Right, and JoelKatz is only asking to verify information that he *already* knows and which is entirely public at this point.
That's not quite true. I would be asking them questions whose answer I know within a range but don't know for sure.
If there is anything about this asic machine that is truly a secret, it probably doesn't need to be disclosed in the exchange. He is merely asking to confirm that this group has the relevant know-how to manufacture such a machine (or even build a prototype).
That is essentially correct. There are certain thing you have to do in the process of designing, building, and testing an ASIC. The questions I would ask would be things they would definitely know. An example question might be "how many gates or equivalent does the ASIC have?" Or "what is the die size?"
Now, I don't know how many gates their ASIC has, but I can calculate how many it would have based on other information (such as the number of hashes and the clock speed). I can then check my gate count against theirs. Similarly, I can calculate the die size based on other information and check its consistency.
Those are very rough approximations really, i think you are well aware of this.