Depends on particular unit (possibly batch).  Depends on room temperature.  Depends on air flow in room.  I run all my fans at 35% and this is OK in my room provided that the temperature in the room stays below 30 degrees C, even on units that are slightly overclocked.

This has happened to me too, several different units.  On one occasion it was correlated with a temporary loss of internet connection.  This affected one unit, but the others worked OK.  The symptoms were that the O/S had crashed.  The miner web interface did not respond, nor would it respond to a ping at its IP address. The only cure was to power cycle.

If you believe Slush's explanation, which I do, then it is seems inappropriate to refer to this situation a "block withholding attack".  A better term would be "block withholding bug". 

As to people who believe that Slush "owes" them something.  We all had access to the same information and had the opportunity to move our hash power months ago, once it was clear that something was wrong.  I stopped mining at the pool some months ago, once it was obvious that the bad "luck" was almost certainly not luck.

Better than "best" would be completely separate implementations with separate code bases, preferably in separate programming languages.  Separate teams evolving a common code base continues the risks of monoculture.

At least there will be something to catch up to and a reason to upgrade one's infrastructure if one needs to run a full node or wants to watch Netflix.

You can be damn sure if this private company started doing something the miners didn't like it would be replaced, probably within one or two days.  As I understand it, the code is all open source and it's just a matter of running similar code at new data centers and then reconfiguring some IP addresses in .conf files.

The example represents a severe performance bug in the bitcoin core software and should be fixed by software changes that make an inefficient operation efficient or restrict very large transactions if this proves too difficult. There is no reason to limit blocksize because of this.

An 80 second delay is significant only if you are mining, as it would represent a significant orphan risk.  This is already true with 0.5 MB blocks for slow internet service, such as the slow upload rate on typical DSL connections.  Most miners who don't have excellent internet connectivity do their mining via mining pools that are well connected and need minimal bandwidth for this purpose, independent of the block size.

The number of blocks nodes received averages out to be approximately the number of blocks these nodes send. "Number of takeoffs approximately equals number of landings." If you have a well connected node with lots of bandwidth then it's possible you will send out more data than you receive, but that is unlikely to happen if you have limited bandwidth.  Thus the network average remains 1 to 1 (except for new nodes).

If a new node starts up then it will have to receive each block once.  If an incompetently run node keeps crashing and losing the entire block database and has no backup then this will happen multiple times.   This part of the problem can easily be fixed by nodes with limited upstream bandwidth deprioritizing transmission of older blocks during time of congestion. (One of many possible network optimizations that can and will appear should they be needed.)

Google's computer is a D-wave adiabatic quantum computer.  It is not a general purpose quantum computer and can not be used to run variants of Shor's algorithm or Grover's algorithm.  This means that it poses no threat to bitcoin, either to signatures or hashes (including proof of work).

Two words:  hardware wallet.
One word:  Trezor

If you are running a Bitcoin Unlimited node, be sure and check that it is open and being counted by the Bitnodes snapshot.  https://bitnodes.21.co/nodes/?q=/BitcoinUnlimited:0.11.2/

You might check this page periodically to make sure that your node hasn't gone missing for some reason, particularly if you've had to restart it.

I am running a node because doing so is the only way to see how bitcoin actually works. I am not running it to help the network in any way, nor somehow magically keep the fees low so I will pay less.  It could be argued as a consequence of my low upstream bandwidth I am actually "hurting" the network by virtue of downloading more that I upload ("leaching").  I do get the benefit of greater security and privacy from running the node.  Also, the same machine is running other servers, and I would probably be running it 24/7 even if I stopped running a bitcoin node.  I have some older machines that I could use for servers, but they consume enough electricity running 24/7 that the newer machine has already paid for itself.

Were I to pay exorbitant transaction fees I would be doing nothing to help the growth of bitcoin.  It would be a foolish decision, because the value of the bitcoins I hold would certainly go down.  Of course, if I were astute, I might pay one large transaction fee and dump all of my bitcoins, shut down my node, and stop wasting my time replying to trolls and sock puppets.

The 8 GB RAM module for the computer that runs my Bitcoin Unlimited node has 8 GB of RAM.  I paid $67 for this memory module 13 months ago.  I note that Amazon is selling the same module today for $35 USD.  It is unreasonable to cripple bitcoin to support users running obsolete hardware.

Just a nit, but there is a subtle point. The primary limiting technological factor for transaction rate is bandwidth and latency.  (This changes the wording, so as to eliminate counting games such as Seg W.)

One does not have to "trust" a protocol.  Actually, in the case of bitcoin, there is no protocol separate from some version of bitcoin software.  One can inspect that software and see in detail what it does and how it does it.  One can build the software from source code if one doesn't trust the downloadable binaries. If one lacks the necessary skills to understand the source code one can hire people to do review on your behalf. 

Now contrast this situation with you you face as the user of the fiat banking system, or a fractional reserve bitcoin exchange such as Mt. Gox.

Maxwell has claimed that he has experience scaling large systems.  He used to work at Juniper. Presumably he worked on some cryptographic aspects of their products as well.  Draw your own conclusions.

Downloading the blockchain was slow until a few months ago, when a change in networking software provided a big speedup, making verification the limit on most computers.  In the pipeline is faster software for validating signatures, which should provide another large speed up.  As the size of the blockchain expands (as it will with time unless Bitcoin dies) there will be more pressure onto software improvements to speed things up.  This pressure will be maximized if the block limit is raised or removed.  Finally, it is possible to prune the block chain and this is in the process of being phased in.

The code speedup is a useful improvement.  However, it does not increase the capacity of the bitcoin system since the capacity is limited artificially by the 1 MB blocksize limit. 

Segregated witness does nothing for light clients.  It may increase the network capacity in a complex way, because it changes what counts in the 1 MB block limit.  It does not reduce bandwidth usage at all, so it amounts to a complex way of partially bypassing the 1 MB block limit.  The same effect could be achieved with much less effort by a simple increase in the block size, say to 2 MB.  Segregated witness also requires all of the clients to be updated in a non-trivial way, including SPV clients which don't care about the block limit. 

What you are seeing is a political game being played by bitcoin core to make it look like they are doing something.  For someone who knows how bitcoin works they are not doing good engineering.  In particular, they are not following the KISS principle. (Keep It Simple Stupid.  Don't do something complex when a simple change can accomplish the same thing.)

Beware of propaganda.  There is controversy regarding these changes.  There are other internet sources where you can see other viewpoints on this issue, such as r/btc.