. Don't put words in the order that are not there. Don't mischaracterize the ORDER as a "request." Some of you are trying to characterize it as a "request" to Apple to "offer to help." PUKE!