Hacker News Reader

Comment by nostrademons | Hacker News Reader

He died in 1986, so everything in the Virginia program happened well after his death. His forced retirement in 1982 was largely because he made a lot of political enemies by making a stink about General Dynamic's flaws in workmanship on the Los Angeles class. His whole management style was to push for accountability and sweat the details.

I think what you're country is seeing is general flaws in quality for basically everything American-made these days. Which, if I'm being totally blunt, are because Americans are lazy and don't care about quality and basically do the minimum that won't get them fired. Hell, I started my career caring about quality and craftmanship and now I do the minimum that won't get me fired too, because there's no point busting your ass for some lazy manager up the chain to take credit for it.

View on HN

Rickover's final testimony to congress in 1982 speaks about this:

Corporate Power A preoccupation with the so-called bottom line of profit and loss statements, coupled with a lust for expansion, is creating an environment in which fewer businessmen honor traditional values; where responsibility is increasingly disassociated from the the exercise of power; where skill in financial manipulation is valued more than actual knowledge and experience in the business; where attention and effort is directed mostly to short-term considerations, regardless of longer-range consequences.

Political and economic power is increasingly being concentrated among a few large corporations and their officers - power they can apply against society, government and individuals. Through their control of vast resources, these large corporations have become, in effect, another branch of government. They often exercise the power of government, but without the checks and balances inherent in our democratic system.

With their ability to dispense money, officials of large corporations may often exercise greater power to influence society than elected or appointed government officials - but without assuming any of the responsibilities and without being subject to public scrutiny. Woodrow Wilson warned that economic concentration could ''give to a few men a control over the economic life of the country which they might abuse to the undoing of millions of men.'' His stated purposes was: ''to square every process of our national life again with the standards we so proudly set up at the beginning and have always carried in our hearts.'' His comments are apropos today.

Summary: https://www.worldfuturefund.org/Articles/rickover.html Full testimony: https://www.jec.senate.gov/reports/97th%20Congress/Economics...

“Show me the incentive, and I'll show you the outcome.” -- Charlie Munger

In other words, if you want to solve a problem you need to align the incentives. If you don't your solution is likely to fail, especially on larger scales.

"measurement disfunction" - measuring something your people do leads to it being the only thing the people do. And some things are hard to measure. When the decided they could make trams run on time by giving penalties to drivers who were late, trams stopped stopping for passengers. True story. Leadership (vision) and arranging for resources to get the job done - that is, I think, the moral of the Rickover story.

There's a quip about organizing sales orgs that resonates with me:

Expect every salesperson in your company to optimize their own compensation. If you want to change their behavior, then change your sales compensation plan to drive the behavior you want.

I don't think this is appreciated enough at the management and CEO level of engineering-based companies.

How many engineering-tied metrics were explicitly part of Boeing's CEO's comp plan? If "Boeing makes less money, but produces safer planes" = more money to the CEO, then that happens.

And yet we're suprised when, absent that, individuals act in their own self interest and optimize for bad engineering outcomes.

Sales organizations are often the most game-theoretic rational folks in an organization: a large fraction of their compensation is directly attributable to recent efforts.

That doesn't make them honest or efficient or trustworthy, just canny and responsive.

By way of contrast, operations can only be seen negatively (if things are working, that's normal, and if things are not, that's bad) and development usually has conflicting incentives (stability! and features! and responsiveness! and bugfixes!) with cycle times and group responsibilites obscuring attribution.

This is why founder mode "works".. it is a way to overcome game theory* .

(I hope that does not mean that resurrecting George Washington is necessary)

*At least in the von Neumann-Morgenstern style

This expands to any department, or really any human activity.

I had an econ teacher who loved the phrase Rules, Incentives, Action, Outcome. She even had pens with RIAO printed out for students. The meaning is that each step led inevitably to the other, but also that you could start at the outcome end, and work your way backwards to what rules created that outcome.

Most of the powerful corporations in 1982 no longer exist or are much smaller because there is great turnover in corporate power. Governments don't lose power unless they lose a war, a rare event today that was common 100 years ago.

Especially look at military contractors, which consolidated significantly since the early 1990s, an intentional change desired by the Department of Defense.

> Governments don't lose power unless they lose a war, a rare event today that was common 100 years ago.

A good point, but of course they do lose power with every election - that's the point, a peaceful transition of power.

"governments" may not, but the political parties running them certainly do. And changes in political parties can change what services government delivers. Heck, you don't even need a change in parties for that to happen.

The USSR would like to enter the chat, but is unable to for existential reasons unrelated to combat.

(Also much of the rest of the Warsaw Pact, and others around the world.)

> reasons unrelated to combat.

Afghanistan is very much a reason why they no longer exist.

One of many modest-sized if not necessarily small cuts. But Afghanistan did not, say, defeat and invade the Soviet Union.

It out-endured them within its own territory. Much as has happened to the US several times, with resulting embarrassment but not regime change.

Military spending on the part of the US, falling global oil prices, multiple internal contradictions and failures, revolt within Warsaw Pact states, and growing internal dissent were far more significant contributors to the ultimate fall.

Britannica and World Atlas give similar accounts:

<https://www.britannica.com/story/why-did-the-soviet-union-co...>

<https://www.worldatlas.com/articles/10-reasons-for-the-colla...>

I have very little sympathy towards the Soviet empire, but the war remained on the Afghan side of the border, and was of a limited scope for most of its duration; construing the Soviet defeat into a direct cause of regime collapse is an extraordinary claim warranting extraordinary proof.

For example, one wouldn't claim that the de-facto American loss of the wars in Iraq and Afghanistan is directly affecting the stability of the American regime, would they?

Coming back to the Soviets, I would concur to a claim that the Afghan War was a failed national project, alongside the Chernobyl tragedy and the ecological disaster of the Aral Sea. The narrative I think proper here is first and foremost is the ordinary Soviet citizen's loss of trust in the Soviet way as such, rather than the specific fear of a foreign invasion.

Point taken, though I'd like to charitably think Americans are 'lazy' the way Haskell is lazy or Japanese Kanban or Jidoka is lazy. As MisterTea points out, the VCs invest in metrics, management pushes on metrics, and productive engineers deliver metrics at the cost of bugs and tech debt. Then we engineers wait for bug reports to percolate down, and we respond only as needed. On Sundays we repent by writing elegant high-quality code that nobody uses. We feel deeply that something is wrong.

> Which, if I'm being totally blunt, are because *American managers* are lazy and don't care about quality and basically do the minimum that won't get them fired.

FTFY.

I care but I am not allowed to. VC only cares about numbers that are larger then previous numbers. They hire management who tells us to make the numbers go up. those numbers go up with production so production is king. Production can be increased with a decrease in quality if we degrade the process or improving the process preserving quality. Improvement takes time and money - two hard things. So degrade it is. It's cheap, fast, and makes management look like a super star. If management doesn't give a shit, why should you?

If your n+1 doesn’t give a shit, why should you stay and enable this carelessness?

Because you get paid no matter what. edit: to add, this is a recent development so of course I am looking to leave.

Very valid reason. But beware that it does not weigh on you and burn you out. Been there several times. It always hurts, but I should have left even earlier each time.

Reply on news.ycombinator.com