Remember when choosing DDR5 meant obsessing over the lowest possible CL latency?

Well, up until it was pretty much too late, they probably were. The issue was not that they were concealing risky assets by bundling them with good ones, the issue was that the way they were doing so hinged on an assumption that was entirely reasonable based on all historical data to that point, but which would prove catastrophically wrong. Specifically, the assumption that you could mitigate default risk through geographic diversification. You can absolutely mitigate the riskiness of a portfolio (or CDO) of risky assets through ensuring those assets are not strongly correlated with one another. At the time, it was well understood that real estate assets were locally correlated, and so a good loan could end up upside down if the local housing market tanks. And so any given CDO rated according to its constituent assets could reasonably be seen as less risky when its loans were sufficiently spread around geographically. The problem was that the subprime crisis hit basically all at once, all over the country. So what on paper looked like a nice, diverse portfolio did not behave like one because the riskier assets everywhere across the country started defaulting all at once. At the time those CDOs were rated, it was entirely reasonable to view this as a pretty unlikely occurrence. Obviously this was wrong. And there were absolutely signs of this that could have been seen, and were seen by some. But that most entities saw the problem too late wasn’t some conspiracy, and where it may be attributable to motivated reasoning, it was (again, at the time) well founded and entirely reasonable motivated reasoning. The zeitgeist in the years since has painted this as a crisis ultimately attributable to greedy recklessness and basically fraud in the financial sector, and certainly the nature of the financial sector’s real estate instruments were the fuel that made the crisis as bad as it was (and merited the additional regulation to follow), the ultimate attribution to what caused the problem was the subprime lending itself and (penultimately) the faulty (but at the time reasonable) assumptions underpinning our assessment of geographic risk diversification.