Schneier on Security

Nick Levinson • January 1, 2023 11:28 AM

@Who?:

Now when I look it up in Google snippets, I see you probably are right: rollback may often be possible, depending on the device. I had thought that a firmware update could change firmware in places and ways that a previous version could not anticipate so that a rollback would miss bits to roll back, but apparently the last pre-update version may be preserved as an image so that rollback could be reliable. But I’m not sure and please don’t rely on the last 2 sentences unless you do your own research and independently agree with them.

There isn’t much posting of hardware questions on this blog and even less of answers other than security implications of broader interest and what there is, other than security issues, may be on this blog by mistake.

Since you seem to want to protect your privacy, you may want to ask someone else whom you trust to keep your identity secret to find out for you, even to post your question for you and to check for answers and pass them along to you. It is possible that your question combined with any prior contact or history would identify you with fair probability, so that may affect your decision on asking. There used to be anonymous forwarding services, but I don’t know if they still exist. There are ways of anonymously reading the Web, such as the Tor browser, but not all websites permit full interaction through Tor.

Nick Levinson • January 1, 2023 9:20 PM

@Phil:

The Intel Management Engine is mentioned on this website. Do this search in Google (although you won’t have the privacy protection of DuckDuckGo, DDG doesn’t offer Google’s search results):

“Intel Management Engine” site:schneier.com

Include the straight quotation marks as shown to push irrelevant results further down.

For info on the “site:” search operator (note that there’s no space on either side of the colon in the search string), see https://developers.google.com/search/docs/monitor-debug/search-operators/all-search-site.

Nick Levinson • January 2, 2023 1:42 AM

@JonKnowsNothing & @Clive Robinson:

Nuclear war is survivable; you both seem to agree on that. JonKnowsNothing’s counterargument, while he agrees on survivability, is that it would be fast and global and survival is not thriving. I’m not sure that there are enough nuclear weapons worldwide to render Earth uninhabitable in a minute; but I agree on the difficulty of surviving. If too many of us die or are incapacitated during the strikes, we could not long sustain the means of supporting 8 billion people now on Earth and therefore we could not sustain most of the survivors. I estimated that if we all switched to gathering and hunting tomorrow morning and stuck with it then, in about a month to a month and a half, not counting murder or the need to find enough potable water near survivors (we can survive only a few days or less without drinkable water), and just because of food supply, out of every 35 people 34 would have to die. How much acreage for hunting-gathering is available and how many people that could support can only be estimated, so numbers may vary; but any reasonable estimate will be dire.

I do not know how to prevent war. I don’t think anyone else knows, either. You can sometimes prevent a particular war, but, in the larger scheme, there are always some people who sometimes want to kill entire societies and almost everyone wants to ensure their own defense against being killed. Therefore, war will be waged from time to time.

All types of weapons will normally be used. There can be limits. A limit could be unilateral when it appears to be too expensive to invent, build, or use even in the face of existential threat or demand (since key resources are never unlimited) or when use would likely fail if the failure would have the consequence that the enemy would gain an increased likelihood of victory. Where a limit would be mutual among belligerents, when they find it in their interest to limit they can and may. One line of discussion among war planners is on the advantage of keeping an enemy head of state alive and in office, i.e., no decapitation, because only the head may have the local political clout to surrender. Hirohito said he made only 2 decisions on his own and that was one of them. The Imperial Japanese generals knew of his decision and probably could have done something like throw fuel oil onto the runway to delay the Swiss ambassador’s take-off (to deliver the message to the Allies to start the surrender process) by a few hours while the generals tried to change the emperor’s mind; but they didn’t.

The claim that “[n]o battle plan survives first contact with the enemy”, even if widely said, is no more than an excuse not to plan and an excuse to ignore commanders. We all know that enemies don’t cooperate and planners know this, too, but not all plans are useless. It may be that militaries that are efficacious delegate a lot of authority to front-line servicemembers, so that it may seem that tactics are unplanned, but even there you’ll likely find planning.

I mentioned EMP because this is a security blog, especially on computer security, and the threat of EMP to computers is substantial.

People who want something that hasn’t been obtained in a long time often say that someone else wasn’t willing to spend the money. Those discussions as I hear them rarely include any numbers about how much it would cost and from where that money would come. For example, we could harden every target; but, never mind unwillingness to spend the money, we don’t even have that kind of money; and we don’t know how to get it in time. Sell buildings? To whom? A few, yes, but try selling a million of them and you get only fire-sale prices. And so on for other ways of raising money. Print money? How long before hyperinflation (>=100%/year) sets in? How would a leadership manage the domestic political fallout of that? Carter lost the presidency over inflation and it wasn’t that severe.

We probably don’t know definitively why Putin hasn’t launched a nuclear weapon. The U.S. is reportedly still being wary of being too provocative of Russia precisely because it is a nuclear power, so I doubt the U.S. military leaders feel confident that Putin and his top generals won’t use them.

Nick Levinson • January 2, 2023 12:38 PM

@Clive Robinson, @Winter, @JonKnowsNothing, @barfa, & @echo:

@Winter & @JonKnowsNothing:

The Thirty Years’ War may have caused 30% population loss, but not worldwide. I gather that was mostly in Germany and among foreign belligerents. According to Wikipedia when I looked (not a reliable source but its sources probably are and I didn’t have time to dig deeper), that war in 1618-1648 caused about 4.5-8 million deaths; the Population Reference Bureau says the world population in 1650 was about 500 million. That death rate was thus under 2%. The same Wikipedia page lists the war that had the most deaths as World War II, at 70 to under 120 million deaths; the same PRB table says the world population in 1950 was 2.499 billion. That death rate was thus slightly under 5%. Both wars were followed by substantial growths in global population and economy, although war is not always how we’d like to get those growths. Among other reasons is that massive population loss ramifies in equally massive fertility loss and therefore a delay in population recovery and thence in economic recovery. So, if the public demands that a war have a supremely good justification, the demand wouldn’t be surprising.

@Clive Robinson:

I never use the cloud. Great name, but I can’t use it when I can’t go online and I think when I want a full choice of OSes. One hopes security is top-notch but I’d like more control of my own security. No, thanks, cloud providers. The one use for it I think is to store securely-pre-encrypted backups (not encrypted merely simultaneously with uploading) as offsite, which would require storing the cloud password offsite; and maybe we wouldn’t have a good enough cipher for that (a one-time pad on just a gigabyte would usually be insane), so maybe the cloud is bad even for that.

If EMP is less of a threat than it used to be, that might explain why I couldn’t locate the medical lit project meant to educate doctors on how to provide half-way decent care based on pre-computer medicine, which at least would be better than, say, medieval medicine. (It also had the problem of expensively having to print and store the lit now at every major hospital and many other loci, along perhaps with some tools and chemicals no longer needed for modern practice, because storing the lit in a computer system would make it unavailable when needed.)

Wine is a nice idea but hopeless, last I heard, because Microsoft updates its products too often and wants the Wine developers to be unable to keep up, so incompatibilities will keep cropping up, enough of them critical to write Wine off.

@Clive Robinson, @barfa, & @echo:

I’m leery of relying too much on analyses of Putin’s private thoughts. There may be plenty of useful evidence of them but he isn’t alone at the top, even if he has sole authority at the top. He has generals and other top leaders and evidently they’re still largely on board with what he’s commanding. He may be micromanaging but micromanagement comes at a cost of time and, if he brings people with him to watch, their time; and skip-level meetings have their own costs.

I suspect much of the resistance in Russia to joining the Russian military to fight in Ukraine is not about objecting to the objective of the fight but about having to join a losing side regardless of objective.

Nick Levinson • January 2, 2023 5:15 PM

@lurker:

The article about the 6 cities was likely published a coupe of years before the invasion into Ukraine that came with a nuclear war threat, so the only connection is about large escalation, and the preparation advice was not about any specific threat.

Operation Tumbler was a 1952 test. It seems to have been more a test of effects, not a test for operational acceptance of a new weapon. The two purposes can overlap, but perhaps a weapon design was not validated enough to justify mass production or the type of weapon the test could have supported was outlawed by treaty.

I agree that recovery is possible, but it is slow and expensive. While a city is recovering from having fallen behind, it is not advancing from where it used to be before recovery was first necessitated. So, if war, in general or a particular one, is not a good idea, it should not be allowed to happen.

Nick Levinson • January 2, 2023 9:28 PM

@JonKnowsNothing, @lurker, @Clive Robinson, @Winter, & @echo:

A military and intelligence problem due to secrecy:

Scholarly research includes relying on the best evidence, such as by interviewing or examining original sources, making those sources available to other scholars, even competing scholars, on their demand, conducting tests that use random sampling and control groups, and publishing the test method so that unrelated, perhaps competing, scholars can replicate the tests.

Intel and military agencies require that same quality in their results for their most important information needs. But, even when they do their best to validate their conclusions, they can’t use most of those best practices. So, inevitably, they have a problem when their choice is between lower quality of results or spilling the beans to the enemy and thus immediately obsoleting the results, harming the nation which the agencies serve. Note that scholars publishing for unlimited readerships do not conform their methods to those that military and intel agencies need.

On nuclear weapons: Maybe it is politically and sometimes legally impossible for any nation to test major nuclear weapons as much as, say, a home appliance manufacturer tests its unclassified consumer products. Henry Kissinger, in his book Diplomacy, pointed to the problem of testing a U.S. ICBM: what if it exits its silo and promptly tips over onto Kansas? Imagine the Russian asking if we’d please fire a no-warhead ICBM at them so they can test their interdiction and counterattack systems: the U.S. military would say something like, you’re pranking us, right?

At least with intel, there may be a chance to have multiple research efforts by leaders who hardly know each other. With major nuclear war systems, there may be no research feasible beyond a certain point, by any nation, not even a first try.

So no nation may be able to use nukes until even an enormous error margin becomes politically acceptable to that nation. That margin includes the prompt death or injuring of many of its own nationals and harm to its physical things and its allies.

The comment on hotheads is interesting if it came originally from a commander who was supposed to command the servicemembers who would be the hotheads. That’s an admission of failure and I’m glad if someone had the courage to so speak and if someone in command was open to listening to it without punishing the speaker.

Nick Levinson • January 6, 2023 8:40 AM

@drafted lurker & @JonKnowsNothing:

IR reflectivity of duct tape:

It depends on the tape’s brand and batch, since, in my observation, it’s not a feature on labels and therefore is not offered, so the manufacturers’ standards may not consider this, so the reflectivity or lack thereof may be accidental. Also, what I see in stores is often not exactly called “duct” tape but may have a brand like “Duck” and simply look like duct tape.

Also, IR is in a frequency or wavelength range, and light sources, transmission media (e.g. air and other reflectors/nonreflectors), and receptors (in this case, cameras) may send, transmit, or receive all or only part of that range, and that, too, may vary by brand and other characteristics.

How much consistency you need may depend on how critical is your concern.

Nick Levinson • January 6, 2023 8:55 AM

@drafted lurker & @JonKnowsNothing:

IR reflectivity:

Also, Windows software to evaluate IR, if there is any, would depend on your computer’s physical ability to sense IR and to pass that info to the software. Many lenses, maybe because of their coatings, may block most or all infrared light. The hardware beyond the camera lens must also be able to handle IR.