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.

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.

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;

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