Why would you want a VM with only a single process?

Why? Long term or short term answer? It is like driving an old diesel car. It might do 500.000km or it might be 300.000km before it dies. Maintenance goes up and average price goes down when you get to the higher mileage. But, I would not bet my business on anything over 250.000km. Prepare for replacement The average shareholder does not drive a car that old.

I think someone is taking too much risk with my money when a crucial part of that business may just break down.

It is not more important than 1M$ to you. That means you have a small business or like risks.

to create a verifiable fast RNG. There may be other parts of the kernel that can be optimized with some HW acceleration.

You don't get to choose business plans, if that is what you are saying.

You can always become an business man yourself and be one of the super rich ruling class yourself. And make better decisions. However: The existing ones already proved that they (or their predecessors) did quite a fine job. But please don't let that stop you to do a better one.

And for some comparison of things, I found this site to most informative (posting this link for the third time)

and some more (e.g. a comparison).

slower turning fans: quietness.

Water cooling is more noisy, by a little bit. The only thing I hear on my system is my harddisk (I run a stock i7-3820 with memory at 2400 for a memory intensive application). If you want to overclock by a large factor, water cooling can do that better.

See here for a comparison between two popular air and water cooling solutions.

I have the TPC-800 which is large and I make sure I always transport it very gently. But there are specifications for the weight it should be able to handle. Those specifications are for shipping conditions. Most motherboards have a special metal plate on the back side of the cpu mount (and it comes with the cooler too).

I would call it 'unverified statement' or 'rumour' instead of bullshit. His uid counts for something as do his previous postings. It is a massive claim though.

s/ath/ad/

A little translation error.

But Sir, I am transporting one way path encryption keys.

There is a comparison of blacklists: http://dnsbl.inps.de/analyse.cgi?type=monthly&lang=en

I send many millions of emails a week for mailing lists. I never get blacklistings 'for nothing'. What is annoying about blacklists is that they see so many spammers that they fail to see that normal people make mistakes (a lot). Since it is 'one strike and you are out' it is difficult to fix a situation for a legitimate situation.

Example: A medium size company mails its 2000 customers, this time in Finland, and we end up with a blacklisting. This customer comes up with a new list: The opt-outs and bounces (he is a sucker). However, we see 6% more permanent/hard-bounces than in the list. This can be because the spamfilters identified the spam and replied with hardbounces or it can be because the hardbounce list is bad (and there are more spamtraps). I already identified the list earlier as 'bad'. If no blacklist would have bothered, I would have allowed an opt-in mailing, asking recipients (again) for permission. Now I have to tell them to go to another ESP because I do not want to risk be a 'repeating offender' in their eyes. For the recipient, the first would be better. This is not a question you can ask a blacklist guy.

I guess the bottom line is, it is hard to communicate with blacklist people to solve legitimate situations. How can you fix that (without offering bad guys a vector for DOS-sing them)?

There is plenty of intelligence in the world and it is not helping. I would call human intelligence mediocre at best.