Dan LuuFeb 17, 2023

I think it's too bad that, for things with bigger scope than incidents, people generally don't postmortem them to figure out what happened.

For one thing, I think it's just sad that a lot of these stories just get lost. But, maybe more importantly, it means it's basically impossible for the lessons learned to spread.

You see a lot of these kinds of analyses if you read military history, but a lot of it is speculative because it's too far in the past to be really sure about the specifics.

Show threadDan Luu

But with tech decisions, you can talk to everyone who was involved in key good or bad decisions and, even taking into account that people are unreliable narrators, I think you can usually get a good picture as to what happened.

I know people who do this kind of digging, but I don't think any of them are interested in writing up events in detail for fairly obvious reasons. If you mention specifics, you're going to make a lot of enemies, and if you don't, it's not really useful or interesting.

Feb 17, 2023 at 8:58amWeb
Show threadDan LuuFeb 17, 2023

I've been thinking about this recently since it's been a good time to catch up with Twitter folks about what decisions led to successes and failures 1-10 years down the line (and I mostly mean pre-acquisition decisions, though post-acq is interesting as well).

If you drill into specifics, it's easy to see that most standard advice is useless, e.g., if you ask, for one of Twitter Platform's most costly mistakes, why it happened, no piece of oft repeated standard advice would've stopped it and

Show threadDan LuuFeb 17, 2023

in fact, the standard advice "focus on your core competencies and outsource everything else" (also discussed in https://danluu.com/nothing-works/) was a (small) factor in making the wrong call

For the major factors, there were a lot of moving parts and no advice that can fit into a thinkpiece could make sense of the complexity of the situation

BTW, I think this is also fairly obvious w.r.t. military history. Of course real situations are complex and no 1000 word blog post is going to save you in battle

Why is it so hard to buy things that work well?

Show threadDan LuuFeb 17, 2023

E.g., if you look at some particular battle, like Waterloo, and ask, "if Napoleon's marshals could've read On War (written from 1816 and 1830, with the benefit of hindsight from the Napoleonic wars) before the battle, would it have helped?

IMO, the answer is no. There were a number of key mistakes, e.g., Marshal Ney's unsupported cavalry charge, not spiking artillery when they had the chance, etc.; no simple piece of abstract knowledge they didn't already have would've prevented these.

Show threadLaurence TrattFeb 17, 2023
@danluu In one sense I agree but in another, if I take your argument to its logical conclusion, On War can't help anyone, ever, which I think neither of us would agree with. But perhaps it's better to say that any piece of advice has a scope it applies to, and that scope is never infinite. For example, careful reading of On War wouldn't have stopped Ney's charge, but (OK, this is a bit tenuous, but you get the idea!) it might have stopped Napoleon offering battle in the first place.
Show threadDan LuuFeb 18, 2023
@ltratt I'm fairly skeptical of "high-level" advice over time that's distilled wisdom of some kind since it generally gives a lot of advice that's right in some circumstances and wrong in others, e.g., to pick a random example, "Thus a great positive success can never be obtained except through positive measures, planned not with a view to a mere state of “waiting-for,” but with a view to a decision, in short, even on the defensive, there is no great gain to be won except by a great stake."
Show threadDan LuuFeb 18, 2023
@ltratt Clausewitz concludes that from Napoleon's Russian campaign, but there are historical instances where you might come to the opposite conclusion and he doesn't really explain how you can tell which situation you're in. I'd say this is in contrast to something like Shaw's Fighter Combat or Kotov's Think Like a Grandmaster, so I don't think this is true of all reading on tactics or strategy, but I think it's common in tech (there's one such article on the front page of HN right now).
Show threadLaurence TrattFeb 18, 2023

@danluu I wholeheartedly agree that all advice must be filtered through one's context, and also pure common sense. But in terms of the specific Clausewitz quote, I'm actually racking my brain to think of a historical counter example, and currently failing...

[For clarity: I don't agree with everything Clausewitz says, and he is oddly myopic in some regards. Most obviously so in naval matters, which is unfortunate given that naval blockades played a big part in later German history.]

Show threadLeon BambrickFeb 17, 2023
@danluu if you outsource something - you can’t wash your hands of it completely. You have to know enough that an expert can’t bullshit you *too* easily. You don’t have to be a practical expert yourself. It’s an order of magnitude less effort, but it’s not zero effort.
Show threadPantaFeb 17, 2023
@danluu if people still wrote personal diaries, that's the kind of information that could at least resurface decades later.