Simple sabotage for software (2023)
erikbern.com
erikbern.com
The way to cripple a company is to get bad people promoted into management and have them optimise something plausible but not profitable. Like what happened in a lot of slow-fail engineering companies - insist on maximising the profit metric to the point where the actual product becomes compromised. That strategy can be intiated from almost anywhere in management too because a constant "what if we optimise for profit" [1] usually does well at meetings.
Being a destructive CTO is almost too easy. Just don't do anything much, weed out any competent people in the level below you and develop a culture of pushing blame aaaaalllllllll the way down the org chart, ideally out of the technical part of the tree, so that nothing broken gets fixed. When people catch on, allow blame to move 1 level up the org chart and do a big shakeup to clear out any institutional knowledge that might have built up in spite of you. That game can go on for years.
[0] I've seen people where "do things through official channels" and "demand written orders" would have made them more productive. We are our own worst enemies.
[1] Yes, the irony here is that naive optimising for profit is typically not profitable.
Turns out the dude has been doing that for about 15 years between various places and now is about to retire, blameless in the eyes of everyone who doesn't understand what happened.
Let’s set asside the fact that the document wasn’t written by the CIA.
The purported goal of this document was to provide practicaly applicable advice to the regular citizen who found themselves under enemy occupation. Most concretely to be given to the French people who did not like the German occupation.
You are talking about the strategy “working” or “not working” as if these are binary things. The goal here was not that these simple steps will bring Germany to their knees but to increase the cost of the occupation. To cause enough deniable friction which bogs down the resources and make everything just a bit more inefficient.
> In my experience workers like that exist naturally
They do. And that is the point. That is what makes these strategies deniable.
> and organisations are great at just sidelining them
If that is your experience I would love to work where you worked. In my experience when someone is following this strategy sidelining only happens slowly and at great costs. One of the many costs is people comitting avoidable blunders when they dismiss real and well reasoned objections in their haste to cut through a sea of useless ones.
> to get bad people promoted into management
Sure. But that takes time. You are thinking on a different time scale than the authors of this document were thinking about. The document was published in Jan 1944. The Normandy landings happened in June the same year and by the end of the next year the war was won. You don’t have time to slowly promote bad people into management. If a dude who read your booklet bumbles about a bit and delays the repair of a train line by days that is a win in this context. Nobody expected that Germany is going to collapse on their own just because enough people sabotage meetings and plug up toilets. (That is by the way also a suggestion from the manual 5.1.b.2. Somewhat less often cited than the points applicable to office work.)
Ones whose only significant effort (if any) is to be a mainstream employee in all other ways since nothing will ever make them productive or capable of efficient operation.
>> and organisations are great at just sidelining them
>If that is your experience I would love to work where you worked.
Me too. That does seem like uncommon good fortune. Too often these are the ones that get promoted into higher levels of mangement, it is so widespread among different companies it goes undetected in the way the CIA intended. Plausibility beats productivity.
What do you mean? The CIA has publicly stated that the document was written by the OSS, its wartime predecessor. [0]
[0] https://www.cia.gov/stories/story/the-art-of-simple-sabotage...
Wild Bill Donovan’s admirers and critics still argue over his legacy, but on one point they agree: His World War II Office of Strategic Service (OSS) became the Petri dish for the spies who later ran the CIA as well as the special operators who conduct some of the most daring raids the world has ever seen.
Four CIA directors — Allen Dulles, Richard Helms, William Colby and William Casey — learned the craft of clandestine warfare as operatives for Donovan’s OSS. Indeed, the daring, the risk-taking, the unconventional thinking, and the élan and esprit de corps of the OSS permeated the new agency.
So would the OSS’s failings: the delusions that covert operations, like magic bullets, could produce spectacular results, or that legal or ethical corners could be cut for a higher cause.Similar to how a branch of the US Public Health Service [0] originally tasked with malaria prevention became Communicable Disease Center (CDC) in 1946-67, subsequently renamed to "Center for Disease Control" and "Centers for Disease Control" (1980), "Centers for Disease Control and Prevention". It hasn't actually been called the "Center for Disease Control" since 1980, although many people (incl. journalists) still call it that.
Also, most countries' Department of Defense/Ministry of Defence were called Ministry of War or Department of War during WWII (and some only controlled the army, not all branches of military). [1] And the White House War Room was renamed the Situation Room in 1961. (RIP George Carlin.)
[0]: https://en.wikipedia.org/wiki/Centers_for_Disease_Control_an...
WWII was before all this, but we have decades of management experience about how to take ordinary people and make them productive in an industrial setting.
The CIA list isn't the inverse of that. They didn't have Dilbert then either, but it looks like the equivalent of mailing over some Dilbert comics. Maybe it is better than nothing and I don't fault them for trying everything but I've never seen evidence that these are actually effective hints at sabotaging an organisation. The office-work stuff, not the physical ideas which I assume are quite effective.
Now that the discipline exists it'd be interesting to get a group of great operations researchers together and have them come up with their own list of ideas then see what the overlap is. It might be quite small. Is it more damaging to have a great worker who gets one or two key thing wrong or a grumpy guts who does poorly at everything? I suspect the former, the sabotage handbook the latter.
This is for the people who, for whatever reason, can't be partisans but can do their part to ensure that they're not collaborating.
People learn very quickly to manipulate the reporting, modeling, and accounting.
Just less than two decades ago we had the entire financial system lending massive amounts of money to the worst possible borrowers and REPORTING massive profits.
Yes but only over a relatively long time period, which the overarching system itself is not incentivizing anyone to look at too carefully.
Naive optimizing for anything, while consuming that metric as fuel, has that effect.
A.k.a. “We will continue having meetings until we figure out why productivity is dropping”.
Not quite right. The Office of Strategic Services did that. The CIA was created only in 1947 several years after the end of the second World War in 1945.
I don’t know what is a big or a small stretch. I just know it is just not true. If you studied the history of the second World War, or the history of the CIA, or the history of the Cold War then that sentence sticks out like a sore thumb. It makes as much sense as describing how many Gatling guns Hannibal mounted per elephant.
I guess in a certain post-truth sense it is a smaller stretch of the truth to say that the CIA published it, than to say that the KGB did it. And in turn attributing it to the KGB would be a smaller stretch of the truth, than attributing it to the elves of Middle Earth. But what is common between all those three is that they haven’t authored this document, because it was the OSS who did it. (As the document itself conveniently, and very prominently states.)
Excellent. Continue the good work.
1) dang has to decide whether this is too snarky or not.
2) Please stay on topic, its about CIA or not.
3) Its hardly "frequently bringing up irrelevant issues".
Have i missed something?
We need to form a sub comittie to figure out if it is.
> If you flag, please don't also comment that you did.
I think in that light, number 1 is to foster an atmosphere of fear where anyone attempting to make things better will be confronted if things don’t go perfectly.
I feel like this, in some capacity, is borderline inevitable in the modern architectures with a bunch of external services, or at the very least will 100% require that your devs are connected to the Internet all the time to be able to do anything, vs systems that are 100% self hostable.
Or even just running a complex Kubernetes cluster with a service mesh and other solutions on the test/staging/prod infra vs loosely mapping to more lightweight options locally, unless you have a super beefy setup. And even then, if everything is split into multiple separate services far enough, you just won't be able to run everything locally, meaning that you need to use some of the components from shared dev environments which will inevitably lead to stepping on each other's heels.
Once you go past something like Docker Compose for local environments, things can go sour.
So I flip it around — production should work exactly like the development environments. If there's a a difference between the two that matters we update production to match. You want to talk to other services by a well-known container name instead of an env var we pass? Sure, whatever we can do that.
It's a miracle that somehow the company remains in business.
(Some may counter this with Hanlon's razor [2]; Never attribute to malice that which is adequately explained by stupidity.)
Plus calling these things "sabotage" isn't right even as a metaphor, and IMO confuses the issue. These people aren't on a third-party payroll, they are just self-serving.
No one is creeping into your office in the middle of the night. You invited them over, let them in, watched them piss in the fish tank, and then gave them a promotion and a raise. Now you're surprised everyone is lining up to fuck up the place? As usual the value judgements are just a distraction in matters of economics, look at the incentives.
Not sure if there's a name attached to this counter-principle, but it's just as important as the original in my opinion.
Deploying infrequently doesn't necessarily sabotage things if it results in well-written and tested code before deployments, which tends to create stable systems. If anything, it's the opposite of "move fast and break things." If you want to sabotage things, urge frequent deployment and minimal testing.
> This is from the 1994 music video Sabotage by Beastie Boys. The lyrics are mostly about technology leadership and developer productivity.
That caught me off guard. Well done.
1. A person who knows little of the topic assumes that talking a lot means knowing a lot. Hence long ChatGPT-like speeches. This has been my personal number 1 red flag since long before LLMs became a thing.
2. An incompetent manager is helpless in a situation when "everybody screwed up a little bit" as they don't pay attention to ownership or don't even understand its importance. Hence creating groups/committees to blur personal responsibility.
3. "Bring up irrelevant issues" is natural consequence of not understanding the topic very well, combined with forcing oneself to take part in the discussion. Pretty sure it's not even intentional. But either way, it's irrelevant from your point of view, but not from the the manager's, so it works.
4. Advocate caution at everything is because if the team screws up and attracts attention from higher ups, people at the most risk will be the lowest performers and the manager. It's a shared interest.
Security and compliance benefit heavily from this approach, and I'm sure it can be extended elsewhere as well (architecture reviews etc).
The article kind of missed that detail, but this form of sabotage doesn't even need to be that effective to deeply sabotage the organization/product, it just has to be bad enough to scare off good talent and keep turnover high. The stuff in the "hiring" section makes it even worse by preventing good people from even getting in the door in the first place, but eventually a company like this would be scraping the bottom of the barrel for candidates even without deliberately bad hiring practices.
Is the sabotaging bit here the "complex" part?
> Encourage a complex dev setup: running a service mesh with a dozen services at a minimum.
...
> Build in-house versions of almost anything that's not a core competency.
So if you need functionality A, B, C ... H, what do you do? Do you build them all in-house or do you have 9 different services each providing the functionality required?
maybe merge the threads?