a) The fire extinguisher was not included.
b) Not enough packing. (leading do many problems with fans falling off, bad heat conduction)
c) Not shipping on the dates expected. (with no talk of compensation so far)
d) Serious problems with hosted miners. (with no talk of compensation so far)

^doesnt seem like KNC has many "inexcusable" issues, but there are 'annecdotal' (in quotes because its not widespread) accounts of:

-capacitors blowing/catching fire (primarily with a specific update 0.93)
-units hashing under 300GH or instability (good chance some small tuning or setup adjustment swill fix many of these)
- units arriving with fans or heatsinks knocked off
- units with poorly-attached heatsinks and/or poor use of thermal paste (i assume a minor issue unless whoever did it was sloppy)

a) LOL
b) supposedly being addressed. We will know in a few days
c) agreed
d) compensation was offered and in some cases received. 1 bitcoin per day not hashing from expected delivery, and 1 month free hosting.

My personal thought on this is that they made a hell of an effort, and hit some unforseen (largely due to optimism) speedbumps. So far, they have been very up on things up to sep 30th, after which they somewhat dropped the ball. According to Bitcoinorama, they've all been working themselves ragged trying to fix everything. I suspect they'll get it all ironed out to the satisfaction of most of their customers. Having followed their development since the day they posted their proposal, I will have to say that there will be some who are not satisfied no matter what. I think those few can be discounted, because there is no scenario under which they would be satisfied.

The extreme variance in hashrate is troubling, and I suspect there are hardware problems that have not been addressed by KnC, though we simply don't know what they're doing. Again, I think they will bend over backwards to make it right. These boys have pride, and skills. That's a dangerous, but usually useful combination of traits.

I blame the people who presold equipment with a given delivery date, and couldn't make working equipment by that date.

...and then people would bitch about not having a specific date. Truth is they are in a race, and the only way to win is be first.

I didn't preorder simply because I didn't have the money. I would have bet on that horse (and not bitfury, so I got it half wrong ). I think KnC has delivered well enough. As in my previous post, I think there are people who you just cannot satisfy, regardless. I would be happy with what they have done, and I'd be faking patience if I were one of the people who's miners didn't make the final announced spec. I'd certainly be on them about it, but not mean. Just "hey, here's what it's doing, what next?".

I think for the next gen of chips, manufacturers would be wiser to work totally in secret and have a consumer grade product ON HAND before opening their order books, even if it's just a small batch. It's a safer model for all involved. But there is danger there, as well, due to the time sensitive nature of this race.

the next generations will follow as:

server grade 28nm bitcoin miners.   used for data centers.  That phase is already going on behind your backs.

the next phase after that is to go toward ~14nm bitcoin miners for data center usage.

I can tell you this...there are a few people making 28nm data center miners right now.

The only other way to be competitive is create a miner product that is reliable, cheap, easy to scale,  and cost efficient.  Hence bitfury 55nm chip.

Sorry that was not me.  I am referring to about 6 companies going 28nm for data center usage.   14nm will be the generation after that.

maybe profit margins will be so low, that another node won't pay for itself. Ever. We maybe stuck at 28nm for a long time

Indeed. New R&D is already a game of diminishing returns, because of the current leaders in the field quickly catching up with the technology gap. And by pricing old tech low enough you can always undercut the pricing for new hardware - making a ROI on R&D impossible - unless there are other factors (production cost of old tech). It simply doesn't make economic sense to throw $100 million dollars at chip development right now for state-of-the art IT, without also dumping $1 billion into bitcoin.

One of the more likely scenarios is that mining will become a defensive business, which is used by large stake holders to govern the bitcoin. This leads to a net loss business for mining itself. There will be profits - but they will go to the companies who produce & maintain the equipment - and the money will come from the entities who have an interest in hashpower and the algorithm.